SELF-HANDICAPPING AND DEFENSIVE PESSIMISM: PREDICTORS AND CONSEQUENCES FROM A SELF-WORTH MOTIVATION PERSPECTIVE

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Andrew J. Martin BA (Hons) MEd (Hons) (Sydney)

FACULTY OF EDUCATION UNIVERSITY OF WESTERN SYDNEY, MACARTHUR

DECEMBER 1998

Andrew J. Martin

ACKNOWLEDGEMENTS

I would like to extend my heartfelt thanks to my chief supervisor, Professor Herb Marsh. He has been the best teacher and role model I could hope for and someone who I felt always cared about my development as an academic. I will look back on my time as his student with great fondness and appreciation for many years to come. I would also like to extend many thanks to my associate supervisor, Associate Professor Ray Debus. I am extremely grateful for the time and advice he has given me in relation to this thesis and also in relation to the broader dimensions of my academic life.

I am grateful to all the Education students from the University of Sydney, the University of Western Sydney, Macarthur, and the University of Western Sydney, Nepean who participated in the survey process. I would especially like to thank the forty students from these institutions who kindly gave their time for in-depth interviews. In the process of collecting the data I was very fortunate to receive the support from many staff at each of the three institutions. Staff I would particularly like to thank are Dr Valentina McInerney, Dr Annette Holland, Kate Johnson, Alison Hine, Jenny Jones, Dr Llian Merritt, Neville Hatton, Margaret Pickup, Dr Stephen Juan, and Helen Watt.

I would like to thank Dr Alan Williamson and Tony Martin for their comments on previous drafts of chapters in the thesis. I am also grateful to Tania Martin, Brett Pickering, and Michael Creighton for their help in the piloting of the survey instruments. I would also like to extend many thanks to Dr Julie Norem for her feedback regarding the measurement of defensive expectations and reflectivity and for her assistance in the development of my thinking about various aspects of defensive pessimism. Similarly, I thank Dr Carol Midgley for her feedback regarding aspects of self-handicapping.

My deepest thanks goes to my wife, Mlane. Her love and support through all the emotional ups and downs of the PhD has been unflagging. This thesis is as much a testament to her stamina as it is to mine.

Statement of Authentication

The work presented in this thesis is, to the best of my knowledge and belief, original, except as acknowledged in the text. I hereby declare that I have not submitted this material, either in whole or in part, for a degree at this or any other institution.

CONTENTS

1. INTRODUCTION

2. SELF-PROTECTION STRATEGIES 2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.3 Self-handicapping Historical roots The variety of studies examining self-handicapping Consequences of self-handicapping Protective or augmenting? Summary of self-handicapping Defensive pessimism The strategic nature of defensive pessimism The distinction between expectations and reflectivity The consequences of defensive pessimism in the short term Longer term effects of defensive pessimism Summary of defensive pessimism Drawing together self-handicapping and defensive pessimism

9 9 11 12 16 19 21 22 23 25 28 29 33 33

3 SELF-WORTH MOTIVATION AND PREDICTORS IN THE PROPOSED MODEL 3.1 3.1.1 Self-worth motivation theory Setting the stage for self-worth protection: The educational context 3.1.2 Setting the stage for self-worth protection: Ability becomes tied to self-worth 3.1.3 Self-handicapping, defensive pessimism, and self-protection 3.2 Factors predicting self-handicapping, defensive expectations, and reflectivity 3.2.1 3.2.2 Motivation orientation Public self-consciousness 4 40 41 52 39 37 36 36 36

3.2.3 3.2.4 3.2.5 3.2.6 3.3 3.4 3.5 3.6

Stability and level of self-concept/esteem Perceived control Views of the nature of intelligence Attributions Integrating the predictors into a higher order structure The generality vs specificity of constructs A need achievement perspective Chapter summary

57 67 74 79 86 89 92 92

4. SUMMARY OF RESEARCH ISSUES AND SPECIFIC FOCI OF THE PRESENT STUDY 4.1 4.1.1 4.1.2 4.1.3 4.2 Quantitative analyses Factors predicting the strategies Outcomes following from the strategies Supplementary quantitative issues Qualitative analyses 94 94 94 95 95 96

5. METHOD 5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.3 5.3.1 Sample and procedure Materials Self-handicapping Defensive pessimism Affective and motivational predictors Academic outcomes Data analysis Brief orientation to CFA and SEM

97 97 98 100 102 103 107 108 109

6. TIME 1 QUANTITATIVE ANALYSES 6.1 6.1.1 6.1.2 6.1.3 6.1.4 Refining the instrument Method Refining the content domain Refining the number of items Refining the number of subscales 5

111 111 112 112 116 119

6.1.5 6.1.6 6.2 6.2.1 6.2.2 6.2.3

Summary of refinement analyses Confirmatory factor analysis of the refined instrument Measurement properties for the entire sample Method Confirmatory factor analyses for Macarthur and Sydney Testing for factor invariance across the three institutions and across sex

124 125 129 129 130

134

6.3

Higher order structure of the data and the role of respondents background characteristics 139 139 144

6.3.1 6.3.2 6.4

The higher order nature of the factors Assessing the role of background variables Examining the Time 1 process model and assessing supplementary issues

146 146 148 153 153

6.4.1 6.4.2 6.4.3 6.5

Time 1 process model Further testing of defensive pessimism A test of the interaction of level and stability of self-concept Summary of Time 1 results

7. EXPLORING THE TIME 1 PROCESS MODEL FROM A QUALITATIVE PERSPECTIVE 7.1 7.2 7.3 7.3.1 7.3.2 7.3.3 7.4 7.5 Qualitative perspectives in a quantitative domain Research issues Method Sample Procedure Data analysis Qualifications relevant to the data analysis Results 165 7.5.1 7.5.2 7.5.3 7.6 Academic strategies Affective and motivational predictors Academic outcomes Summary of chapter 6 166 176 189 194 161 162 162 164 156 156 158 161

8. DISCUSSION OF TIME 1 RESULTS 8.1 8.1.1 8.1.2 8.1.3 8.1.4 8.1.5 8.1.6 8.2 What factors predict the use of the three strategies? Task-orientation External attributional orientation Uncertain personal control Performance orientation Level of self-concept Beliefs about the nature of intelligence To what extent do self-handicapping, defensive expectations, and reflectivity predict academic outcomes? 8.2.1 8.2.2 8.2.3 8.3 Self-handicapping, self-regulation, and persistence Defensive expectations, self-regulation, and persistence Reflectivity, self-regulation, and persistence Active and self-presented self-handicapping and defensive expectations 8.4 8.5 8.5.1 8.5.2 8.6 Domain specificity of constructs Time 1 qualitative analyses Self-handicapping Defensive expectations Issues to be addressed at Time 2

197 198 198 200 202 206 212 213

214 215 216 218

219 221 224 224 225 227

9. TIME 2 QUANTITATIVE ANALYSES 9.1 9.1.1 9.1.2 9.2 9.2.1 9.2.2 9.2.3 9.2.4 9.2.5 Results First order CFA of Time 2 subscales Time 2 higher order CFAs Time 2 process model Testing interactions in the Time 2 process model Assessing a model in which defensive expectations predict reflectivity 7 Method Participants Materials and Procedure

229 229 229 229 230 230 234 234 242

243

9.2.6 9.3

The longitudinal data (Time 1 - Time 2) Summary of chapter findings

243 253

10. EXPLORING SHIFTS ACROSS TIME 1 AND TIME 2 FROM A QUALITATIVE PERSPECTIVE 10.1 10.2 10.2.1 10.2.2 10.2.3 10.3 10.3.1 10.3.2 Results Shifts in self-handicapping and defensive expectations Concomitant shifts in affective and motivational constructs identified in process models 10.3.3 10.3.4 10.4 Distinguishing features of first and second year university The role of cultural factors Summary of Time 2 qualitative findings 263 272 277 280 Qualifications for the Time 2 qualitative study Method Sample Procedure Data analysis 255 257 258 258 258 260 260 260

11. DISCUSSION OF TIME 2 FINDINGS 11.1 11.1.1 11.1.2 11.2 11.2.1 11.2.2 The Time 2 process model Generality of process models across Time 1 and Time 2 The role of anxiety in predicting academic strategies The longitudinal data Predictors of academic grades Factors predicting affective and motivational constructs at Time 2 11.2.3 11.2.4 11.3 11.4 Test-retest parameters Correlations among the strategies Qualitative findings Summary of chapter

283 283 283 284 286 287

288 290 291 292 295

12. A QUADRIPOLAR NEED ACHIEVEMENT REPRESENTATION

OF SELF-HANDICAPPING, DEFENSIVE EXPECTATIONS, AND REFLECTIVITY 12.1 Brief orientation to need achievement theory and a quadripolar representation 12.2 12.2.1 12.2.2 12.3 12.4 Results Time 1 Time 2 Discussion Summary of chapter 297 301 301 304 305 308 297

13. GENERAL DISCUSSION 13.1 Synthesis of self-handicapping, defensive expectations, and reflectivity from a self-protection perspective 13.2 13.2.1 13.2.2 13.3 13.3.1 Assessing the strategies in the context of existing theory Defensive expectations and reflectivity Self-handicapping Implications for education The role of the predictors in addressing students defensive manoeuvring 13.3.2 13.3.3 13.4 13.4.1 13.4.2 13.4.3 13.4.4 13.4.5 13.4.6 A need achievement perspective Summary of educational implications Limitations of the study and directions for future research Self-report data Measures used in the study The content domain and nature of the sample Statistical issues Consequences of the three strategies The qualitative data

309

309 311 311 315 317

318 324 327 328 328 330 333 336 336 337

14. CONCLUSION

341

REFERENCES

345

APPENDICES

373

10

LIST OF TABLES

Table 5.1 Table 5.2

Sample items from strategy subscales Sample items from affective and motivational predictor subscales

98

99 100

Table 5.3 Table 6.1

Sample items from academic outcome subscales Cronbachs alpha coefficients and correlations between math and verbal domains: raw; corrected for reliability; CFA uncorrelated uniquenesses; and CFA correlated uniquenesses

113 115 117

Table 6.2 Table 6.3 Table 6.4

Means, SDs, and t values for math/verbal subscale pairs Item analysis for active self-handicapping scale Correlations among central strategy subscales and both predictor and outcome subscales

121

Table 6.5

Correlations among strategy subscales and perceived control over future success and failure subscales 122

Table 6.6

Correlations among strategy subscales and esteem-relevant competence valuation subscales 123 124 126

Table 6.7 Table 6.8 Table 6.9

Correlations among strategy subscales and attribution subscales Factor structure of the final 22 factor model for Nepean (n=204) Correlations among central strategies and both predictor and outcome subscales for Nepean (n=204)

127

Table 6.10 Table 6.11 Table 6.12

Correlations among predictors and outcomes for Nepean (n=204) 128 Factor loadings for the 22 factor model: Macarthur and Sydney Correlations among central strategies and both predictor and outcome subscales: Macarthur and Sydney 132 131

Table 6.13

Correlations among predictors and outcomes for Macarthur and Sydney 133 134 135

Table 6.14 Table 6.15 Table 6.16

Invariance tests across the three institutions Invariance tests across sex Factor structure of the 22 factor model for the entire sample (n=584)

136

Table 6.17

Correlations among strategies and both predictors and outcomes 11

for entire sample (n=584) Table 6.18 Correlations among predictors and outcomes for the entire sample (n=584) Table 6.19 Table 6.20 Fit statistics for the seven strategy models Strategy factor loadings and correlations based on two factor higher order CFA Table 6.21 Predictor factor loadings and correlations based on six factor higher order CFA Table 6.22 Correlations among background variables and other first and higher order constructs in the model Table 7.1 Table 7.2 Table 7.3 Research issues to be addressed in Time 1 qualitative analysis Time 1 interview participants Summary of findings in relation to Time 1 self-handicapping interview data Table 7.4 Summary of findings in relation to Time 1 defensive expectations interview data Table 7.5 Summary of performance orientation findings for students high and low in self-handicapping and defensive expectations Table 7.6 Summary of external attributional orientation, uncertain personal control, and task-orientation for students high and low in selfhandicapping and defensive expectations Table 7.7 Summary of self-regulation and persistence findings for students high and low in self-handicapping and defensive expectations Table 9.1 Table 9.2 Factor loadings of Time 2 subscales Correlations among Time 2 strategies and both predictors and outcomes Table 9.3 Table 9.4 Table 9.5 Table 9.6 Correlations among Time 2 predictors and outcomes Time 2 higher order factor loadings Time 2 correlations among higher order and first order factors Correlations generated in the CFA of the parsimonious Time 1 - Time 2 data set Table 10.1 Table 10.2 Qualitative research issues addressed at Time 2 Sample details for Time 2 qualitative study 12

137

138 141

142

144

147 160 161

171

175

183

190

195 231

232 233 235 236

246 256 259

Table 10.3 Table 12.1

Summary of Time 2 qualitative findings Time 1 stimulus coordinates for self-handicapping, defensive expectations, reflectivity, and self-concept

282

302

Table 12.2

Time 2 stimulus coordinates for self-handicapping, defensive expectations, reflectivity, and self-concept 303

13

LIST OF FIGURES

Figure 1.1 Figure 1.2 Figure 1.3

Basic structure of the proposed process model to be examined Constructs incorporated into proposed model Structure of the proposed longitudinal model explored in the study

2 5

Figure 2.1

The proposed process model incorporating the association between self-handicapping and academic outcomes 19

Figure 2.2

The proposed process model incorporating the association between both defensive expectations and reflectivity and academic outcomes 32

Figure 3.1

The proposed process model incorporating the association between both motivation orientation (comprising ego-orientation, task-orientation, and avoidance-oriented performance) and esteem-relevant competence valuation and self-protection strategies 52

Figure 3.2

The proposed process model incorporating the association between public self-consciousness and self-protection strategies 57

Figure 3.3

The proposed process model incorporating the association between both level and stability of self-concept and selfprotection strategies 67

Figure 3.4

The proposed process model incorporating the association between perceived control and self-protection strategies 74

Figure 3.5

The proposed process model incorporating the association between implicit theories of intelligence and self-protection strategies 80

Figure 3.6

The proposed process model incorporating the association between causal attributions and self-protection strategies 85

Figure 3.7

Proposed higher order structure of affective and motivational predictors and its role in the proposed process model 88 110

Figure 5.1 Figure 5.2

Representation of higher order factor structure

CFA/Measurement component and structural component of SEM 110

14

Figure 6.1

Seven models used to explore the multidimensionality of strategies 140 149 159

Figure 6.2 Figure 7.1 Figure 9.1

Significant structural relations in Time 1 process model Model identified through quantitative analysis Significant structural relations (and higher order factor loadings) in Time 2 process model

237

Figure 9.2

Significant structural relations (and higher order factor loadings) in Time 2 Process Submodel 1 - Defensive Expectations 239

Figure 9.3

Significant structural relations (and higher order factor loadings) in Time 2 Process Submodel 2 - Self-Handicapping 240

Figure 9.4

Significant structural relations (and higher order factor loadings) in Time 2 Process Submodel 3 - Reflectivity 241 245

Figure 9.5 Figure 9.6 Figure 9.7

Basic structure of the hypothesised longitudinal model Significant structural relations in longitudinal model Significant structural relations in Longitudinal Submodel 1 Defensive Expectations

250

Figure 9.8

Significant structural relations in Longitudinal Submodel 2 Self-Handicapping 251

Figure 9.9

Significant structural relations in Longitudinal Submodel 3 Reflectivity 252

Figure 12.1

Quadripolar model of need achievement (Covington, 1992, 1997) 297

Figure 12.2

Hypothesised quadripolar model of need achievement incorporating the present studys strategies 300

Figure 12.3

Time 1 multidimensional mapping of the three strategies and self-concept 302

Figure 12.4

Time 2 multidimensional mapping of the three strategies and self-concept 303

15

APPENDICES

Appendix A Appendix B Appendix C Appendix D Appendix E Appendix F Appendix G

Academic Process Questionnaire - First administration Item wording and subscale statistics 29 factor CFA loadings (Nepean, n=204) Academic Process Questionnaire (revised) 22 factor CFA for Nepean and Macarthur (n=393) Modelling latent interaction effects Interview consent form - Time 1 and Time 2 qualitative studies

374 391 403 404 411 414

417

Appendix H

Semi-structured interview schedule - Time 1 qualitative study 418

Appendix I

Categories and Nudist node addresses - Time 1 qualitative study 421 424

Appendix J Appendix K

Subscales added to questionnaire at Time 2 Assessing the Elliot and Church (1997) performanceavoidance subscale

425

Appendix L

Semi-structured interview schedule - Time 2 qualitative study 426

Appendix M

Categories and Nudist node addresses - Time 2 qualitative study 428

16

Only man behaves with such gratuitous folly. It is the price he has to pay for being intelligent, but not, as yet, quite intelligent enough - Aldous Leonard Huxley (1932)

17

ABSTRACT

The present study examines predictors and consequences of self-handicapping and defensive pessimism (comprising defensive expectations and reflectivity) from a selfworth motivation perspective (Beery, 1975; Covington, 1984a, 1992; Covington & Beery, 1976). Consistent with self-worth motivation theory, self-handicapping and defensive expectations are proposed as two strategies students use to protect their selfworth in the event of potential failure, and in some cases to enhance their worth in the event of success. Using longitudinal data derived from undergraduate students from three institutions in their first and their second years at university, quantitative analyses, involving confirmatory factor analysis and structural equation modelling, tested models in which a set of affective and motivational factors was proposed to predict self-handicapping, defensive expectations, and reflectivity. These three strategies were in turn proposed to predict a variety of academic outcomes including self-regulation, persistence, future academic plans, and grades.

In students first (n=584) and second years (n=489) at university, two process models were examined and the patterns of relationships emerging at both times were broadly congruent: External attributional orientation, performance orientation, uncertain personal control, and anxiety, all positively predicted self-handicapping, defensive expectations, and reflectivity, while task-orientation was found to negatively predict self-handicapping and defensive expectations and positively predict reflectivity. In turn, self-handicapping and defensive expectations negatively predicted persistence and self-regulation, while reflectivity positively predicted these outcomes. A pivotal finding in the additional longitudinal model (n=328), beyond those derived in the process models, was the negative effect of Time 1 self-handicapping on subsequent academic grades. Application of the quadripolar model of need achievement (Covington & Omelich, 1991) provided an integrative conceptual rationale to assist substantive interpretations suggesting that reflectivity involved both a success orientation and a motive to avoid failure, defensive expectations primarily reflected failure avoidance, while self-handicapping actually bordered on failure acceptance.

18

In-depth interviews with students were conducted at Times 1 (n=24) and 2 (n=16) which expanded quantitative findings about the self-protective process from students personal perspectives. Time 1 interview data highlighted the idiosyncratic ways in which students self-protect through self-handicapping and defensive expectations, illuminated their consequences, and provided rich detail about the precise nature of the many factors associated with these strategies. Time 2 qualitative analyses shed light on upward and downward shifts in self-handicapping and defensive expectations over time. Time 2 qualitative data also provided insights into the concomitant shifts in the affective and motivational factors underpinning these strategies and the perceived academic and personal difficulties which students felt accounted for the identified shifts in self-handicapping and defensive expectations.

Both quantitative and qualitative data hold implications not only for current understanding and existing theory regarding self-handicapping and defensive pessimism, but also for educational practice and research dealing with these phenomena. On a more general level, data derived from the various studies provide new perspectives on defensive manoeuvring and the lengths to which students will go to protect their self-worth, the many reasons they pursue their studies in such a fashion, and the diversity (and complexity) of consequences that follow from such behaviour.

19

1. INTRODUCTION

For some students, the motive to protect self-worth is paramount and can sometimes be more important than the need to learn and perform successfully. According to the self-worth theory of motivation (Beery, 1975; Covington, 1984a, 1992, 1997; Covington & Beery, 1976), the need to protect self-worth arises primarily from a fear of failure and the implications this failure may have for ones private and public sense of ability and subsequent self-worth. Individuals who see failure as reflecting poorly on their ability are inclined to protect the self because ability is typically equated with self-worth (Covington, 1984a, 1992, 1997). Students can use a variety of strategies to deal with threats to their self-worth. Two such strategies - self-handicapping and defensive pessimism - are the focus of the present study.

Numerous motivational and affective factors are here proposed to underpin the need to strategically manoeuvre in a self-protective manner. Each of these factors is suggested to be a facet of the motive to protect the self and are argued to render the individuals self-worth particularly vulnerable in the event of failure. Moreover, because selfprotective strategies constitute students cognitive and behavioural reactions to achievement scenarios that pose the threat of failure, it follows that these strategies ultimately impact on important educational outcomes. Thus the characteristic way in which students react to potential failure and the implications this failure has for their ability, can influence their achievement-related behaviour and indeed, achievement itself.

It seems, then, that there exists a process by which students self-protect in the academic domain and the purpose of the study is to explore this process in detail. Essentially, this process is one in which a variety of motivational and affective factors render students vulnerable to the ability-related implications of failure. In response to this, students strategically manoeuvre so as to alter the meaning or implications of this anticipated failure. In turn, these strategies impact on important educational outcomes. Indeed, this process is very much akin to that outlined in a model proposed by Buss and Cantor (1989). According to them, individuals dispositions and characteristic orientations influence the strategies they use to negotiate demands in their 20

environment, and these strategies in turn influence their behaviour within this environment. In the context of the present study, these dispositions and characteristic orientations are juxtaposed by a set of motivational and affective factors proposed to be indicants of a self-worth motivation. The strategies refer to the means of self-worth protection. The behaviours refer to the behavioural and educational outcomes that follow from the implementation of these strategies (see Figure 1.1).

Motivational and affective predictors

Self-worth protection strategies

Educational outcomes

Figure 1.1 Basic structure of the proposed model to be examined The central foci of this process model are self-handicapping and defensive pessimism, two strategies students use to negotiate achievement situations and protect their selfworth. Self-handicapping involves the choice of an obstacle to successful performance that is typically used as an excuse in the event of poor performance (Berglas & Jones, 1978; Jones & Berglas, 1978). Self-handicappers are able to invoke the obstacle as the reason for the poor performance and are thereby able to deflect the cause of this poor performance away from their ability. Defensive pessimism involves setting unrealistically low expectations (defensive expectations) and thinking through a variety of possible outcomes (reflectivity) before upcoming performances. Defensive expectations and reflectivity are protective strategies in the sense that they steel the individual in the event of failure, thereby cushioning the blow to the self-worth (Norem & Cantor, 1986a, 1986b). Defensive expectations are also protective in that by setting unrealistic expectations prior to performance, the individual has generated lower and safer standards against which to be judged. To date, little work has systematically studied these strategies together and the present investigation seeks to do so with particular reference to self-worth motivation theory as a means of unifying substantive and empirical claims.

21

It is further proposed that giving rise to these strategies is a variety of motivational and affective constructs that render individuals self-worth particularly vulnerable in achievement scenarios. These include: (a) motivation orientation: ego- and task-orientation are proposed to reflect students emphasis on ability versus effort as providing feelings of success and their inclination to do one or the other influences their tendency to be motivated to protect the selfworth; (b) esteem-relevant competence valuation: related to ego-oriented students tendency to feel successful when demonstrating ability is the extent to which students selfesteem is dependent on the demonstration of competence; (c) avoidance-oriented performance: the extent to which students are motivated to perform out of a need to avoid poor performance and its negative consequences is suggested to influence students proclivity to protect the self; (d) public self-consciousness: given that there exists in many students a need to not only protect their private sense of ability but also their public ability image, public self-consciousness is argued to underlie a motive to protect the self; (e) attributions: students tendency to attribute the cause of success and failure to ability, effort, or external factors can also underpin a self-protection motive. For example, seeing failure as due to ability, and success as due to external factors, promotes a perceived lack of control over outcomes and this induces insecurity regarding upcoming performances, a concomitant fear of failure, and a threat to ones self-worth; (f) level and stability of self-concept: students level and stability of self-concept are proposed to be important in underpinning this self-worth motivation framework. For example, the extent to which students are confident in meeting upcoming challenges and the extent to which their self-concept is stable, conceivably hold implications for their tendency to protect the self; (g) uncertain control: related to this shaky self-concept is students uncertainty regarding their capacity to avoid failure or maintain success. Under conditions of such uncertainty, students may be particularly concerned with failure and the need to protect the self; and, (h) implicit theories of intelligence: individuals beliefs about the nature of intelligence could influence their inclination to self-protect. For example, an entity 22

view in which students see intelligence as due to ability and immutable may render them vulnerable to self-worth protection concerns. For students holding an entity belief, poor performance reflects adversely on their intelligence and ability, with consequent implications for their self-worth.

The study not only assesses the constructs that underpin the tendency to engage in self-handicapping or defensive pessimism but also examines the effects of selfhandicapping and defensive pessimism on educational outcomes. Whilst these strategies might be construed as adaptive in the sense that they can protect the individuals self-worth, it may be that they are not particularly adaptive in terms of important academic outcomes in the longer term. On the other hand, having laid the foundations for self-protection, the student may be in a stronger position to focus on actual performance rather than its self-worth implications. To test these competing contentions, the present investigation examines a variety of outcomes proposed to follow from self-handicapping and defensive pessimism. These include persistence, self-regulation, future plans for study, and academic grades.

Thus the complete model to be explored (see Figure 1.2) is one which incorporates a variety of motivational and affective predictors that are proposed to influence the tendency to self-handicap or be defensively pessimistic. In turn, these strategies are hypothesised to impact on a variety of educational outcomes. The model presented in Figure 1.2 is primarily based on a single wave of data. Importantly, however, the study extends this model across two time points such that this self-protection process is also assessed in a longitudinal fashion (see Figure 1.3). This involves surveying a sample of undergraduates in both their first and second years at university about the variety of constructs proposed in Figures 1.2 and 1.3.

23

Motivational and affective predictors

Self-protection strategies

Academic outcomes

Motivation orientation

Public selfconsciousness Esteem-rel competence valuation Self-concept level & stability Defensive pessimism

Persistence Selfhandicapping Selfregulation Grades (end of 1997)

Perceived control

Future academic plans Intelligence beliefs

Attributions

Figure 1.2 Constructs incorporated into proposed process model

24

TIME 1 (1997)

TIME 2 (1998)

Motivation orientation Public selfconsciousness Esteem-rel competence valuation Self-concept level & stability

Motivation orientation

Public selfconsciousness Esteem-rel competence valuation Self-concept level & stability Selfregulation Defensive pessimism Grades (end of 1997) Perceived control Future academic plans Defensive pessimism Selfregulation

Selfhandicapping

Persistence

Selfhandicapping

Persistence

Perceived control

Intelligence beliefs

Intelligence beliefs

Future academic plans

Attributions Attributions

Figure 1.3 Structure of the proposed longitudinal model explored in the study

25

In addition to these central issues, the quantitative component of the present study addresses a variety of supplementary issues that have not been given direct attention in previous research. These include closer consideration of (a) defensive expectations and reflectivity and the extent to which they operate as separate constructs, (b) the active and self-presented dimensions of self-handicapping and defensive expectations, (c) the domain specificity of the constructs involved in the central model (e.g., whether students use the same strategies in different university subjects), (d) a need achievement framework that clarifies the relationships amongst self-handicapping, defensive expectations, and reflectivity, (e) the psychometric properties of a number of recently proposed scales, and (f) the higher order nature of the affective and motivational predictors proposed to underpin (to varying degrees) a self-worth motivation.

Whilst the emphasis of the present study is on the process and longitudinal models, it is considered that the substantive issues involved lend themselves quite well to a qualitative approach. Accordingly, qualitative studies are conducted at Time 1 (first year) and Time 2 (second year). The Time 1 qualitative study involves interviewing students high or low in self-handicapping and defensive expectations, with a view to exploring major relationships derived in the Time 1 quantitative process model. The Time 2 study involves interviewing students who evince large shifts in selfhandicapping or defensive expectations across Times 1 and 2 and exploring their accounts of these identified shifts and possible reasons for them. These qualitative data are not intended to test the quantitative data per se. Rather, their purpose is to illuminate, from students personal perspectives, key findings derived in the quantitative models and shed light on issues which the Times 1 and 2 quantitative analyses could not assess.

The foreseen yields of the present study, therefore, are multifold: it provides timely confirmatory measurement data on self-handicapping and defensive pessimism, the factors underlying them, and the academic outcomes that follow from them; in contrast to the bulk of research which typically examines either predictors or consequences of defensive manoeuvring, the study specifies a more encompassing 26

process model in which both predictors and consequences of defensive manoeuvring are explored; the study extends the self-worth motivation research by explicitly locating both self-handicapping and defensive pessimism within it and also by proposing a factor structure comprising constructs hypothesised to reflect a self-worth motivation; the study is among the first to draw together self-handicapping and defensive pessimism under a common empirical process-like framework; in examining the relative salience of factors giving rise to self-handicapping and defensive pessimism, the study contributes to current understanding of the defensive manoeuvring related to achievement and motivation in the educational context; conceptual issues are assessed in the context of longitudinal quantitative and qualitative methods that provide data which to date have typically been generated using (primarily single wave) correlational or experimental methods; the qualitative components of the study contribute to the somewhat limited body of qualitative work that examines defensive manoeuvring in the educational context; a recently developed model of need achievement is used to clarify the relative positioning of self-handicapping, defensive expectations, and reflectivity, along failure-avoidance and success-orientation axes; the study provides data on a variety of supplementary issues including (a) the content specificity of key constructs, (b) the self-presented dimensions of selfhandicapping and defensive expectations, and (c) the clarification of the roles of defensive expectations and reflectivity; and findings arising from the study hold a number of implications for interventions designed to address students defensive manoeuvring in the academic setting.

27

2. SELF-PROTECTION STRATEGIES

The importance of protecting the self from the implications of failure has been emphasised over the past two decades (Beery, 1975; Covington, 1984a, 1992, 1997; Covington & Beery, 1976). Covington (1984a, 1992) argues that because failure is interpreted by many students as being indicative of low ability and because ability is equated with self-worth, failure holds implications for these students self-worth. Thus many students go to great lengths to avoid failure or to alter its meaning. Two strategies they can use to do this are self-handicapping and defensive pessimism. As is explicated more fully below, self-handicappers alter the meaning of failure by deflecting its cause away from their ability and on to factors such as a lack of effort which are less likely to threaten their self-esteem and consequent worth. Defensive pessimists alter the meaning of failure by cognitively and affectively steeling themselves for the event of failure as well as setting lower and safer standards against which to be judged. The present study, then, proposes that self-handicapping and defensive pessimism share a common motivational base and seeks to explore the extent to which this is empirically manifested both in terms of the factors that underpin them and in terms of their consequences.

2.1 Self-handicapping Self-handicapping involves the choice of an impediment or obstacle to successful performance that enables individuals to deflect the cause of failure away from their competence and on to the acquired impediment. Through such strategic manoeuvring, individuals are able to avoid disconfirmation of a desired self-conception (Rhodewalt & Davison, 1986). Examples of self-handicapping include the strategic reduction of effort, procrastination, or the choice of performance-debilitating circumstances. In the event of failure, the individual has a ready excuse for it: For example, the lack of effort is seen as the cause rather than the individuals lack of ability. Selfhandicapping, then, is a strategy that is directly relevant to the area of education and can be used to protect the student in the face of potential academic failure.

28

A small body of research has examined the self-presented dimensions of selfhandicapping (Arkin & Baumgardner, 1985; Leary & Shepperd, 1986). Self-presented self-handicapping is distinct from the above, more active, form of self-handicapping in that individuals exaggerate the obstacles in their path to success or report that these obstacles are present when in fact they are not. Thus self-presented self-handicapping typically takes the form of self-reports of performance-related hindrances such as the exaggeration of obstacles to success, excuses of ill-health, or test anxiety. Exaggerated test anxiety, for example, provides a ready excuse for potential failure (Greenberg, Pyszczynski, & Paisley, 1985). Thus active self-handicappers actually set a handicap in place that can inhibit performance, whereas self-presented self-handicapping is a claim that a handicap is present when it may not be or when it is claimed to be present to a greater degree than is actually the case.

Self-handicappers primarily rely on two attributional principles: Discounting (or protection) and augmentation (Kelley, 1972). The protection principle holds that failure under particular circumstances, such as the presence of an obstacle or impediment, is not proof of incompetence. In essence, by creating an impediment to performance, the self-handicapper minimizes the implications of failure, because failure is discounted - that is, it is attributed to the obstacle rather than to low ability (Baumeister & Scher, 1988, p.8). Indeed, Covington and Omelich (1979b) found that following failure, students were judged as less incompetent when they invested little effort but had an excuse available. The augmentation principle holds that success, in spite of the presence of performance-related hindrances, is proof of competence. Essentially, then, in addition to protecting themselves from the implications of failure, self-handicappers are able to capitalise on the favourable implications of successful performance (Baumeister & Scher, 1988). The self-handicapper, according to Berglas (1987), is wise to these attributional principles and exploits them in order to preserve a competence image (p. 310).

In many instances, then, self-handicapping is a conscious strategy engaged in by the individual for protective or augmenting purposes. It must be noted, however, that whilst this conscious self-handicapping represents the primary phenomenon studied in the research to date, it can also be a more unconscious reaction to impending failure. 29

The position adopted in the present study, however, is one which emphasises the strategic nature of defensive manoeuvring and thus is understood to be a more conscious, rather than unconscious, process. Moreover, as is discussed below, the weight of evidence suggests that self-handicapping primarily serves a protective function, and it is therefore conceptualised and operationalised in the present study as a defensive strategy. Whilst in some cases self-handicapping may have augmenting consequences, this is not the focus of the study. According to the most recent review of self-handicapping, by acquiring or claiming a handicap, and blocking the expression of ability, an individual can diminish lack of ability as the most plausible attribution for flawed performance or failure (Arkin & Oleson, 1998, p. 314).

2.1.1 Historical roots Higgins (1990) traces the fundamental underpinnings of self-handicapping back to Adlers psychology of self-esteem which contends that individuals motivations are derived from the need to compensate for subjective feelings of inferiority and weakness. Adler was among the first to identify individuals symptoms as unconscious creations of the ego aimed at safeguarding their esteem losses following failure. Higgins identifies two publications (Goffman, 1959; Heider, 1958) as marking the next important stage in conceptualising preceding self-handicapping theory. Goffman provided an important perspective on impression management, a process in which individuals attempt to manipulate their overt behaviours so as to regulate the impression others have of them. Heider proposed the naive theory of causal attributions which provided a framework by which to think about the conditions under which observers see characteristics of individuals and under which conditions these individuals are seen to be responsible for their behaviour. Indeed, the theory of selfhandicapping advanced by Jones and Berglas (1978; Berglas & Jones, 1978) essentially constituted a synthesis of Adlers, Goffmans, and Heiders theoretical positions. As noted above, Kelley (1972) further contributed to self-handicapping theory through his outline of the discounting and augmenting principles of attribution.

According to Higgins (1990), following Jones and Berglas (1978; Berglas & Jones, 1978), the next important development in the self-handicapping research involved Snyder and Smith (1982) who extended self-handicapping to encompass 30

characteristics within the individual. Until 1982, self-handicapping was characterised as behaviour incorporating external agents such as ingestion of drugs or alcohol. The Snyder and Smith extension of self-handicapping gave rise to the conception of selfhandicapping also incorporating self-reports of ill-health or anxiety. Indeed, the present study extends this notion somewhat by exploring self-presented selfhandicapping which refers to reports of impediments to success in the students life that are either non-existent or exaggerated.

2.1.2 The variety of studies examining self-handicapping To date, self-handicapping has been assessed in a variety of ways. The majority of research has examined active self-handicapping in an experimental setting. In such studies, participants are presented with situations in which they anticipate failure and are typically given an opportunity to withdraw effort, ingest a performancedebilitating drug, or select performance-inhibiting conditions under which to be evaluated. These impediments to success are then able to be used as alibis for anticipated poor performance. A smaller body of research examines self-presented self-handicapping. In these studies, participants who are led to anticipate poor performance are given the opportunity to exaggerate performance-related hindrances that may be used as excuses for poor performance. A third body of research has assessed trait self-handicapping in which self-handicapping scales are administered to respondents. Data derived from the scales are typically studied using correlational techniques which assess the relationship between self-handicapping scores and other trait-like measures. Whilst the empirical focus of the present study is on the latter, trait-like self-handicapping, the review of literature encompasses the variety of ways in which self-handicapping occurs and is assessed. The following represents a summary of studies that illustrate the diversity of research in the area. Further detail is presented in Chapter 3.

31

2.1.2.1 Active self-handicapping Individuals can actively self-handicap in a variety of ways. Studies assessing such self-handicapping usually lead participants to anticipate failure or induce some form of insecurity regarding their self-worth and then provide them with opportunities to choose an impediment to successful performance. Active self-handicapping has been found under a variety of conditions ranging from drug ingestion (Berglas & Jones, 1978; Gibbons & Gaeddert, 1984; Higgins & Harris, 1988; Kolditz & Arkin, 1982; Tucker, Vuchinich, & Sobell, 1981) to the selection of performance-hindering circumstances (Frankel & Snyder, 1978; Greenberg, 1985; Rhodewalt & Davison, 1986; Shepperd & Arkin, 1989a; Snyder, Smoller, Strenta, & Frankel, 1981).

Withdrawal of effort as a means of active self-handicapping is perhaps more relevant to the educational context. Research investigating the strategic withdrawal of effort has focused on conditions in which participants self-esteem is under threat or in which they are unsure about their ability to successfully meet upcoming challenges. For example, self-handicappers have been found to withhold effort before important competitions posing a threat to self-esteem but do not do so before unimportant meets (Rhodewalt, Saltzman, & Wittmer, 1984). Similar to the above studies of performance-inhibiting conditions, participants who anticipate failure invest less effort than those who anticipate success (Pyszczynski & Greenberg, 1983; see also Baumeister, Hamilton, & Tice, 1985).

2.1.2.2 Claimed and self-presented self-handicapping In addressing the issue of self-presented self-handicapping, DeGree and Snyder (1985) allude to Adlers contention that trauma does not cause success or failure. Rather, individuals make out of trauma what suits their purpose. Specifically, individuals can use past experiences in order to avoid present responsibility and to maintain their competence-image. Consistent with this, DeGree and Snyder hypothesised that (a) individuals would accentuate the adversity of events in their lives when this was seen as a viable excuse and (b) this claimed self-handicapping should arise in situations of evaluative threat. Consistent with hypotheses, they found that study participants emphasised traumatic life events that had occurred in the past when an uncertain evaluation was expected and when they believed their traumatic past might represent a 32

viable excuse for possible failure. Similarly, Smith, Snyder, and Perkins (1983) argued that symptomatic complaints and behaviors may serve as strategies for reducing the negative implications of poor performance, thereby theoretically maintaining the individuals sense of self-esteem (p. 787). They found that hypochondriacal individuals reported more physical ill-health and current physical symptoms to explain poor performance in an evaluative setting than individuals who believed poor health was not a viable excuse and those in a non-evaluative setting.

In a study designed specifically to determine the preferred form (active or claimed/ self-presented) of self-handicapping, Hirt, Deppe, and Gordon (1991; see also Smith, Snyder, & Handelsman, 1982; Snyder, Smith, Augelli, & Ingram, 1985) provided subjects with the opportunity to actively self-handicap (withdraw practice effort) or claim/self-present a self-handicap (report high stress) before a test. They found that when participants were of the belief that one particular handicap was viable then that handicap was the one adopted. More importantly, they found that when both stress and withdrawal of practice were seen as viable handicaps, participants preferred the claimed/self-presented stress self-handicap. They hypothesised that participants preferred to choose a claimed/self-presented self-handicap because while this provided them with a ready excuse for poor performance, it did not compromise their chances of success. It seems, then, that psychological symptoms can serve to control the causal attribution of potentially negative performance feedback about self-relevant dimensions and thereby prevent loss of self-esteem (Smith et al., 1982, p. 315).

2.1.2.3 Trait measures of self-handicapping Whilst most of the research into self-handicapping has operationalised it behaviourally, a smaller body of research has assessed such defensive manoeuvring through psychometric scales. These measures are intended to quantify the dispositional tendency to actively self-handicap. A variety of self-handicapping scales have been devised to assess this dispositional tendency. The original Self Handicapping Scale (SHS - Jones & Rhodewalt, 1982) was a 20-item scale found to be predictive of self-handicapping behaviour (Rhodewalt et al., 1984). Strube (1986), however, identified a number of problems associated with the original scale. For example, the reliability was not particularly high (=.62). The original scale also 33

yielded six factors and Strube found a major division between the first and subsequent factors and contended that a one-factor scale might be the best measure of selfhandicapping. Accordingly, Strube (1986) took the ten items that loaded most highly on this factor, formed the shortened SHS, and demonstrated acceptable predictive and concurrent validity.

One point to note with the SHS, however, is that the protective (or augmenting) dimensions of self-handicapping were not explicitly invoked by component items. A later scale devised by Midgley, Arunkumar, and Urdan (1996) explicitly incorporated into items the self-protective dimensions of self-handicapping and found this scale to be a reliable measure of self-handicapping. Thus, whilst the earlier scales assessed behaviour that could impede successful performance, they could not ascertain whether such impediments were specifically intended for self-protective purposes. For example, an individual may withdraw effort before an exam because of other important commitments and not because of a need to have an excuse available in the event of poor performance. The Midgley et al. scale, on the other hand, unambiguously assesses self-handicapping behaviour in the sense that each item explicitly asks about the self-protective intent of individuals behaviour and it is this scale which is adapted for use in the present study.

Using this Midgley et al. (1996) scale as a guide for further item construction, operationalisation of self-handicapping in the present study incorporates all three of the aforementioned dimensions of self-handicapping (active, self-presented, and trait). Essentially, this involves collecting data on active and self-presented selfhandicapping through students self-reports such that students respond to items addressing (a) the ways they actively impede their success for protective purposes (in the case of active self-handicapping) and (b) their tendency to exaggerate these impediments (in the case of self-presented self-handicapping). 2.1.3 Consequences of self-handicapping1
Subsequent to the present studys data collection, two studies were published dealing with the consequences of self-handicapping. Zuckerman, Kieffer and Knee (1998) found that high selfhandicappers were more inclined than low self-handicappers to use coping strategies reflecting a negative focus and withdrawal, were lower in academic achievement, and over the longer term evinced
1

34

Not a great many studies have examined the consequences of self-handicapping behaviour. Of those conducted, however, the findings have been mixed. While some have demonstrated positive consequences, others show that self-handicapping can be quite counterproductive. To note, however, is that the outcome constructs under focus in the relevant studies are not entirely comparable: Some have examined performance, some have focused on anxiety and perceptions of efficacy, and others have assessed intrinsic motivation.

In relation to anxiety and perceptions of efficacy, Leary (1986) found that study participants were less aroused (less anxious) in a social interaction when they were told that a noise would interfere with their ability to interact than when they were told that it would not (see also Harris & Snyder, 1986). Leary also found that socially anxious people in the loud-noise condition described themselves positively (see also Arkin & Baumgardner, 1985, in Snyder, 1990). Rhodewalt and Hill (1995) too, found that participants were higher in self-esteem following self-handicapping (see also Isleib, Vuchinich, & Tucker, 1988; Rhodewalt, Morf, Hazlett, & Fairfield, 1991), while Mayerson and Rhodewalt (1988) found that self-handicappers felt that their ability image was intact following failure. Similarly, Drexler, Ahrens, and Haaga (1995) found that participants who chose to self-handicap on a spatial task were less likely to experience a decrease in positive affect than those who chose not to selfhandicap. In these studies, then, self-handicapping served an adaptive purpose in that it reduced anxiety and evoked more positive self-appraisals.

It has also been shown that performance improves for those in a condition in which they believe music will inhibit their performance (Snyder et al., 1981). According to Snyder et al. (1981), the distraction condition improves performance because low effort is no longer needed as an excuse. The subject is free to act on achievement motivation uninhibited by fear of failure clearly attributable to low ability (p. 29). In terms of self-reports of handicaps, Smith et al. (1982) found that test-anxious study participants who reported elevated anxiety as a handicap did not evince performance decrements relative to other participants (see also Greenberg et al., 1985). In a similar
poorer adjustment. In another study, Eronen, Nurmi, and Salmela-Aro (1998) found that selfhandicapping was associated with dissatisfaction in ones academic life and low well-being.

35

vein, Thompson (1993) found that when given a face saving opportunity, self-worth protective individuals actually improved performance (see also Sanna & Mark, 1995).

Self-handicapping may also enhance the engagement and enjoyment of a task. In a study by Deppe and Harackiewicz (1996) to explore this issue, participants high or low in self-handicapping engaged in a task in which they were given the opportunity to practice or not practice (self-handicap). High self-handicappers who did not practice became more involved in the task and consequently enjoyed the task more than high self-handicappers who did practice. They found in a second experiment that self-handicapping enabled individuals to become more involved in the activity such that they tended not to focus on performance concerns. They concluded that selfhandicaps may provide the breathing room that some individuals need to become involved in a task and to experience that task as pleasant (p. 874).

Others, however, have proposed that self-handicapping is not so adaptive. Covington (1984a , 1992) argues that in the long run, self-defeating tactics such as selfhandicapping are not successful: Despite the fact that these tactics may temporarily reduce the immediate anxiety and distress of failure, in the long run they are selfdefeating because they tend to set up the failures that students are trying to avoid (1984a, p. 91). For example, high self-handicappers have been found to withdraw effort, perform more poorly, and report more distracting cognitions than low selfhandicappers in high difficulty tasks (Rhodewalt & Fairfield, 1991; see also Poehlmann & Oleson, 1995, in Arkin & Oleson, 1998). Rhodewalt and Davison (1986) found that study participants who received non-contingent success feedback and chose to self-handicap performed more poorly on an ability test than those who elected not to self-handicap (but participants receiving non-contingent failure feedback who self-handicapped performed better), while Rhodewalt et al. (1984; see also Rhodewalt, 1984, in Rhodewalt, 1990) found self-handicapping to be associated with underachievement. Midgley et al. (1996) found that self-handicapping mediated the relationship between negative attitudes and grade point average (GPA) such that negative attitudes about education positively predicted self-handicapping which

36

negatively predicted GPA (see also Midgley & Urdan, 1995)2. Similarly, Garcia et al. (1995) found that relative to students low in self-handicapping, high self-handicappers were significantly lower in intrinsic goal orientation, self-regulation, and grades.

The findings, then, are mixed. Whilst self-handicapping may be detrimental to academic outcomes, other research has demonstrated that there are benefits that can be derived from using the strategy. As outlined earlier, however, studies of the consequences of self-handicapping have employed outcome variables that are not parallel and so it is difficult to make generalised predictions. It can be tentatively proposed, however, that because the present study operationalises self-handicapping in a way parallel to work by Midgley and colleagues, self-handicapping may yield similar effects. Specifically, self-handicapping is predicted to be negatively associated with academic outcomes. Because of the diversity of outcome constructs employed in studies to date and because the present study is aimed at clarifying the consequences of self-handicapping, a variety of outcome constructs are incorporated which reflect the diversity of outcome constructs examined to date. Accordingly, future academic plans, persistence, and self-regulation (reported at Time 1), and academic grades (reported at Time 2) are incorporated in the present study as outcome constructs. The first component of the proposed process model incorporating these constructs is shown in Figure 2.1.

In another study published subsequent to the present studys data collection, Urdan, Midgley, and Anderman (1998) found self-handicapping to be negatively associated with students GPA.

37

Self-protection strategies

Academic outcomes

Persistence

Active selfhandicapping Selfregulation Self-pres selfhandicapping Future academic plans Grades (end of 1997)

Figure 2.1 The proposed process model incorporating the association between selfhandicapping and academic outcomes Note. Broken paths estimated at Time 2 2.1.4 Protective or augmenting? It is unclear as to whether self-handicapping is prompted more by a need to protect individuals self-esteem (the protection hypothesis) or to augment individuals in the eyes of themselves and/or others (the augmentation hypothesis). Indeed, much of the research does not clearly distinguish between the two (Arkin & Oleson, 1998). Shepperd and Arkin (1989a) hypothesised that self-handicapping is probably motivated more by a need to protect the self because its correlate, public selfconsciousness (see Chapter 3), has typically been found to be aimed at avoiding disapproval rather than garnering approval (see Carver & Humphries, 1981; Fenigstein, 1979). Findings by Mayerson and Rhodewalt (1988) also support the protection hypothesis (see also Rhodewalt, 1990). They found that after failure feedback, self-handicappers reported that the handicap not only inhibited their performance but their low performance did not reflect poorly on their ability. Low self-handicappers, on the other hand, reported that low performance was due to low ability. In contrast, they also found that after participants received success feedback there were no differences in attributions between self-handicappers and non self38

handicappers, representing evidence contrary to the augmentation hypothesis. Also consistent with the protection hypothesis, Isleib et al. (1988) found that subjects who drank alcohol before a test and received failure feedback reported that the alcohol interfered with their performance and that failure was not related to their ability. In another study, Rhodewalt et al. (1991) found that individuals high in selfhandicapping discounted ability after failure. In a second study, they showed that after success, the presence of a self-handicap led to no more positive self-esteem than success without a self-handicap. Thus, self-handicapping served no enhancing or augmenting purpose but it did protect self-esteem in the event of failure. They concluded that there existed a greater readiness among subjects to adopt a selfprotective rather than self-enhancing posture (p. 130).

Other commentators argue that the extent to which self-handicapping serves a protective or enhancing purpose depends on individuals level of self-esteem. For example, Baumeister, Tice, and Hutton (1989) found that those high in self-esteem were motivated to enhance their public and self-image whilst low self-esteem individuals were more concerned with protecting their image. Similarly, Roth, Snyder, and Pace (1986) found that high self-esteem individuals were more likely to use selfenhancing strategies and more likely to present themselves in a more positive manner that actually bordered on the unrealistic. In another study assessing self-esteem and self-handicapping, Tice (1991) found that when a high score on an important task was characterised as meaningful, high self-esteem participants self-handicapped

(presumably for self-enhancement purposes), whereas when failure on a task was considered meaningful, low self-esteem participants self-handicapped (for selfprotective purposes). In this case, then, the joint effects of self-esteem, task importance, and the implications of the outcome influenced the tendency to handicap in an enhancing or protective way. Findings were replicated in a second study. Tice concluded that low self-esteem people self-handicap primarily to protect their image, whereas high self-esteem people self-handicap primarily to augment their image (p. 724).

Other research has examined dispositional correlates of self-handicapping that can be related to protective and acquisitive tendencies. One such dispositional measure, the 39

Modified Self-Monitoring Scale (Lennox, 1988), assesses acquisitive self-presentation and protective self-presentation. Ferrari (1992) examined the correlation between the short form of the SHS (Strube, 1986) and these two dimensions of self-monitoring. He found that the SHS was correlated .29 with acquisitive self-presentation and .30 with protective self-presentation, indicating that self-handicapping can present both protective and enhancing opportunities.

The focus of the present study is upon the protective dimension of self-handicapping. In fact, this is consistent with the Adlerian perspective on strategic manoeuvring which holds that individuals are primarily motivated to manoeuvre so as to protect their feelings of inferiority. Indeed, Jones and Berglass (1978; see also Snyder, 1990, for a review) original contentions were that self-handicappers sought to protect themselves from the horrible . . . suspected truth (p. 203) that they are not competent and hence unworthy of love. This protection perspective of selfhandicapping is also consistent with more recent research. Midgley and colleagues, for example, have explicitly invoked in their self-handicapping items the need for selfhandicapping as an excuse in the event of poor performance (Midgley et al., 1996; Midgley & Urdan, 1995)

2.1.5 Summary of self-handicapping Self-handicapping can take active and self-presented forms. Moreover, it has been proposed that there exists a dispositional tendency to engage in self-handicapping and psychometric scales have been devised to quantify this. The present study unifies these three dimensions by not only assessing the active form of self-handicapping using a self-report scale but devises a scale to assess the self-presented form of selfhandicapping. In terms of the consequences of self-handicapping, the evidence is mixed, and it is another purpose of the present investigation to clarify this issue. In general, the weight of evidence suggests that self-handicapping is more a protective strategy than an augmenting one - and this is the stance adopted in the present study. As yet, the factors that give rise to self-handicapping have not been discussed directly and this is the focus of Chapter 3. Before addressing these factors, however, the literature in relation to the second strategy in the proposed model - defensive pessimism - is reviewed. 40

2.2 Defensive pessimism In contrast to the large volume of self-handicapping literature, the literature in relation to defensive pessimism, a more recent formulation, is not so extensive. Defensive pessimists3 set unrealistically low expectations (defensive expectations) prior to an event in which some form of evaluation takes place and think through a variety of possible outcomes before performance (reflectivity; Norem & Cantor, 1986a, 1986b; Norem & Illingworth, 1993). In adopting a defensively pessimistic approach, individuals are able to cognitively work through potential failure and thereby steel themselves for this outcome (Norem & Cantor, 1986a, 1986b; Norem & Illingworth, 1993). It is this thinking-through process that enables defensive pessimists to feel less anxious, more in control, and better equipped to invest effort into the task before them. With the increase in effort that accompanies the strategy, performance is often subsequently unimpaired (Garcia et al., 1995; Norem & Cantor, 1986b). Thus, defensively pessimistic students perform in ways that run counter to expectancy-based research: These low expectations do not become self-fulfilling prophecies, nor does the anxiety, although real, lead directly to performance deficits. In fact this strategy may be thought of as a method by which individuals are able to cope with their anxiety, in effect, change it from a debilitating to a motivating force (Norem & Cantor, 1986b, p. 1209).

A defensively pessimistic strategy can also cushion the individual against debilitating anxiety prior to stress-provoking tasks and motivate continued persistence in the face of that stress (Cantor & Norem, 1989, p. 93). In fact, setting lower expectations can also serve to establish performance standards that are less difficult to achieve (Showers & Ruben, 1990; see also Atkinson, 1957; Birney, Burdick, & Teevan, 1969), and may even lower the threshold for satisfactory performance (Baumgardner & Brownlee, 1987). As Covington notes, by dropping the lower bounds of what one will accept - students can continue to evade feelings of failure (1992, p. 90). Covington also implies that there is a self-presented and public
While defensive pessimism comprises defensive expectations and reflectivity, the bulk of research conducted to date deals primarily with the defensive expectations component and refers to these defensive expectations as defensive pessimism. Thus, in the present discussion of previous research, the term defensive pessimism is used and often invariably refers to defensive expectations.
3

41

dimension to strategic reductions in expectations: The student who publicly announces before each examination that he or she will be satisfied with just a passing grade is taking crafty advantage of this strategy (1992, p. 90). This self-presented dimension of defensive expectations is further explored in the present study.

Norem and Cantor (1990b) differentiate defensive pessimism from general pessimism and depression because defensive pessimists set low expectations whilst acknowledging past successes. This contrasts with realistic pessimists who have not done well in the past and depressive pessimists who tend to discount past successes. Showers and Ruben (1990) argue that while defensive pessimists may demonstrate a pattern of low expectations, moderate levels of anxiety, and extensive effort and preparation before a stressful event, maladaptive pessimistic strategies may be characterized by low expectations, uncontrolled anxiety, and a tendency to withdraw effort and disengage from the situation at hand (p. 387).

2.2.1 The strategic nature of defensive pessimism According to Norem and Cantor (1990b), individuals use of cognitive strategies influence the way they appraise, plan, retrospect, and invest effort: Strategies such as self-handicapping, defensive pessimism, and depressive self-focus, constitute characteristic ways that individuals try to reach desired end states and avoid undesirable outcomes in particular task settings (p. 192). Specifically, defensive pessimists seem to use expectations to gain a sense of control in ways that influence motivation and performance and the conjunction of expectations and efficacy/control as cognitive bridges between motivation and action is particularly helpful in illuminating the conditions under which negative expectations do not lead to negative outcomes (Norem & Cantor, 1990b, p. 192).

Cantor and Norem (1989; see also Cantor, Norem, Niedenthal, Langston, & Brower, 1987) argue that defensive pessimism and optimism are strategic in that pessimists and optimists display distinct patterns in appraising and reacting to achievement scenarios. Typically, defensive pessimists are more anxious, work through worst-case scenarios, do not deny control over failure and success after the task, and increase effort. The strategic nature of defensive pessimism is emphasised in the fact that 42

defensive pessimists manoeuvre protectively before a performance situation and as a result do not feel the need to engage in post hoc defensive posturing such as denying control after the fact, as is the tendency of optimists. According to Norem & Cantor (1990a), there is no need to resort to self-protecting attributions because the initial strategic construction of the situation has prepared them for the possibility of failure (p. 49).

2.2.1.1 Expectancy regulation A crucial component of defensive pessimism is the part played by expectations for upcoming performance situations. Expectations are a means by which individuals are able to negotiate evaluative scenarios and also regulate the impression that they make on others. In fact, it has been found that individuals actually try to change expectations of them held by others so as to protect themselves in evaluative situations (Baumgardner & Brownlee, 1987). Two ways to do this are to (a) communicate ones low expectations for an upcoming performance or (b) disconfirm others expectations by deliberately failing in the face of high expectations. Baumgardner and Brownlee found that socially anxious people deliberately fail when there are high expectations of them. They also found that individuals high in anxiety, when led to believe that initial performance would influence others high expectations of them, performed significantly poorer than other groups. These individuals, then, regulated others expectations of them by failing strategically and thereby created lower and safer standards against which to be judged. According to Baumgardner and Brownlee (1987), success may carry certain costs for the future self-presentational goals of the individual . . . the individual who lacks confidence in his or her abilities may be particularly threatened by the high standards his or her successes might impose (p. 525). For example, socially anxious people who are concerned about meeting external standards may be more comfortable with low standards (Schlenker & Leary, 1982) and so by disconfirming the high expectations of them, the socially anxious person may think subsequent success more likely since it is now determined by performance against a lower standard.

Thus, it is not uncommon for individuals to use expectations strategically to negotiate events that are affectively risky (Norem & Cantor, 1986b). Expectations can serve to 43

protect the individual from evaluative threat and the implications of failure both privately and publicly. Accordingly, it is argued here that the defensive expectations component of defensive pessimism is a means by which individuals deal with performance that is threatening to their self-esteem. Through these defensive expectations, failure against a lower standard is less likely, or, if failure does occur, individuals have steeled themselves for it, particularly in relation to the implications it has for their self-worth.

2.2.2 The distinction between expectations and reflectivity As outlined earlier, the two components that comprise defensive pessimism are defensive expectations and reflectivity. One purpose of the present study is to assess the dispositional tendency for students to engage in defensive expectations and reflectivity, not only as they operate individually, but also as they operate jointly in the form of defensive pessimism.

The thinking-through or reflectivity process has been found to be central to defensive pessimists success on tasks. In fact, Norem and Illingworth (1993) found that when study participants were asked to list their thoughts while engaging in tasks, defensive pessimists reported feeling better and less anxious than when they were distracted from reflectivity. Optimists, on the other hand, felt better and less anxious when they were in a distraction condition. It is proposed here, however, that there may be an important difference between simply thinking about various outcomes and actually expecting those outcomes. It is further proposed that reflectivity may be an adaptive component of defensive pessimism that may enhance individuals control, whereas engaging in defensive expectations may be a not so adaptive component of defensive pessimism.

It has been shown that negative thinking is not a problem until it becomes an expectation. In fact, Goodhart (1986) went so far as to argue that negative thinking is more adaptive than optimism provided that explicit negative expectations do not follow from the negative thinking. Goodhart examined the effect of negative and positive thinking on performance and also the effect of expectations/predictions in each of the positive and negative thinking scenarios. Results suggested that negative 44

thinkers performed better than positive thinkers: They seemed to try harder to avoid realising negative possibilities. Importantly, this finding applied only to the participants who did not predict or form expectations about their performance.

In another study, Sherman, Skov, Hervitz, and Stock (1981) asked all participants before an upcoming anagram task to explain failure or success hypothetically. Half the participants were also asked to make explicit expectations for the upcoming event. Those who explained failure and then expected it performed more poorly than those who explained success and expected it. When expectations were not made, those who explained failure performed better than those who explained success. It seems, then, that raising the possibility of failure without forming concrete failure expectancies motivates better performance (Sherman et al., 1981, p. 142). In explaining this, Sherman et al. argued that when individuals explain hypothetical failure, they generate cognitions about their abilities and motivation consistent with failure. If these individuals are then asked to make explicit expectancies, these cognitions are harnessed which increases the likelihood of failure.

The positive effects of bringing negative possibilities to mind (rather than forming negative expectations) were also demonstrated by Showers (1992). She found that negative thinking for defensive pessimists resulted in more social engagement. Moreover, their partners in the social situation rated them as being more sociable and felt that they invested more effort into the task than the positive thinking pessimists. According to Showers, this suggests that some benefits of the pessimistic perspective may accrue simply from bringing negative possibilities to mind, rather than from believing that they are likely to occur (p. 478). Indeed, simply bringing possibilities to mind seems more closely aligned with reflectivity than defensive expectations.

In sum, then, it is proposed that the difference between simply thinking about an outcome and actually expecting an outcome is one of the key differences between reflectivity and defensive expectations. It is further proposed that the two components of defensive pessimism may represent different strategies and it may even be the case that different outcomes follow from them. For example, based on the evidence presented above it might be predicted that reflectivity has positive effects on academic 45

outcomes, whilst defensive expectations have a negative influence. Thus one component of the present study assesses their effects as separate constructs.

However, proposing defensive expectations and reflectivity as two independent constructs runs contrary to recent theorising about defensive pessimism. But how they actually work together in representing defensive pessimism has not been made clear in the research to date. A small body of work has focused on two possible processes: (a) defensive expectations and reflectivity interact such that students high on both dimensions can be classified as defensive pessimists (Norem & Illingworth, 1993) and (b) defensive expectations relax students sufficiently that they can engage in reflectivity which then impacts positively on performance (Norem, personal communication, March 26, 1997, March 7, 1998), which in the context of the present study, implies that defensive expectations predict reflectivity. Because of this lack of clarity, the present study primarily deals with defensive expectations and reflectivity as main effects, however, the feasibility of possibilities (a) and (b) is also explored in the due course of analyses.

Importantly, it is the defensive expectations component of defensive pessimism which constituted the original emphasis (Norem & Cantor, 1986a, 1986b) of the theory and it was only in later conceptualisations (e.g., Norem & Illingworth, 1993) that reflectivity was formally incorporated into the construct. Whilst reflectivity underlies defensive pessimism (the specific extent to which this is the case is one aim of the empirical analysis), it is argued that defensive expectations is the more central construct and the one which is used more strategically in a self-protective manner. Accordingly, defensive expectations as a construct is given relatively greater emphasis in the subsequent discussion and empirical treatment of defensive pessimism.

Notwithstanding this, defensive expectations and reflectivity as two main effects are correlated in the central model, and, as discussed below, this is statistically equivalent to a model recently proposed by Norem (personal communication, March 26, 1997, March 7, 1998) in which defensive expectations predict reflectivity.

46

2.2.3 The consequences of defensive pessimism in the short term4 In assessing the effects of defensive pessimism, a good deal of the research has contrasted defensively pessimistic students with optimists. For example, Norem and Cantor (1986a) found that defensive pessimists set significantly lower expectations than optimists but there were no significant differences between the two groups in performance on tasks. In a second study, Norem and Cantor (1986b) found that defensive pessimists performed successfully despite their lower expectations, and results also suggested that when the defensive pessimistic strategy was interfered with in the form of encouragement, there was a significant decline in performance.

Other research has contrasted three groups: defensive pessimists, optimists, and aschematics (those neither defensively pessimistic nor optimistic) who do not differ substantially in previous performance. Norem and Cantor (1990b) found that while defensive pessimists felt more anxious, were more negative about their performance, were more absorbed in the task, and reported feeling less in control than optimists, they performed just as well as optimists. Importantly, then, the anxiety and the lack of control did not impair the defensive pessimists performance. They further found that the more reflective the defensive pessimists were, the better their performance (see also Norem & Illingworth, 1993), while the more reflective the optimists, the worse their performance. These results were confined to the academic domain and did not generalise to the social domain.

Cantor and Norem (1989) reported that compared to optimists, defensive pessimists felt less in control, were more stressed, experienced less enjoyment, and perceived less progress when they were studying. They also examined social interactions of defensive pessimists and found that whilst they were as socially effective as the optimists, they reported feeling more fatigued after social engagement. Moreover, the defensive pessimists were found to be more variable in emotional reactions across situations than optimists, who appeared more emotionally well-balanced.

Subsequent to data collection, one study of defensive pessimism was published (Eronen et al., 1998). Results indicated that defensive pessimists were significantly higher in grades than optimists at Time 1 of the study but these differences were not statistically significant at Time 2. Importantly, however, Eronen et al. did not assess defensive pessimism directly. Rather, they inferred defensive pessimism from a cluster analysis of numerous constructs.

47

Additionally, Cantor et al. (1987) found that relative to optimists, defensive pessimists perceived good grades as more difficult to achieve, and while they were higher in anxiety, this anxiety was positively associated with performance.

The weight of evidence suggests that the negative expectations held by defensive pessimists do not impair performance. Rather, defensive pessimists performance in some cases is actually enhanced. Notwithstanding the generally positive (or unimpaired) influence of defensive pessimism on performance there seem to be other costs involved in using the strategy. Importantly, the longer term effects of defensive pessimism also attest to these costs.

2.2.4 Longer term effects of defensive pessimism Whilst most research in the area of defensive pessimism has assessed the short-term impact of this strategy, a smaller body of research has investigated its longer term effects. Norem and Cantor (1990b) report findings of a three-year-follow-up of defensive pessimists and found that while the grades achieved by defensive pessimists were in the short-term comparable to those of optimists, three years later, defensive pessimists GPAs were lower than the other two groups (optimists and aschematics). In the third year, defensive pessimists also experienced more global life stress, more psychological symptoms, and less satisfaction with their lives.

According to Norem and Cantor (1990b), it is difficult to know precisely what led to these longer term effects. Notably, whilst the defensive pessimists in the first year held unrealistically low expectations, after three years they actually held higher expectations. In relation to this, it will be recalled that interference with the defensive pessimistic strategy in the form of encouragement/optimism has been found to adversely affect defensive pessimists performance (Norem & Cantor, 1986b) and thus it may have been the change in expectation (disruption of the strategy) that led to the decline in performance three years later.

Another reason Norem and Cantor (1990b) provide to explain the change in academic fortunes of the defensive pessimists is that the emotional ups and downs of academic pessimism . . . may take a heavy toll on well-being over time . . . the continual 48

struggle to gain control and manage anxiety may begin to water down the rewards of success and the intrinsic pleasure of the activities themselves (p. 202). According to Cantor and Norem (1989), defensive pessimism may be a liability over time and the rocky emotional terrain (p. 105) associated with it leads to a variety of academic and personal costs over time. Cantor and Norem also make the point that defensive pessimists were globally pessimistic when they arrived at college and that through defensive pessimism they gained control in the short term. Over the course of time, however, this strategy weakened and their global pessimism became more salient than the defensive expectations and reflectivity. Yet Cantor and Norem are inclined to dismiss this interpretation in the light of evidence that the individuals high in defensive pessimism in the academic domain were not pessimists in other domains, suggesting that they were not globally pessimistic.

The negative effects of defensive pessimism may also be partly a consequence of students inclination to self-verify. According to Swann (1987; see also Swann & Read, 1981), individuals prefer self-confirmatory feedback and think and behave in ways that promote the survival of their self-conception (p. 1039). For the most part this is because (a) they see such information as diagnostic (Snyder & Swann, 1978), (b) a stable self-conception helps them organise experience, and (c) they resist changes in their conception of self that would have them venture into new territory or engage in tasks that threaten failure. Individuals can self-verify by behaving in ways consistent with their self-view or behaving in ways that control how others see them (Swann, 1987). Importantly, individuals with negative self-views are also inclined to prefer and seek out negative feedback so as to confirm their negative self-view (Swann, Pellham, & Krull, 1989). In terms of defensive pessimism, it may be that in the longer term, defensive pessimists behave in ways that verify their pessimistic selfview. According to Swann (1987), although there are surely painful consequences associated with verifying negative self-conceptions . . . failing to verify them may have even more painful consequences (p. 1039; see also Swann, Stein-Seroussi, & Giesler, 1992).

There is another possibility which has not been explored. In the longer term, the defensive expectations component of defensive pessimism may become more salient 49

than the reflectivity component. It has been shown that negative expectations about an event impact adversely on outcomes whereas simply thinking about negative outcomes can actually be adaptive (Goodhart, 1986; Sherman et al., 1981). It may be that in the longer term defensive pessimists succumb to their negative expectations which become more salient than their reflectivity. As Covington (1992) notes, students do only as much as they expect of themselves (p. 90) and so their negative expectations in time become inimical to adaptive functioning.

The present study, then, seeks to explore further the association between defensive pessimism and academic behaviours. Consistent with the literature separating thinking (reflectivity) and expectations (defensive expectations), the present study also deals with them as separate constructs. Thus to parallel the relationships proposed in relation to self-handicapping and academic outcomes, the link between both defensive expectations and reflectivity and academic outcomes such as future academic intentions, persistence, self-regulation (reported at Time 1) and academic grades (reported at Time 2) are incorporated into the second component of the proposed model (see Figure 2.2).

It will be noted in Figure 2.2 that in a similar manner to self-handicapping, defensive expectations (the component of defensive pessimism given emphasis in the present study) is separated into active and self-presented dimensions. To date this has not received empirical attention and it is intended to parallel the distinction made in the self-handicapping literature. Active defensive expectations refer to unrealistically low expectations which are actually held by the individual whereas self-presented defensive expectations refer to individuals claims that they hold unrealistically low expectations when in fact they do not or when unrealistically low expectations are claimed to be present to a greater degree than is actually the case.

50

Self-protection strategies

Academic outcomes

Active selfhandicapping Persistence Self-pres selfhandicapping Grades (end of 1997)

Active defensive expectations Self-pres defensive expectations

Selfregulation

Future academic plans

Reflectivity

Figure 2.2 The proposed process model incorporating the association between both defensive expectations and reflectivity and academic outcomes Note. Broken paths estimated at Time 2 As noted earlier, while the present study deals with defensive expectations and reflectivity as correlated main effects, two other possibilities have been implied. The first is that defensive expectations and reflectivity interact (Norem & Illingworth, 1993). The second is that defensive expectations predict reflectivity (Norem, personal communication, March 26, 1997, March 7, 1998). Concerning the former, supplementary analyses directly examine the extent to which affective and motivational factors predict the interaction term as well as the extent to which this interaction predicts academic outcomes. In terms of the second, the present model, in which defensive expectations are correlated, is statistically equivalent to a model in which defensive expectations predict reflectivity. Notwithstanding this, in addition to a central model in which the two are correlated and operate as main effects, a supplementary model is tested that explicitly explores their relationship in the form of a beta path (rather than a correlation).

51

2.2.5 Summary of defensive pessimism Defensive pessimism is comprised of defensive expectations and reflectivity. In the present study, both strategies are considered primarily as main effects. Notwithstanding this, in supplementary analyses, defensive pessimism is examined in terms of (a) how defensive expectations and reflectivity interact and (b) the extent to which defensive expectations predict reflectivity. The findings in relation to the consequences of defensive pessimism are mixed: In the short term, defensive pessimism seems to be associated with high achievement, however, in the longer term, defensive pessimism seems less adaptive. The present study draws on the distinction between expecting an outcome and simply thinking about it and proposes that defensive expectations are negatively associated with academic outcomes while reflectivity is positively associated with these outcomes.

2.3 Drawing together self-handicapping and defensive pessimism For the most part, the literature has not explicitly drawn comparisons between selfhandicapping and defensive pessimism and so the contentions to follow are in parts speculative and in a number of cases not supported directly by empirical evidence. Indeed, assessing their empirical links is a central purpose of the present study. Arising from the preceding review of self-handicapping and defensive pessimism, the following features common to each of the two strategies are proposed: Both defensive pessimism and self-handicapping are used strategically to enable individuals to protect (or in some cases enhance) their self-worth. Importantly, however, it is predicted that of the two self-protective strategies, self-handicapping is more deleterious to key academic outcomes. As Norem and Cantor (1990a) note, defensive pessimism (and optimism), as strategies, are thus less likely to interfere with good performance than is self-handicapping, yet they provide similar cushioning or protection for self-esteem (p. 50). The phases involved in the process of self-handicapping and defensive pessimism are somewhat similar. According to Berglas (1987), self-handicappers first experience anxiety, they employ a self-handicapping strategy, their anxiety is alleviated, and they are able to perform successfully or in the case of failure are 52

protected by the availability of an excuse. According to Norem and Cantor (1986b), the defensive pessimist also experiences anxiety, strategically utilises expectations, achieves a sense of control, is better placed to succeed as a consequence, or has cushioned the blow of potential failure. Both strategies can be seen as something of a mixed blessing. Defensive pessimism seems to lead to success in the shorter term, yet is associated with diminished performance in the longer term. Self-handicapping too may lead to some success in the shorter term, yet according to Covington (1984a) in the longer term, is less adaptive. Self-handicapping can take self-presented or active forms. In the case of the former, the individual exaggerates circumstances or personal dispositions that might interfere with success in an upcoming task. In the latter, the individual chooses performance-inhibiting conditions under which to be evaluated or withdraws effort. Whilst not previously identified, it is quite conceivable that defensive expectations can also be self-presented or active. In the case of selfpresented defensive expectations, individuals may communicate to others that they are more pessimistic than they actually are. In the case of active defensive expectations, the individual actually holds unrealistically low expectations. The present study seeks to clarify the active and self-presented dimensions of selfhandicapping and defensive expectations There has been some suggestion that defensive pessimism is a form of impression management. Defensive pessimism is positively correlated with protective selfmonitoring, a dispositional motivation to avoid the disapproval of others and has been proposed to be a means by which individuals can protect their public image (Showers & Ruben, 1990). Self-handicappers are also concerned with impression management in that they are concerned with avoiding disapproval or negative evaluation by others (Shepperd & Arkin, 1989a; Strube, 1986). Both self-handicapping and defensive pessimism seem to activate negative selfschemas. On self-handicapping, Garcia and Pintrich (1994) report that self53

handicapping may be a strategy in which one engages when negative self-schemas are activated in the working self-concept (p. 136). Similarly, defensive pessimists can be thought of as being driven by negative failure self-schemas, perhaps arising from being in highly competitive environments that make salient differences in ability (Garcia & Pintrich, 1994, p. 137).

On these bases, self-handicapping and defensive pessimism are proposed to have a good deal in common. It is argued that the most fundamental way in which the two strategies are similar is in their objective to protect the self. A conceptual basis that underpins this need to protect the self is self-worth motivation theory and this is the broad perspective that the present study adopts in proposing an academic process model in which a set of affective and motivational factors give rise to selfhandicapping and defensive pessimism which in turn predict academic outcomes.

54

3 SELF-WORTH MOTIVATION AND PREDICTORS IN THE PROPOSED MODEL

3.1 Self-worth motivation theory The present study is concerned with identifying factors that give rise to the use of selfhandicapping and defensive pessimism and the consequences that follow from these strategies. As established in the preceding chapter, self-handicapping and defensive pessimism are two strategies that, for a large part, are self-protective and geared towards protecting individuals competence in the event of failure. Individuals motivation to protect the self is directly addressed by self-worth motivation theory and given this, the present study invokes this theory as a means of connecting selfhandicapping and defensive pessimism. Moreover, self-worth motivation theory is also a basis from which to select conceptually-cohesive constructs to incorporate as predictors in a model of the academic process as it relates to self-handicapping and defensive pessimism.

3.1.1 Setting the stage for self-worth protection: The educational context According to Covington (1992), to be able is to be worthy, but to do poorly is evidence of inability and is reason for despair. The essence of achievement motivation according to self-worth theory resides in this formula (p. 79). Classrooms promote failure avoidance and self-protection largely because there are very few rewards available to students. Reward tends to be a zero-sum game in which one student gains while the others miss out (Covington & Beery, 1976) and so in effect, students succeed at the expense of others. The means by which reward comes to influence failure-avoidance and self-protection is fairly simple. Students quickly come to learn that achievement gains reward and this reward is seen as evidence of approval. Importantly, this approval is also seen to be representative of the worth of the person. Covington and Beery (1976) invoke the commodity theory of success (Brock, 1968) to demonstrate how the scarcer the rewards, the more valued they become and that when they become highly valued, competition sets in. More importantly, the more valued rewards become, the greater is their evidence of high ability and conversely the more a lack of reward indicates evidence of a lack of ability. The implication is that the

55

majority of students will have to deal with perceived failure because rewards are scarce.

Thus classrooms become largely failure-oriented and because failure implies a lack of ability and a lack of ability reflects poorly on the worth of a person, students in the classroom become failure-avoidant. According to Covington (1984a, 1992), there are only a few ways to deal with the pressure of an environment oriented more towards avoiding failure than striving for success. Two strategies of particular relevance to the present study are (a) setting unrealistically low goals and (b) manipulating the scene a priori such that failure does not reflect poorly on ones ability (Covington, 1984a; Covington & Beery, 1976).

3.1.2 Setting the stage for self-worth protection: Ability becomes tied to selfworth The underlying theme in the preceding discussion is that the reason why failure is to be avoided at all costs is because failure implies low ability and low ability implies low self-worth. According to Covington (1989), the one main, even preeminant, reason that students achieve in school is to protect a sense of worth, especially in competitive situations (p. 88). Because the value of a person is typically measured in terms of ability and performance relative to others, it is not surprising that the students sense of esteem often becomes equated with ability - to be able is to be valued as a human being, but to do poorly in school is evidence of inability, and reason to despair of ones worth (Covington, 1992, p. 16).

The chief means by which an individual comes to equate ability with worth stems from the assumption in society that ones worth is measured on valued criteria such as academic success (Beery, 1975). Because of this, people feel the need to excel academically so as to gain the respect of others. As Beery (1975) notes,

if individuals have come to believe that they must possess an extremely high level of ability to be okay, to be worthwhile, and that anything less is tantamount to stupidity and makes them worthless, then any kind of evaluative situation is, naturally, going to be extremely threatening [and] . 56

. . given these internalised assumptions, their natural response, their most readily available coping response, is to make sure that they have an excuse ready in advance to make sure that their limitations in performance do not reflect on their ability (p. 200).

The achievement-related bases of self-worth begin very early in life when young children are aware of the brightest student in the class, are of the belief that ability is the best predictor of success, and see that success is the basis upon which people should be rewarded (Covington & Beery, 1976). Thus, ability becomes the focus of their scholastic experience. From here, the individuals sense of worth is threatened by the belief that his value as a person depends on his ability to achieve, and that if he is incapable of succeeding, he will not be worthy of love and approval (Covington & Beery, 1976, p. 6). Indeed, this is particularly relevant to young childrens experience which is very much grounded in the need for love and approval. Thus it is not surprising that they quickly come to value success derived from ability and in the threat of its absence, attempt to avoid failure. In fact, in one of the earliest studies of self-handicapping, Jones and Berglas (1978) cited the link between competence and both parental love and approval as a primary determinant of self-handicapping behaviour. In a later review of self-handicapping, Berglas (1990) claimed that the etiology of self-handicapping behaviors can be traced to parenting styles that blocked or inhibited the development of a sound, experience-based competence image (p. 154). Notwithstanding this, it is important to also recognise that ability comes to be valued not only because self-worth is bound up with it but also because it is instrumental in bringing about success (see Covington & Beery, 1976; Covington & Omelich 1982a, in Covington & Omelich, 1984b)

By the time students reach college, ability is the most predominant factor in determining how they define themselves academically (Covington & Omelich, 1984b) and value it to a greater extent than they value effort (Harari & Covington, 1981). In fact, there is a developmental trend from elementary school to college such that young students equate high effort with positive outcomes and high ability while older students see outcomes as due primarily to ability and actually devalue effort (Harari & Covington, 1981). Moreover, the most substantial contributor to their feelings of self57

worth is their estimate of their ability (Covington & Omelich, 1984b) and while success following effort yields pride and teacher reward (Covington & Omelich, 1979c), the preferred path to success is through ability (Brown & Weiner, 1984; Covington & Omelich, 1979b; Nicholls, 1976).

In fact, effort comes to be devalued primarily because of the implications it holds for individuals ability. Covington and colleagues (Covington & Omelich, 1979b; Covington, Spratt, & Omelich, 1980) have shown that students who fail having tried hard experience the greatest shame whereas failure following little study evokes least shame. It is not surprising, then, that students employ strategies such as selfhandicapping which are aimed at reducing effort attributions for failure - indeed, Covington and Omelich (1979b) have shown that diminished effort coupled with an excuse elicits less shame than diminished effort with no accompanying excuse. As Covington (1989) reports, the central, activating principle behind such self-defeating tactics is that effort represents a potential threat to the students sense of worth and self-esteem, because a combination of intense effort and failure implies lack of ability (p. 89).

3.1.3 Self-handicapping, defensive pessimism, and self-protection Self-handicapping, it has been argued, is a strategy aimed at protecting ones selfworth. Jones and Berglas (1978) report that self-handicapping derives from an abnormal investment in the question of self-worth (p. 203). They argued, consistent with the self-worth motivation perspective, that during childhood, self-handicappers are not confident of their parents unconditional love of and value for them. As a consequence, they come to learn that love and value follow from their ability to perform and please others. Under such circumstances, failure threatens the individuals self-worth and according to Jones and Berglas, he who tries and fails loses everything. He who fails without trying maintains a precarious hold on the illusion of love and admiration (p. 204). Indeed, in a subsequent elaboration of selfhandicapping, Snyder and Smith (1982) emphasised the self-worth basis of selfhandicapping by arguing that self-protective strategies are important insofar as they serve to foster an underlying sense of self-esteem and competence (p. 107).

58

But what happens when there is both a need to succeed and a desire to protect the self? According to Norem and Cantor (1986a), under such circumstances, students use other strategies. One of these is defensive pessimism which is employed when the individual is faced with a performance situation in which failure is threatening to self-esteem or self-image and success is highly desirable and important (Norem & Cantor, 1986a, p. 349). They go on to argue that unlike the self-handicapper, for the individual utilising this [defensively pessimistic] strategy, the attributional implications are important, but they do not override the significance of the performance itself (p. 349-350).

In sum, then, the dimension underpinning the constructs in the present study is the motive to protect the self. This, it is argued, is the primary factor giving rise to selfhandicapping and defensive pessimism. Whilst this constitutes the very broad conceptual basis of self-handicapping and defensive pessimism, there is a need to identify the specific constructs which directly predict self-handicapping and defensive pessimism. Self-worth motivation theory provides a useful framework with which to guide the selection of a set of constructs used to predict these strategies.

3.2 Factors predicting self-handicapping, defensive expectations, and reflectivity Adopting self-worth motivation theory as a basis for exploring the processes involved in self-handicapping and defensive pessimism brings into consideration a range of constructs pertinent to that theoretical position and which has relevance to a set of constructs predicting the two strategies central to it. As outlined earlier, the present study examines a model in which a variety of motivational and affective factors (underpinning a self-worth motivation) predict self-handicapping and defensive pessimism which in turn are proposed to predict academic outcomes such as selfregulation, persistence, future plans, and academic grades (see Figures 1.2 and 1.3 in Chapter 1). Below is a consideration of each of the motivational and affective predictors with particular focus on their relationship to self-worth motivation and their relevance to self-handicapping and defensive pessimism. Importantly, in a number of respects, self-worth motivation theory does not make direct predictions about the relevance of these predictor constructs. Rather, the rationale for selecting these

59

predictors for inclusion in the model is often inferred from - and not a direct consequence of - self-worth motivation theory.

3.2.1 Motivation orientation Motivation orientation in the present study is concerned with ego-orientation, taskorientation, and avoidance-oriented performance. Ego-orientation refers to the tendency of individuals to be concerned about their ability or competence and the tendency to feel successful when they have demonstrated superior ability relative to others. Task-orientation refers to the tendency of individuals to feel successful and gain satisfaction in mastering what they have set out to do (Duda & Nicholls, 1992; Nicholls, 1989). Avoidance-oriented performance has recently been proposed to constitute another component of motivation orientation and is concerned with the tendency of the individual to be motivated to perform so as to avoid failure, its negative consequences, and the demonstration of a lack of ability (Elliot & Church, 1997; Middleton & Midgley, 1997; Skaalvik, 1997a, 1997b). The review to follow deals firstly with ego-orientation and task-orientation, their emphasis on ability and effort respectively, and their relationship to self-handicapping and defensive pessimism. Following this, avoidance-oriented performance is reviewed, along with avoidance-oriented performers fear of poor performance, its negative consequences, and the consequent inclination to manoeuvre defensively. 3.2.1.1 Ego- and task-orientation5 The self-worth theory of motivation is primarily focused on individuals who behave in ways that protect their sense of ability and competence. Individuals operate in this way because they are inclined to equate self-worth with ability and so threats to their ability represent threats to their self-worth. One could predict, therefore, that individuals who are concerned about their ability relative to others or are concerned about demonstrating competence (ego-oriented individuals) would be more inclined to
5

In much of the literature, the terms ego-involvement, ego-orientation, performance goals, performanceorientation, and extrinsic orientation are used interchangeably (as are task-involvement, task-orientation, learning goals, mastery-orientation, and intrinsic orientation). Consistent with Greenwald (1982), it is acknowledged here that while there are differences between the terms, they can be subsumed under a broad banner of ego- and taskrelated dimensions that have as their core the notion that individuals are primarily concerned with demonstrating competence or gaining mastery. Whilst the present study is empirically concerned with ego- and task-orientation (the dispositional dimensions of ego- and task-involvement respectively), the review of literature adopts a broader

60

engage in strategies to protect their self-worth. As Midgley et al. (1996) propose, individuals who are concerned about proving that they are competent may choose to use these [self-handicapping] strategies to promote their image as able to others, especially if they perceive that they may perform poorly (p. 425). In contrast, individuals who are task focused are more concerned about mastery than demonstrating competence and so are not expected to engage in strategies designed to protect their self-worth.

3.2.1.1.1 Ability and effort and the perceived causes of success According to Jagacinski and Nicholls (1984, 1987), competence perceptions for taskinvolved individuals are based on mastery, learning, and the application of high effort. On the other hand, competence perceptions for ego-involved individuals are based on externally-referenced cues such as outperforming others or doing as well as others but with less effort. Jagacinski and Nicholls found that when ego-involved individuals used high effort and discovered that others did as well with less effort, they experienced greater embarrassment and guilt. Also, when they invested less effort than others and succeeded, ego-involved individuals reported a greater sense of accomplishment. Egoinvolvement, then, is associated with (a) the perception that an investment of effort may reflect a lack of ability, (b) the view that competence is primarily comprised of ability rather than effort, and (c) the perception that ability (rather than effort) is the main determinant of successful performance. These beliefs render ego-involved individuals self-worth particularly vulnerable in situations which might be challenging and/or threaten failure. Task-involved individuals, on the other hand, tend not to consider that investment of effort implies a lack of ability and are therefore more inclined to invest, rather than withdraw, effort if they feel that failure threatens.

In a study reflecting these perceptions, Duda, Fox, Biddle, and Armstrong (1992) found that task-oriented subjects saw success as being due to effort whereas egooriented subjects saw success as being due to deceptive tactics and/or ability. Duda and Nicholls (1992; see also Duda, 1993) also found that in a factor analysis of motivation orientation and beliefs about success in the classroom, task-orientation
approach and addresses the various ego- and task-related dimensions as are seen to be relevant to self-protection generally and both self-handicapping and defensive pessimism in particular.

61

loaded on one factor along with effort-related beliefs about success. On the other hand, ego-orientation loaded on one factor along with ability-related beliefs about success in the classroom.

Given, then, that ego-oriented individuals see the causes of success in various domains as being due to ability and given that ability is generally perceived as being relatively fixed and uncontrollable (see review of implicit theories of intelligence in Section 3.2.5), these students may be particularly vulnerable in achievement scenarios in which there is the possibility of failure. Moreover, on the basis of ego-oriented individuals belief that ability is the cause of success, they may be particularly vulnerable to the possibility that failure reflects low ability (Harackiewicz, Barron, & Elliot, 1998). Low ability, according to self-worth motivation theory, implies low selfworth. Thus, ego-oriented individuals would be inclined to engage in strategies aimed at protecting self-worth.

Related to ego-oriented students emphasis on ability is the issue of the extent to which individuals self-esteem is dependent on their feeling and demonstrating competence (esteem-relevant competence valuation). From both self-worth motivation and motivation orientation perspectives, individuals whose self-esteem is primarily dependent on their experience and demonstration of competence are vulnerable to threats to this self-esteem under conditions in which some form of evaluation of that competence takes place. Thus a construct akin to ego-orientation is esteem-relevant competence valuation. Esteem-relevant competence valuation has not received direct empirical attention in research to date and thus represents a new perspective on factors that not only underpin self-worth motivation but also predict self-handicapping and defensive pessimism.

62

3.2.1.1.2 Ego-orientation, task-orientation, and self-handicapping Similar to the present study, Midgley and Urdan (1995) sought to examine predictors of academic self-handicapping but focused on the middle school years. Amongst a number of hypothesised predictors were school task and school performance foci, individuals task and relative ability goals, and extrinsic goals. Correlations between these variables revealed significant positive associations between both school performance focus and extrinsic goals and self-handicapping. Negative correlations were obtained between both school task focus and individuals task goals and selfhandicapping. In separate regression analyses, amongst the significant predictors of self-handicapping were extrinsic goal orientation and school performance focus6. In a final regression analysis that included the significant predictors from previous analyses, the remaining significant predictors of self-handicapping included extrinsic goal motivation. More directly relevant to ego-orientation, Midgley et al. (1996) found self-handicapping to be positively correlated with ego-orientation. Similarly, another study in the educational context found that ego-oriented high school students tended to engage in avoidance strategies while task-orientation was positively associated with the practice of mastery strategies and negatively correlated with avoidance strategies (Lochbaum & Roberts, 1993).

Rhodewalt et al. (1984) investigated the withdrawal of effort and the report of physical problems and outside commitments as means of self-handicapping for competitive athletes. They found that before an important competition (a high egoinvolving condition) that posed a threat to self-esteem, high self-handicappers withheld effort (see also Smith et al., 1983; Snyder et al., 1985, in relation to selfhandicapping in response to evaluative settings and personally important tasks which pose a threat to the ego) whereas they did not do so in unimportant meets (low egoinvolving condition). In a second study, similar results were obtained for professional golfers.

In a study published subsequent to the present studys data collection, Urdan et al. (1998) found that class ability goals and the tendency for teachers to be ability-goal oriented both significantly positively predicted self-handicapping.

63

In a correlational study of individuals self-handicapping and achievement goals, Rhodewalt (1994) found that high self-handicappers were more likely to endorse performance goals. In respect to task-related dimensions, Rhodewalt found that low self-handicappers were more inclined to pursue learning, rather than performance goals. In another study, Pyszczynski and Greenberg (1983) found that participants reported investing less effort (i.e., self-handicapped) on a high ego-task than participants on a low ego-task (see also Greenberg, 1985).

It is important to note, however, that whilst the evidence pertaining to ego-orientation presented here suggests that it is a somewhat maladaptive orientation, there is evidence which shows that ego-related dimensions are not necessarily inimical to successful functioning in the educational environment. In a review of motivation orientation research, Harackiewicz et al. (1998) report that in terms of cognitive engagement, learning strategies, self-regulation, and performance, ego-related dimensions can be adaptive (see also Elliot & Church, 1997; Harackiewicz, Barron, Carter, Lehto, & Elliot, 1997; Meece, Blumenfeld, & Hoyle, 1988; Midgley & Urdan, 1995; Roeser, Midgley, & Urdan, 1996; Skaalvik, 1997a, 1997b). Middleton and Midgley (1997) suggest that the divergent findings pertaining to ego-related dimensions may be due to empirical confusion in past research between performanceapproach goals and performance-avoidance goals (hence the empirical distinction between ego-orientation and avoidance-oriented performance in the present study). Harackiewicz et al. (1998) also make the point that the extent to which ego-related dimensions are adaptive or not so adaptive may depend on the context. For example, Harackiewicz and colleagues have shown that in a competitive setting and for those high in achievement orientation, performance goals may be beneficial to motivation and performance (Elliot & Harackiewicz, 1994; Epstein & Harackiewicz, 1992; Harackiewicz & Elliot, 1993). The present position is one which recognises that egoorientation can be adaptive but is essentially an orientation which grounds success and failure in terms of ability (see Harackiewicz et al., 1998; Middleton & Midgley, 1997). From a self-worth motivation perspective this motivates the student to manoeuvre defensively (e.g., self-handicap), particularly in contexts (such as university) in which relative ability is salient.

64

In summary, then, the ego-relevance or ego-involving characteristics of a task are predicted to influence the tendency to self-handicap. Whilst there is not as much evidence to make firm hypotheses about the role of task-orientation, it is tentatively proposed that because task-orientation is generally associated with effort and adaptive learning strategies (Ames & Archer, 1988), the degree to which individuals are taskoriented is predicted to be negatively associated with self-handicapping.

3.2.1.1.3 Ego-orientation, task-orientation, and defensive pessimism Individuals who are ego-oriented are concerned with demonstrating their ability and outperforming others. It is quite a blow to their self-worth when they fail because ability is seen to be the cause of failure and low ability implies low self-worth. Another strategy that is argued to be aimed at protecting ones sense of ability is defensive pessimism, and more specifically, defensive expectations. It is contended that by creating lower and safer standards, individuals maximise their chances of relative success (relative, that is, to the lower standard) thereby avoiding judgements that they lack ability. Thus, ego-oriented individuals who are concerned with attaining success and protecting their ability image could be predicted to set unrealistically low expectations before an upcoming performance. These expectations are easier to attain and thus individuals avoid that which most threatens their ability perceptions - failure.

The empirical basis of the proposed link between motivation orientation and defensive expectations is more tentative and rests on a somewhat indirect body of research. It has been found that task-oriented individuals choose challenging, interesting tasks and persist with them whereas ego-oriented individuals low in perceived ability tend to choose very easy or very difficult tasks (Duda, 1992). Similarly, Koestner, Zuckerman, and Koestner (1987) found that ego-involved individuals choose significantly less challenging activity levels than task-involved individuals. Less challenging task choice could reflect defensive expectations. Ego-orientation has also been found to be positively correlated with negative thoughts about sporting activities, whereas task-orientation is negatively correlated with performance worry (Newton & Duda, 1993). Likewise, Diener and Dweck (1978) found that children who reflected a performance goal approach in the face of difficult tasks tended to focus on failure, while quite the opposite profile was found for children with a learning goal profile. 65

It is inferred from this relatively indirect evidence that ego-oriented individuals are more likely to be defensively expectant, whereas task-oriented individuals are predicted to be less pessimistic regarding upcoming performance. Predictions in relation to reflectivity are difficult to make. It may be that because thinking through various outcomes is self-protective (in that it prepares individuals for the eventual outcome) it may be positively associated with ego-orientation and negatively associated with task-orientation. Alternatively, it might be that reflectivity is more an effort-oriented, self-regulatory strategy aimed at enhancing control rather than protecting the individual and as such might be more positively associated with taskorientation and negatively associated with ego-orientation.

3.2.1.2 Avoidance-oriented performance According to Skaalvik, Valans, and Sletta (1994), ego-involvement can be separated into two dimensions, a more adaptive performance-oriented construct or a more avoidance-oriented construct. Skaalvik et al. report that

the core of ego-involvement is that students in the learning situation are preoccupied with themselves and how they are perceived by others. For some students this may result in trying to be best, to beat others or to establish their superiority over others, whereas for other students it may result in trying not to be poorest, to avoid looking stupid and to avoid negative reactions from others (p. 232-233).

In the past few years an avoidance orientation (performance-avoidance goals - Elliot & Church, 1997; Elliot & Harackiewicz, 1996; performance-avoid goal orientation Middleton & Midgley, 1997; Midgley, Kaplan, Middleton, & Maehr, 1998; defensive ego-involvement - Skaalvik et al., 1994; self-defeating ego-orientation - Skaalvik, 1997a, 1997b) has been identified by various researchers. Of the avoidance orientations identified, the operationalisation in the present study (referred to as avoidance-oriented performance) most closely reflects that proposed by Midgley and colleagues (Middleton & Midgley, 1997; Midgley et al., 1998). It is the Midgley et al. construct which is most congruent with the self-worth related notion of a motivation 66

to perform based on a fear of poor performance and its implications for ones ability image. The Skaalvik construct is more akin to public self-consciousnesss, while the construct proposed by Elliot and Church is more akin to worry and negative affect (see Midgley et al., 1998). In addressing avoidance-oriented performance, it is important to distinguish between this recent construct and the work avoidance construct originally included in the Motivation Orientation Scale (Duda & Nicholls, 1992). Avoidanceoriented performers in the context of the present study are those who work but do so primarily to avoid poor performance and its negative consequences. Work avoiders, on the other hand, are not motivated to work at all (see Duda & Nicholls, 1992; Middleton & Midgley, 1997).

It is proposed that avoidance-oriented performers are such because they hold doubts about their ability and contemplate the possibility of failure. According to Elliot and Harackiewicz (1996), for avoidance-oriented performers, the prospect of potential failure is likely to elicit anxiety, encourage self-protective withdrawal of affective and cognitive resources, disrupt concentration and task-involvement, and orient the individual toward the presence of failure-relevant information (p. 463). Importantly, however, by manoeuvring in a self-protective fashion, failure is not seen as failure in a strict sense but one mitigated by the circumstances in which it occurs. Essentially, then, when failure is perceived to be impending, avoidance-oriented performers are proposed to become motivated to act in a defensive fashion and two forms of this defensive manoeuvring are the primary foci of the present study.

Because avoidance-oriented performance has only recently received empirical attention there are no data linking it to self-handicapping and defensive pessimism (notwithstanding this, Middleton & Midgley, 1997, do predict that avoidance-oriented performance underpins self-handicapping). The data reviewed here, then, are indirectly associated with self-handicapping and defensive pessimism and so predictions regarding avoidance-oriented performance are advanced tentatively.

3.2.1.2.1 Avoidance-oriented performance and self-handicapping Given that failure is the basis of the threat to individuals self-worth it follows that individuals are motivated to avoid the implications of this failure. This motivation is 67

argued to be the crux of avoidance-oriented performance. It could therefore be inferred that individuals who are high in the motive to avoid failure and its negative consequences are more likely to engage in self-protective strategies such as selfhandicapping. As Skaalvik (1997b) posits, avoidance-oriented performers may develop a wide range of defensive strategies to avoid failure or at least the implications of failure, for instance, looking stupid (p.72).

Recent psychometric studies of an avoidance orientation have yielded data that could be indirectly associated with the tendency to self-handicap. For example Middleton and Midgley (1997) found that performance-avoidance goals were positively associated with avoidance of help-seeking and negatively associated with achievement. Skaalvik (1997a, 1997b; see also Skaalvik et al., 1994; Elliot & Harackiewicz, 1996) found self-defeating ego-orientation to be negatively correlated with achievement and motivation. Elliot and Church (1997) found performance avoidance goals to negatively predict achievement and intrinsic motivation. Elliot and Harackiewicz (1996) also showed that performance avoidance goals were negatively associated with intrinsic motivation. Also, procrastination, one form of selfhandicapping, has been associated with a fear of failure. Solomon and Rothblum (1984) found that a large proportion of a sample of undergraduates procrastinated before assignments and exams and a factor analysis of the reasons for their procrastination indicated that the first factor, accounting for almost half of the variance, reflected a fear of failure.

3.2.1.2.2 Avoidance-oriented performance, expectations, and pessimism As Teevan and Smith (1975) note, the individual who fears failure tends to set goals in such a way as systematically to reduce the probability of perceived failure. It is thus less the actual performance of the individual than his level of aspiration which will determine later feelings of success or failure (p. 968). According to Fried-Buchalter (1992), failure fearers tend to establish goals for themselves that are either unrealistically high or unrealistically low, thereby precluding any real test of ones actual ability (p. 369). Indeed, Tseng and Carter (1970) found that high school students with a strong need to avoid failure held significantly lower occupational aspirations. 68

Competitive trait anxiety may also reflect a fear of failure disposition (Passer, 1983). Passer explored aspects of trait anxiety and how individuals high on the trait appraised competition. It was found that individuals high on this dimension expected to perform less well in an upcoming competition, worried more frequently about making mistakes, not playing well, and losing, and expected more negative evaluation following failure than individuals who were low on the trait. Passer concluded that several generalised expectancies pertaining to fear of failure and fear of evaluation could serve as mediators of childrens competitive trait anxiety (p. 186).

Whilst these results do not relate specifically to defensive expectations, they do demonstrate a tendency for avoidance-oriented performers to regulate aspirations and expectations regarding future performances. Inferring from this, it is suggested that a motive to avoid poor performance and its negative consequences (avoidance-oriented performance) may underlie a tendency to strategically set unrealistically low expectations for upcoming performances in a bid to mitigate the implications of potential failure. Again, given the paucity of research in relation to reflectivity, predictions in relation to it are difficult to make and so the relationship between avoidance-oriented performance and reflectivity is a more exploratory component of the study.

3.2.1.3 Summary of motivation orientation In the context of the present study, motivation orientation comprises ego-orientation, task-orientation, and avoidance-oriented performance. Owing to their tendency to see success and failure as due to their ability, ego-oriented students are particularly vulnerable in the event of failure and to threats to their ability and subsequent selfworth. Consequently, ego-orientation is proposed to be positively associated with a tendency to engage in self-worth protection strategies such as self-handicapping and defensive expectations. Avoidance-oriented performers are concerned with avoiding demonstrations of a lack of ability and because of this, it is proposed that they too are particularly inclined to engage in self-worth protection. Accordingly, avoidanceoriented performers are predicted to be more inclined to self-handicap and hold defensive expectations. Task-oriented students, on the other hand, see effort as the 69

primary basis of academic outcomes, and as a consequence, their ability is not under threat to the extent that it is for ego-oriented students and avoidance-oriented performers. Therefore, task-orientation is proposed to be negatively associated with self-handicapping and defensive expectations. Because there is no empirical basis upon which to relate reflectivity to motivation orientation, no predictions are made. Motivation orientation (comprising ego-orientation, task-orientation, and avoidanceoriented performance) is incorporated into the process model in Figure 3.1.

Also incorporated in Figure 3.1 is esteem-relevant competence valuation. This too, is a more exploratory component of the study as there are no direct empirical bases from which to make specific predictions. Nonetheless, it is contended that individuals tendency to base their self-esteem on competence is somewhat akin to ego-oriented individuals tendency to focus on ability. An emphasis on competence renders individuals high in esteem-relevant competence valuation particularly vulnerable under conditions in which some evaluation of their competence takes place. Accordingly, it is predicted that individuals high in esteem-relevant competence valuation are more inclined than individuals low on the trait to engage in selfhandicapping and defensive expectations.

70

Motivational and affective predictors

Self-protection strategies

Academic outcomes

Active selfhandicapping Persistence Motivation orientation Self-pres selfhandicapping

Active defensive expectations Esteemrelevant comp valuation Self-pres defensive expectations

Selfregulation

Grades (end of 1997)

Future academic plans Reflectivity

Figure 3.1 The proposed process model incorporating the association between both motivation orientation (comprising ego-orientation, task-orientation, and avoidance-oriented performance) and esteem-relevant competence valuation and self-protection strategies Note. Broken paths estimated at Time 2

3.2.2 Public self-consciousness According to Nasby, the self-schema can be divided into one that organises and controls the processing of covert or private information about the self and one that organises and controls the processing of overt or public information about the self (1989, p. 117). Nasby further proposed that private self-consciousness is positively related to the former while public self-consciousness is positively related to the latter. It has been found that individuals high in public self-consciousness tend to concern themselves with how they appear to others, what public impression they create (Froming and Carver, 1981), how the self is being presented, and the reaction of others to that presentation (Fenigstein, 1979). 71

3.2.2.1 Public self-consciousness and self-handicapping It has been argued that self-handicapping is not only a strategy aimed at protecting the private self-esteem but also an impression management strategy designed to protect or enhance the self in a publics eye. Kolditz and Arkin (1982) were the first to specifically examine this issue. They proposed that self-handicapping is primarily an impression management strategy and that it should be more likely to occur when in the presence of an audience. Kolditz and Arkin found that in an insoluble test condition, participants tended to choose a performance-debilitating drug but only when the experimenter (an audience) was present and when they believed the experimenter would have access to their score on the second test. Similarly, MelloGoldner and Wurf (1997) found that in public conditions in which participants were cognitively focused on an audience, the likelihood of self-handicapping increased (see also Montgomery, Haemmerlie, & Zoellner, 1996).

Strube (1986) found that the shortened SHS correlated moderately with public selfconsciousness, social anxiety, and other-directedness. In other work, Ferrari (1992) found that the shortened SHS correlated moderately with public self-consciousness and social anxiety. Midgley and Urdan (1995) examined dispositional correlates of academic self-handicapping and found that public self-consciousness was significantly associated with academic self-handicapping - although the relationship was weak. In another study, Schill, Morales, Beyler, Tatter, and Swigert (1991) correlated the shortened SHS with Schills (1990) measure of self-defeating personality. They found that three of the four significant correlations were concerned with the others dimension of the self-defeating scale and implied an other-directedness or public orientation associated with self-handicapping.

These findings raise the question as to whether self-handicapping is more a public or a private strategy. Baumeister and Scher (1988) make the point that self-defeating behaviour can often have impression management advantages. In their review of selfdefeating behaviour they concluded that focus on oneself, especially public attention to oneself, was repeatedly associated with self-destructive choices (p. 12). Indeed, Tice and Baumeister (1990) found that individuals high in self-esteem practiced less 72

than those low in self-esteem in a public condition and seemed to be engaging in an impression management strategy aimed at protecting the public rather than the private self.

Indeed, it could be predicted that when people fail only in private and have not yet spoiled their public image, they will self-handicap for an upcoming performance in order to maintain and protect that public image (Baumgardner, Lake, & Arkin, 1985). If this is the case, then self-handicapping can be seen as a public protective strategy occurring when a positive public image or an unknown public image may be spoiled. In a study by Baumgardner et al., participants were given failure feedback. When they believed the experimenter was unaware of the failure, they claimed a mood that was a self-handicap for an upcoming test. When the experimenter knew of the previous failure, they did not self-handicap. These results confirm the protective impressionmanagement view of self-handicapping.

In a study of the effects of public self-consciousness, ego-involvement, and their interaction on the tendency to self-handicap, Shepperd and Arkin (1989a) argued that self-handicapping is designed to protect individuals public image rather than their private image. Consistent with hypotheses, they found that individuals high in public self-consciousness listened to music they believed to be performance debilitating, but only when the task was characterised as being a valid indicator of academic success. Similarly, Ferrari (1991) found that procrastination was positively associated with public self-consciousness. Ferrari concluded that procrastinators are anxious individuals who actively avoid evaluative information, particularly if there is an audience . . . view their self-worth as based solely on ability at a task . . . [and] prevent judgements of performance and, hence, judgements of ability (p. 456-457).

In another investigation of the public dimension of self-handicapping, Hobden and Pliner (1995) recruited a group of perfectionist students who they claimed were highly likely to self-handicap due to their excessive fear of failure. According to Hewitt, Flett, Turnbull-Donovan, & Mikail (1991), perfectionism can be operationalised into private and public dimensions. Private perfectionists are those who have an intrinsic need for perfection, whereas public perfectionists are socially prescribed perfectionists 73

who are motivated to appear perfect to others. They predicted that socially (publicly) prescribed perfectionists would self-handicap only in a public condition. Findings supported this prediction and confirmed the publicly-oriented dimension of selfhandicapping.

There is a reasonable body of evidence, then, to support the view that public selfconsciousness and self-handicapping are related. The balance of evidence suggests that self-handicappers, whilst being to some degree concerned with protection of a private self-esteem (see Berglas & Jones, 1978), are also very much concerned with protection of the public image. Public self-consciousness is therefore predicted to be positively associated with self-handicapping. As Self (1990) notes, when there is an audience that requires convincing, it may appear safer to actively impede ones performance in order to insure that the audience will be convinced that one could not have done better (p. 52).

3.2.2.2 Public self-consciousness and defensive pessimism If it is to be assumed that there is a self-worth protection basis to defensive pessimism, it is not unreasonable to propose that consistent with other self-protection strategies, a concern with the public image may also underlie it. One factor that has been associated with the tendency to manage impressions or self-present is Snyders selfmonitoring construct. It has been posited that self-monitoring is comprised of two dimensions: acquisitive and protective (Lennox, 1988). Acquisitive self-monitoring involves an active self-presentation style geared toward accommodating oneself to the situation in order to accumulate status and social approval (p. 286). These people tend to be optimistic, want to get ahead rather than get along, and are high in selfesteem. On the other hand, protective self-monitoring involves an anxious, passive attempt to avoid social disapproval and gain acceptance by conforming to external expectations (p. 286). According to Polak and Prokop (1989), there is a striking resemblance between defensive pessimism and protective self-monitoring where both behavior styles are associated with lower self-esteem and are oriented toward the avoidance of disapproval and disappointment (p. 286). Importantly, they found that protective self-monitoring was correlated with defensive pessimism (r=.70) to a far greater extent than acquisitive self-monitoring was correlated with defensive 74

pessimism (r=.12). They concluded that persons who are concerned with avoiding disapproval and rejection and who view academic achievement as important to social approval may be motivated to harness their anxiety by means of defensive pessimism (p. 288).

Whilst technically not the same as defensive pessimism, worrying is along similar affective lines in that it evokes a negative self-schema tied to ones expectations. Pruzinsky and Borkovec (1990) have argued that because social-evaluative fears are predominant among worriers, we expected worriers to be high on public selfconsciousness (p. 508). Consistent with this reasoning, they found that worriers were higher in public self-consciousness. Whilst this does not directly enable predictions regarding defensive pessimism, it does suggest that publicly self-conscious individuals would be more inclined to entertain negative thoughts about upcoming performances.

3.2.2.3 Summary of public self-consciousness It is clear that individuals who are publicly self-conscious are concerned with the impression they are conveying and seek to avoid negative evaluation. It is on this basis, and previous empirical findings, that public orientation is predicted to be associated with self-handicapping and defensive expectations. The issue of reflectivity is not clear primarily because it has not been widely assessed to date nor has it been associated with a public orientation. It may be that because publicly-oriented individuals tend not to focus on the self, they would be unlikely to be reflective. On the other hand, because publicly-oriented individuals are concerned with their public image, they may be likely to entertain all possible outcomes in order to prepare themselves publicly for the given outcome. At any rate, public self-consciousness is incorporated into the cumulative academic process model (see Figure 3.2) as a factor proposed to give rise to self-handicapping and defensive expectations, while the relationship between public self-consciousness and reflectivity is a more exploratory component of the model.

Motivational and affective predictors

Self-protection strategies

Academic outcomes

Active selfhandicapping

75

Persistence Motivation orientation Self-pres selfhandicapping

Esteemrelevant comp valuation

Active defensive expectations

Selfregulation

Grades (end of 1997)

Public selfconsciousness

Self-pres defensive expectations Future academic plans Reflectivity

Figure 3.2 The proposed process model incorporating the association between public selfconsciousness and self-protection strategies Note. Broken paths estimated at Time 2

3.2.3 Stability and level of self-concept/esteem The review of literature to follow deals primarily with the issue of level and stability of global self-esteem. However, because this study addresses specific academic domains (viz. maths and verbal) and because self-esteem is a more global construct (Marsh, 1988b, 1990), it is proposed that a general self-esteem measure may not be the most appropriate means by which to assess competence-related appraisals of the self. In contrast, self-concept is typically operationalised as a domain-specific measure of perceived competence, skill, and confidence (Marsh, 1988b, 1990) and as such is proposed to be more relevant to the present studys issues. Whilst no research has addressed the issue of self-concept and both self-handicapping and defensive pessimism, there is direct and indirect evidence in relation to self-esteem and these two strategies. Thus the conceptual review deals primarily with self-esteem reflecting more generalised appraisals of the self - and extrapolates from this evidence to propose an association between self-concept and self-protection strategies.

3.2.3.1 Stability of self-concept/esteem

76

Research into self-concept and self-esteem has typically explored both their level and their stability. Without doubt, however, the bulk of work has examined the level of the self-concept and esteem, whilst only a smaller body has explored their stability. Selfconcept/esteem stability refers to the magnitude of short-term fluctuations in appraisals of the self (Kernis, Grannemann, & Barclay, 1992, p. 622; see also Rosenberg, 1986). A conceptually similar construct, self-concept clarity or certainty, refers to the extent to which the contents of an individuals self-concept . . . are clearly and confidently defined, internally consistent, and temporally stable (Campbell, Trapnell, Heine, Katz, Lavallee, & Lehman, 1996, p. 141). Certainty, then, is also comprised of stability to some degree. In fact, the measure of trait certainty that Campbell and colleagues (1996) have devised incorporates items that explicitly reflect stability much along the lines of Rosenbergs (1965) measure of self-esteem stability. Because of the conceptual overlap between the two and because recent measures of certainty explicitly include items measuring stability, both stability and certainty are examined in this review of literature. Importantly, however, it is stability that is the focus of the empirical component of this study.

3.2.3.1.1 Sources of instability Individuals with an unstable self-concept/esteem tend to over-rely on others evaluations of them (Rosenberg, 1986). These individuals tend to place greater emphasis on how they are perceived by others and this can create instability in a number of ways. First, the impression others have of us is often shrouded in ambiguity and because others are also inclined to engage in impression management, we can never truly ascertain how we are viewed by them. Arising from this is some instability regarding the self-concept. Second, the variety of people in individuals lives all have differing views (to varying degrees) of them. Because individuals depend on these appraisals in helping to shape self-perceptions (Kinch, 1963), it follows that they may develop a self-concept that is not particularly stable. Third, because individuals are concerned with the impression they are making, they may present varied images depending on the audience (Juvonen & Murdock, 1993) and these variations may contribute to some diffuseness or instability of self-concept. Finally, Kernis and Waschull (1995) found that the more individuals base their self-worth on competence, the more unstable is their self-esteem. 77

3.2.3.1.2 Stability of self-esteem and self-handicapping It has been suggested that not only is the level of self-esteem important in influencing the tendency to self-handicap, so too is the certainty and stability of that self-esteem. For example, Berglas and Jones (1978; see also Jones & Berglas, 1978) argued that it may not be chronically low or high self-esteem individuals who are likely to selfhandicap but those whose self-esteem is uncertain. As Harris and Snyder (1986) note, the person who is most likely to self-handicap is one who experiences at least some level of self-esteem, but perhaps an uncertain self-evaluation plagued by doubts of performing successfully in an esteem-threatening, evaluative situation (p. 451). Consistent with this, Arkin and Oleson (1998) report that doubt contributes to individuals tendency to self-handicap and that this arises from a lack of belief in a core, stable self (p. 318). In the classroom, Riggs (1992) noted that students preoccupation with their competence image and the extent to which their selfappraisals are unstable are crucial to the development of a disposition toward selfhandicapping (p. 252). According to Riggs, if students belief in their competence is precarious, the fear of exposing a lack of competence may lead them to adopt protective strategies. Indeed, Harris and Snyder (1986) found that study participants uncertain in their self-esteem tended to withdraw effort as a self-handicapping strategy. Similarly, Newman and Wadas (1997) found that individuals who were unstable in self-esteem were more inclined to self-handicap.

3.2.3.1.3 Self-esteem stability and defensive pessimism Attempts to link self-concept stability with defensive pessimism are restricted by the paucity of relevant evidence and so the predictions made are advanced tentatively. It is contended that individuals whose self-concept is unstable are less certain of their ability to meet challenge successfully. Thus they are unsure about whether they can avoid failure and in response to this may set lower and safer standards. As has been argued, defensive pessimists create lower and safer standards against which to be evaluated and this may be a means of meeting the challenge. They will therefore be seen (and see themselves) in some way as less of a failure against the safer standard. Also, in the face of uncertainty, defensive expectations may act as a cushion (Norem & Cantor, 1986b). Moreover, given the link between self-concept instability and self78

worth protection (Kernis & Waschull, 1995), a proposed link with defensive expectations is feasible. In relation to reflectivity, because individuals unstable in selfconcept are unsure as to how they can meet upcoming challenges, they are unsure as to the outcome. Because they are unsure of the outcome, it might be predicted that they will contemplate all possible outcomes as a means of preparation and gaining control. Again, however, given the lack of existing evidence in relation to reflectivity, its function in the model is exploratory.

3.2.3.2 Level of self-esteem

3.2.3.2.1 Level of self-esteem and self-handicapping Other research has examined self-handicapping as it relates to the level (rather than the stability) of self-esteem. Given that self-handicapping occurs when ones competence image is threatened, it is plausible . . . that individual differences in trait self-esteem may be associated with differential tendencies to engage in selfhandicapping (Tice & Baumeister, 1990, p. 444). Self-handicapping findings in relation to level of self-esteem, however, have been mixed. Whilst some research has found individuals high in self-esteem to be more likely to self-handicap, other work suggests that individuals low in self-esteem are more likely to do so. Yet another body of work has suggested no systematic relationship between the two. Before examining this research, it is again important to note that studies of self-handicapping and defensive pessimism typically operationalise self-appraisals in the form of global selfesteem. The present study, however, assesses self-protection within specific academic domains (viz. maths vs verbal) and because of this, operationalises self-appraisals in the form of domain-specific self-concept. Thus predictions regarding domain-specific self-concept based on the self-esteem literature are made more cautiously than if the present study utilised global self-esteem as the self-appraisal measure.

In evidence attesting to the positive association between self-esteem and selfhandicapping, Harris, Snyder, Higgins, and Schrag (1986) found that study participants high in self-esteem were more likely than participants low in self-esteem to nominate protective excuses for failure. Alloy (1982) examined self-handicapping tendencies amongst depressed (lower self-esteem) and non-depressed (higher self79

esteem) participants and found that the non-depressed subjects were more likely to engage in self-enhancing or defensive strategies so as to present themselves more favourably or protect themselves from the implications of failure. Tice and Baumeister (1990) found that when given no feedback, high self-esteem participants practiced less than low self-esteem participants whereas when given success feedback these differences were eliminated. On the other hand, Ferrari (1991) found that study participants who chose to self-handicap were lower in self-esteem. Also, Midgley and Urdan (1995) found low self-worth significantly correlated with and predicted selfhandicapping. Similarly, Weary and Williams (1990) found that depressed study participants were more likely to self-handicap than non-depressed participants and that this was motivated to avoid further losses in self-esteem. This evidence is supported by findings of Midgley et al. (1996) who found that negative self-esteem was significantly associated with protective self-handicapping whereas positive selfesteem was not.

Other work, however, has found self-esteem to be relatively independent of selfhandicapping. In a recent study, Rhodewalt and Hill (1995; see also Rhodewalt et al., 1991) found that after being informed of failure, participants were more likely to nominate a self-handicapping excuse irrespective of the level of their self-esteem. Similarly, Rhodewalt and Fairfield (1991) found that irrespective of level of selfesteem, high self-handicappers withdrew effort in a task in which they anticipated failure. In fact, Jones and Rhodewalt (1982, in Rhodewalt, 1990) have suggested that when associations between self-handicapping and level of self-esteem have been found, it is due to shared method variance in which respondents are asked in both scales to concede to something unflattering about themselves. Moreover, Berglas (1987) argued that it was not so much level of self-esteem that influenced the tendency to self-handicap, but its certainty and stability. Indeed, these findings may be consistent with those of Tice (1991) who found that individuals high in self-esteem would self-handicap as much as those low in self-esteem, but for different reasons: Individuals high in self-esteem would self-handicap to enhance success, whereas individuals low in self-esteem would self-handicap to protect themselves in the event of failure. At any rate, in a discussion of future directions in self-concept research, Hattie and Marsh (1996) suggested that further work be conducted into strategies 80

designed to protect, enhance, and/or maintain the self-concept. Amongst the variety of strategies they addressed was self-handicapping. 3.2.3.2.2 Level of self-esteem and defensive pessimism7 Perhaps it is not surprising that self-esteem is negatively associated with defensive pessimism. Norem and Cantor (1986b), for example, found that defensive pessimists were significantly lower in self-esteem than optimists. A wide variety of work on general pessimism and optimism has also acknowledged the role of self-esteem. Amongst school-aged children, self-esteem and pessimism are negatively associated (Fischer & Leitenberg, 1986). Amongst college students, pessimism is negatively associated with self-esteem while optimism is positively associated with it (Davis, Hanson, Edson, & Ziegler, 1992; see also Weinglert & Rosen, 1995). In work on expectations, self-esteem also plays a part. Abel (1996) found that self-esteem is positively associated with expectancies of success, while Drwal and Wiechnik (1984) found that compared to those low in self-esteem, individuals high in self-esteem held higher expectancies for success. They also found that expectancies of individuals high in self-esteem were more resistant to failure that those of individuals low in selfesteem.

In a study published subsequent to data collection, Eronen et al. (1998) found that defensive pessimists were significantly lower in self-esteem and higher in depression than optimists.

81

3.2.3.2.3 The issue of level of self-concept Whilst the preceding evidence deals primarily with global self-esteem and its association with self-handicapping and defensive pessimism, it is important to note that the present study addresses these two strategies in a domain-specific context. Specifically, the constructs involved in the process model are assessed in both the maths and verbal academic domains. Thus, as noted earlier, global self-esteem may not be the most appropriate means by which to assess appraisals of the self. Selfconcept, on the other hand, is a domain-specific self-appraisal and instruments have been devised that specifically target its multidimensional domain-specific dimensions (Marsh, 1990, 1992).

The indirect evidence regarding self-concept and achievement behaviour suggests that students high in self-concept are not likely to engage in self-protective behaviour. In summarising his findings of the previous decade, Marsh (1990) reported that of the studies examining the relationship between self-concept and achievement, all correlations between verbal self-concept and verbal achievement indicators were positive and significant as were correlations between math self-concept and math achievement. Moreover, in terms of the causal ordering of self-concept and achievement, Marsh (1990) concluded that the effects of academic self-concept are causally predominant (p. 122) over school grades.

It has been proposed that constructs such as self-efficacy are similar to multidimensional domain-specific measures of self-concept (Norwich, 1987) and are distinct from global self-esteem. The theory and evidence in relation to self-efficacy (Bandura, 1982, 1986) also suggest that individuals high in self-efficacy are not inclined to engage in maladaptive self-protective behaviour. First, self-efficacious individuals tend to choose challenging activities that ultimately enhance development in a given area and optimise the likelihood of success on future tasks. Second, whereas highly self-efficacious individuals are more likely to expend effort and more likely to persist in the face of failure, individuals low in self-efficacy tend to readily abandon challenging tasks. Third, individuals low in self-efficacy display a tendency to dwell on deficiencies and view tasks as more difficult than they actually are. 82

Finally, individuals low in self-efficacy tend to react to negative outcomes with resignation or apathy, whereas individuals high in self-efficacy react with activism.

In sum, then, the evidence suggests that individuals high in self-concept or selfefficacy behave in ways in which expectations are high and challenge is embraced. Moreover, the application of effort and the readiness to persist in the face of challenge suggest that individuals high in self-concept or self-efficacy are unlikely to behave in a defensive fashion in the form of defensive expectations and self-handicapping. As Bandura (1986) reports,

people who see themselves as efficacious set themselves challenges that enlist their interest and involvement in activities; they intensify their efforts when their performances fall short of their goals, make causal ascriptions for failures that support a success orientation, approach potentially threatening tasks non-anxiously, and experience little in the way of stress reactions in taxing situations (p. 395).

Taken together, the weight of the direct evidence in relation to self-esteem and defensive manoeuvring is mixed, while the indirect evidence in relation to selfconcept/efficacy and achievement behaviour suggests that individuals higher on these dimensions are unlikely to behave in self-defeating ways.

3.2.3.3 Interaction of level and stability of self-concept/esteem It may be that genuinely high self-esteem individuals are not threatened by negative feedback because they are confident of their competence. On the other hand, defensively high self-esteem individuals who are uncertain of their high self-esteem may also be high in the need for approval, be threatened by negative feedback, and act in ways to restore or maintain their high self-esteem (McFarlin & Blascovich, 1981).

Some of the research into self-esteem stability has interacted stability with level of self-esteem. Kernis, Grannemann, and Barclay (1989) argued that individuals with high but unstable self-esteem have the most to lose from a self-esteem threat. Although they possess a positive self-concept, it is fragile and sensitive to evaluative 83

information (p. 1014). If they do not react in a defensive manner, they may be forced to revise their self-appraisal and this might represent too great a threat to them. On the other hand, stable high self-esteem individuals have little reason to feel threatened by provocations, because they are secure in their self-views and less sensitive to evaluative information (p. 1014). Consistent with this, it has been speculated that individuals high but unstable in their self-esteem are concerned with maintaining a positive self-view and that this should reflect itself in heightened sensitivity to evaluative events that have potentially positive self-relevant implications (Kernis et al., 1992, p. 624).

On the other hand, unstable low self-esteem individuals are concerned about avoiding a continuously negative self-view [and this] will be related to the greater use of strategies that reduce the diagnostic implications of failure (Kernis et al., 1992, p. 624) - indeed, this is the primary objective of self-handicapping and to some degree, defensive expectations. Therefore, for individuals low and unstable in self-esteem these might be two strategies that may help them avert developing stable negative views (Kernis et al., 1992, p. 637). Moreover, individuals stable and low in selfesteem have a chronic dislike for themselves and given that self-protective strategies are used to protect oneself against negative self-feelings, it seems illogical that they would be used by people whose negative self-feelings are unlikely to change (Kernis, Cornell, Sun, Berry, & Harlow, 1993, p. 1191).

In a study that assessed the effects of the interaction of level and stability of selfesteem, Kernis et al. (1993) examined the effect of positive and negative interpersonal feedback on undergraduates high or low and stable or unstable in self-esteem. They found that individuals high but unstable in self-esteem reacted most favourably to positive feedback and most unfavourably to negative feedback. Unstable high selfesteem individuals also minimised the relevance of the negative feedback, reduced the credibility of the source of that feedback, and offered more excuses for that performance than stable high self-esteem individuals. Unstable low self-esteem individuals did not react more favourably to positive feedback than stable low selfesteems, nor did they react more negatively to unfavourable feedback - they accepted this feedback and offered no excuses. Thus individuals high but unstable in self84

esteem were more likely to engage in self-protection and self-enhancement following negative and positive interpersonal feedback respectively, whilst individuals low but unstable in self-esteem did not do so.

Thus for high self-esteem individuals, instability reflects fragility in positive selffeelings, which promotes externally directed self-protective strategies (Kernis & Waschull, 1995, p. 132). The interaction of level and stability of self-concept, then, whilst not central to the study, is incorporated into supplementary analyses. However, whilst individuals high and unstable in self-esteem appear to be more inclined to engage in self-protection strategies, the evidence is slim and so no specific predictions are made in relation to the precise nature of the interaction.

3.2.3.4 Summary of stability and level of self-concept/esteem In terms of stability, the weight of evidence suggests that individuals with shaky self-appraisals may be more inclined to engage in self-protective strategies such as self-handicapping and defensive expectations. While direct evidence in relation to the level of self-esteem and self-handicapping appears mixed, the indirect findings in relation to self-concept and self-efficacy suggest that individuals high on these dimensions are unlikely to engage in self-defeating behaviour. In view of this equivocal evidence, then, no specific predictions are made in relation to level of selfconcept and self-handicapping. The evidence in relation to level of self-esteem and defensive pessimism is more clear-cut and it is predicted that self-concept and defensive expectations are inversely related. In terms of reflectivity, there is no direct evidence and so again, this is incorporated as an exploratory component of the study. Level and stability of self-concept are incorporated into the cumulative process model shown in Figure 3.3. Whilst not indicated in the model, the effects of the interaction between level and stability of self-concept are also examined in supplementary analyses.

85

Motivational and affective predictors

Self-protection strategies

Academic outcomes

Active selfhandicapping Public selfconsciousness Self-pres selfhandicapping Motivation orientation Persistence

Active defensive expectations

Selfregulation

Grades (end of 1997)

Esteemrelevant comp valuation Level and stability of self-concept

Self-pres defensive expectations Future academic plans Reflectivity

Figure 3.3 The proposed process model incorporating the association between both level and stability of self-concept and self-protection strategies Note. Broken paths estimated at Time 2

3.2.4 Perceived control Perceived control over future outcomes is also a factor associated with self-protection (Thompson, 1994). In an achievement context, perceived control refers to the extent to which individuals feel they are able to avoid failure and achieve success. Perceived control is relevant to self-worth motivation in that individuals who feel they have little or no control over outcomes are increasingly uncertain as to whether they can avoid failure or bring about success. Given that it is failure which is most threatening to individuals competence image and self-worth, those who perceive little control over outcomes may be more likely than individuals who perceive control to engage in strategies that can protect them in the event of failure. Two dimensions of control are examined in the review of literature. The first relates to feedback contingencies, an operationalisation of control examined in much of the self-handicapping literature. The second examines control in terms of self-reports of control.

86

3.2.4.1 Contingency of feedback and self-handicapping Non-contingent feedback refers to the administration of reward or punishment to individuals independent of their behaviours or responses. Individuals who come to see no relationship between their responses and feedback develop a perception of uncertain control over future outcomes (Perry & Dickens, 1984; Self, 1990; see also Skinner, Wellborn, & Connell, 1990 in relation to the classroom). In the educational context,

simply rewarding children for various behaviours without regard for the effects of reinforcement schedules and without corrective feedback may have the paradoxical effect of decreasing the probability of desired behaviors, or at least, of creating a context in which the student becomes unclear as to what behaviors are desirable (Kleinhammer-Tramill, Tramill, Schrepel, & Davis, 1983, p. 62).

Similarly, it has been proposed that cognitive, motivational, and emotional deficits produced from either positive or negative non-contingent outcomes create expectations that interfere with subsequent learning, reduce motivation, and lower self-esteem (Perry, Magnusson, Parsonson, & Dickens, 1986, p. 97). Consistent with this, Perry et al. (1986) found that individuals in a high non-contingent feedback group reported less control than a low non-contingent group (see also Skinner et al., 1990).

By the time students reach university, they have experienced a great deal of noncontingent feedback. For example, Brophy (1981) found that teachers can be quite inaccurate in administering praise. Brophy found that only eleven percent of students correct responses are praised properly. As one component of a qualitative observational analysis of school climate and achievement, Brookover, Schweitzer, Schneider, Beady, Flood, and Wisenbaker (1978) found that in low achieving schools, students were not often reinforced for their performance. At other times, slower students were positively reinforced for incorrect answers and there was also some confusion in that the same reinforcement could often be administered for both incorrect and correct answers. Indeed, in a review of classroom conditions which give 87

rise to self-worth protection, Thompson (1994) identified non-contingent feedback to students as a prime contributor.

In a study exploring individuals choice of a performance-inhibiting drug when trying to protect a tenuous but positive competence image, Berglas and Jones (1978) found that sixty percent of participants who received non-contingent success feedback selected a performance-inhibiting drug, whilst only twenty percent of participants who received contingent success feedback did so. Moreover, Berglas and Jones contended that non-contingent success decreases confidence in ones ability whilst contingent success has the opposite effect. When confidence in ones ability is low, one is more likely to take steps to protect the uncertain self-esteem in the face of an upcoming performance. As has been discussed, one way to do this is to engage in selfhandicapping. As Baumeister and Scher (1988) argue, the central cause of selfhandicapping appears to be some form of induced insecurity about future performances, especially when coupled with high external expectations for success (p. 8). According to them, insecurity arises from non-contingent feedback.

Thus students receiving non-contingent success feedback are uncertain how they attained success and are therefore uncertain as to how to maintain it. In the classroom, because the worth of students is dependent on the demonstration of competence, they do not want to jeopardise this worth by displaying low competence. Consequently, those uncertain about their capacity to repeat their demonstration of competence will take steps to protect what competence gains they have made. According to Riggs (1992), wanting to protect the private image won from their previous success, and afraid that they will not be able to repeat that success, these individuals may selfhandicap so that a future failure can be attributed to the handicap rather than to the lack of competence (p. 251).

In a study by Rhodewalt and Davison (1986; see also Hobden & Pliner, 1995), participants who had previously received contingent or non-contingent failure or success feedback were given a choice to perform a test in the presence of performance-debilitating music, performance-facilitating music, or neutral music. Participants in the non-contingent success group tended to choose the debilitating 88

music option. Similarly, Higgins and Harris (1988) found that heavy social drinkers who received non-contingent success feedback drank more than contingent success and contingent failure participants. Also, non-contingent failure participants drank more than those in the contingent success group. In a study of self-handicapping through self-reported pain, Mayerson and Rhodewalt (1988) found that study participants receiving non-contingent success feedback reported greater pain prior to a second test than participants receiving contingent feedback. Similarly, it has been found that when individuals receive success feedback that is not contingent on their effort, they are more likely to choose extremely difficult performance goals (Greenberg, 1985). In sum, then, non-contingent success feedback has been found to increase the likelihood of self-handicapping behaviour (see also Kolditz & Arkin, 1982; Tucker et al., 1981). Importantly, whilst non-contingent success feedback is most likely to lead to self-handicapping, the evidence in relation to non-contingent failure feedback is more equivocal. For example, whilst Rhodewalt and Davison (1986) found that non-contingent failure does not evoke self-handicapping behaviour, evidence in a study by Weidner (1980) suggested otherwise.

Taken together, these findings demonstrate that non-contingent feedback (primarily that which involves success) can influence the tendency to self-handicap. Noncontingent success feedback elicits an uncertainty regarding ones ability to repeat previous success and non-contingent failure feedback elicits uncertainty regarding ones ability to avoid further failure. Essentially, then, the issue of non-contingent feedback is about the control individuals perceive they have over future success and failure and when perceived control is low, individuals see it as more expedient to risk failure with a handicap as an alibi, than to strive for success with no chance of attribution deflection.

89

3.2.4.2 Self-reported control and self-handicapping Self-reported control is the second form of control considered in this review of literature and is the form of control operationalised in the empirical component of the present study. However, no research to date has examined the issue of self-reported control and self-handicapping. Accordingly, much along the lines of other constructs in the present study, the predicted association between self-reported control and selfhandicapping is inferred from related research. Schmitz and Skinner (1993) found that control correlated with effort and that the more effort exerted, the better the performance. Patrick, Skinner, and Connell (1993) reported that perceived control strongly predicted persistence, attention, effort, and participation in a sample of 8-10 year old students. Moreover, unknown control negatively predicted behaviour and emotion. Skinner et al. (1990) showed that the more contingent a teachers reinforcement, the more control over future outcomes the students reported. In addition to this they demonstrated that engagement (persistence and effort) and achievement were associated with higher perceived control. Harter and Connell (1984) also found that uncertain control was negatively correlated with achievement, mastery motivation, competence evaluation, and competence affect (feeling good about ones work). Similarly, Connell (1985) reported that students high on an uncertain control dimension were likely to score significantly lower on IQ tests and teachers ratings of their competence. These students were also low in mastery orientation, perceived competence, and autonomous judgement. Taken together, this set of indirect evidence suggests that students low in perceived control are not inclined to engage in behaviour consistent with an adaptive achievement orientation. On this basis, it can be inferred that such students are more likely than students high in perceived control to engage in counterproductive achievement behaviour such as selfhandicapping.

3.2.4.3 Perceived control and pessimism It has been proposed that in situations in which people are uncertain about obtaining important outcomes, they strategically manage their cognitions in an attempt to influence their future affective responses to the outcomes in question (Pyszczynski, 1982, p. 387) so as to maximise positive affect and minimise negative affect. One manifestation of such strategic manoeuvring can be related to expectations in which 90

unexpected positive outcomes are more satisfying than expected positive outcomes and expected negative outcomes are not so hard-hitting as unexpected negative outcomes (Feather, 1969). Indeed, this tendency is analogous to defensive expectations wherein individuals set unrealistically low expectations before an upcoming performance and in this way establish safer standards against which to be judged whilst also cushioning the self from the disappointment that failure engenders.

When presented with each of four outcomes (controllable positive outcomes, controllable negative outcomes, uncontrollable positive outcomes, and uncontrollable negative outcomes), participants in a study by Zackay (1984) were required to estimate the probability of it happening to them and the probability of it happening to someone else. Zackay found participants judged controllable negative events as less likely to happen to them than to another person and controllable positive events as more likely to happen to them than to another person. Participants, then, were more optimistic in relation to controllable events. Consistent with this, Forsyth and McMillan (1981) found that people hold more positive expectancies when they attribute outcomes to controllable causes. In another study, Weinstein (1980) found a strong negative correlation between the controllability of a negative event and the extent to which undergraduate subjects saw it as happening to them. Weinstein (1980) concluded that if an event was one they perceived to be controllable and if they were committed to a particular outcome, the majority of factors they brought to mind were ones that increased the likelihood that it would turn out the way they would like (p. 819).

In terms of control as a function of feedback contingencies, Albersnagel, Arntz, and Gerlsma (1986) found that non-contingent feedback was associated with lowered expectancies for future success. Similarly, Schunk (1983) found that following noncontingent reward, students set unrealistically low expectations to ensure success and such self-serving tendencies were probably less pronounced among performancecontingent children who developed greater skill and likely were certain of their capabilities (p. 517).

3.2.4.4 Summary of perceived control 91

While the connection between self-reported control and self-handicapping has not been made in the literature to date, it is one which may not be unreasonable (particularly given the strong association between non-contingent feedback and selfhandicapping) and which is explored further in the present study. In terms of the indirect evidence, low control is also associated with low effort, less participation, and diminished engagement, three features characteristic of self-handicapping. According to Rhodewalt (1990),

the self-handicapper who is caught in a hopeless situation may entertain the belief that he or she could be effective if it were not for the handicap. This line of speculation leads to the nomination of control-related individual differences as candidates for the prediction of self-handicapping (p. 95; see also Arkin & Baumgardner, 1985).

The evidence also suggests that perceived control can influence the extent to which an individual is pessimistic or holds negative expectations regarding future events. There are some data which also suggest that perceptions of low control may be specifically related to defensive expectations (see Schunk, 1983). No evidence, however, is available to enable predictions about perceived control and reflectivity. Given that reflectivity is a means by which individuals gain control through their consideration of a diversity of outcomes, it may be that those with perceptions of low control over future outcomes may seek to redress this through reflectivity. Notwithstanding this possibility, however, the influence of self-reported control on reflectivity is an exploratory component of the study. Perceived control is incorporated into the expanding process model shown in Figure 3.4.

92

Motivational and affective predictors

Self-protection strategies

Academic outcomes

Public selfconsciousness

Active selfhandicapping Persistence

Motivation orientation

Self-pres selfhandicapping

Esteemrelevant comp valuation Level and stability of self-concept

Active defensive expectations

Selfregulation

Grades (end of 1997)

Self-pres defensive expectations Future academic plans

Perceived control

Reflectivity

Figure 3.4 The proposed process model incorporating the association between perceived control and self-protection strategies Note. Broken paths estimated at Time 2

3.2.5 Views of the nature of intelligence Implicit theories of intelligence refer to the beliefs individuals hold about the nature of intelligence. Individuals tend to see intelligence as something that is a fixed and immutable entity or as an incremental dimension that can be changed or improved upon. It is contended that individuals theoretical perspective on the nature of intelligence may be important in determining how they attain feelings of competence and self-worth as well as being relevant to the use of academic strategies such as selfhandicapping and defensive pessimism.

93

3.2.5.1 Implicit theories and self-worth The implicit theories about intelligence to which individuals adhere can render their self-worth vulnerable, particularly under conditions of evaluation in which there is the possibility of negative judgements about ones performance. According to Dweck and Leggett (1988),

within an entity theory individuals are not simply judging a momentary level of ability. Rather, they may be judging what they perceive to be an important and personal attribute. Thus an entity theory may place ones intelligence on the line in evaluative situations, magnifying the meaning and impact of negative judgements (p. 263).

Individuals self-worth stems in part from the goals that they adopt in response to their theories of intelligence. According to Dweck (1991), it is interesting to think of the different theories as different self-concepts and of their allied goals as ways of building and maintaining self-esteem within those self-conceptions (p. 208). Thus, entity theorists feel good about themselves when they demonstrate relative ability whereas incremental theorists feel good about themselves when they develop mastery (Dweck & Leggett, 1988). The problem with the entity view, then, is that it puts selfesteem and self-development in conflict with each other (Dweck, 1991, p. 209). Thus entity theorists self-esteem needs counter their learning needs, whereas for the incremental theorist, self-esteem needs are consistent with their learning needs because self-esteem increases when learning and mastery increase (Dweck, 1991).

The implicit theory one adopts also has implications for the tendency to engage in protective behaviour. Importantly, individuals who believe intelligence is increasable, pursue the learning goal of increasing their competence, whereas those who believe intelligence is a fixed entity are more likely to pursue the performance goal of securing positive judgements of that entity or preventing negative judgements of it (Dweck & Leggett, 1988, p. 262-263). Incremental theorists tend to adopt learning goals and embrace tasks that involve the possibility of making errors, are difficult, and which, therefore, involve some possibility of failure. Entity theorists, on the other hand, tend to choose tasks in which success is assured, do not provide 94

challenges, minimise the risk of failure, and protect them from threats to their competence image. In short, entity theorists should adopt goals that tend to involve positive judgements of their intelligence or avoid negative judgements of their intelligence . . . low confidence should make them . . . attempt to conceal their perceived lack of ability from an evaluator (Dweck & Bempechat, 1983, p. 245; see also Clark & Tollefson, 1991). Indeed, it is entity theorists emphasis on ability that renders them vulnerable to self-worth protection. Because they are dominated by the conviction that ability alone is a necessary condition for success and that lack of ability is a sufficient explanation for failure (Covington, 1992, p. 82), any performance demands are directly relevant to their ability image and consequent selfworth. In the face of this, it is proposed that entity theorists are most inclined to engage in self-protection (see also Covington 1984a).

Consistent with this, Zhao and Dweck (1992, in Dweck, Hong, & Chiu, 1993) found that following failure in hypothetical situations, entity theorists were more likely to engage in ability-related derogation (I would think that I was dumb) and selfderogation (I would feel I was worthless) than incremental-view theorists. Hong and Dweck (1993, in Dweck et al., 1993) found that following failure, entity theorists took longer to react to ability-related words due, they claimed, to a heightened emotional reaction that ability posed for them following failure. Indeed, Bempechat, London, and Dweck (1991) found that amongst older primary school children, following failure, those with an entity view took longer to recover and complete subsequent problems. In the social domain, entity theorists are more impaired by failure in social encounters (Lepore, Kiely, Bempechat, & London, 1989) and also seem to be more sensitive to rejection and concerned with popularity (Benenson & Dweck, 1986).

3.2.5.2 Implicit theories and self-handicapping In a study of self-handicapping and implicit theories of intelligence, Rhodewalt (1994) examined the views of ability held by high and low self-handicappers. He found that self-handicapping was negatively correlated with a learning/task approach to study and positively correlated with a performance/ego approach. When interpreting these findings, Rhodewalt contended that high self-handicappers tend to have a fixed-entity view of ability and low self-handicappers have an incremental view of ability. 95

According to Rhodewalt, the implications of failure are so damaging [to the fixedentity theorist] that it is to be avoided at all costs. Those with an incremental orientation, on the other hand, have much less to lose when they encounter negative feedback. In fact, such information is useful because it suggests that self-corrective action is needed (p. 82).

Some indirect evidence also suggests that an entity view of intelligence should be associated with self-handicapping while an incremental view should be negatively associated with it. In an investigation of study skills in high school students, Jones, Slate, Blake, and Sloas (1995) found that students with an incremental view had more adaptive study skills in terms of help seeking, organisation of work, and punctuality of assignments. Similarly, Stipek and Gralinski (1996) found that entity views correlated with superficial engagement while incremental or effort-related beliefs correlated with active engagement. Jones, Slate, Marini, and DeWater (1993) obtained a significant negative correlation between an entity view of intelligence and academic study skills, particularly as they related to an investment of effort, information processing, and note-taking. Inferring from these results and also from the direct evidence in the Rhodewalt (1994) study, it is predicted that an entity view of intelligence is positively associated with self-handicapping, whilst an incremental view of intelligence is negatively associated with it.

3.2.5.3 Implicit theories and (defensive) expectations In a review of self-worth motivation amongst college students, Covington (1997) highlighted the link between ability attributions and lowered expectancies for success. When dealing with the issue of negative expectations, Covington reported that lowered expectations occur because among adults, at least, ability is typically perceived to be a fixed, immutable factor, and because ability is also believed to be the preemptive cause of academic success (p. 66; see also Harari & Covington, 1981). Indeed, Covington (1989) recommended that amongst the variety of steps educational institutions can take to minimise the self-worth implications of students academic experience is the promotion of an incremental view of intelligence. More indirectly, Koestner, Aube, Ruttner, and Breed (1995) examined the issue of implicit theories and expectations in a sample of children with some mental retardation. These 96

children typically have low confidence in their abilities and this tends to lead to maladaptive motivational patterns. They argued, however, that low confidence in ones abilities will not lead to helpless behaviours if children possess an incremental theory of their capacities (p. 58). Children in their study were encouraged to think about ability in incremental terms and to avoid normative comparisons with other classmates. Koestner et al. found that when this incremental approach was emphasised, these children exhibited the same motivational pattern as a matched sample of children without mental retardation in the sense that they embraced high challenge, implying some level of optimism.

Whilst these results do not constitute direct evidence in establishing a link between intelligence beliefs and defensive expectations, they do suggest that incremental theorists may be more optimistic and aspire to more positive outcomes. Additionally, given that entity theorists perceive ability and intelligence to be fixed, they are less likely to see potential failure as surmountable. When this is the case, individuals may be more inclined to play it safe and engage in protective strategies to mitigate the implications of potential failure. Furthermore, given that entity theorists see outcomes as reflecting ability more than effort, failure in this context is more threatening and so self-protection concerns may be salient. Consistent with this rationale, then, it is tentatively predicted that an entity view is associated with self-protection generally and with defensive expectations more specifically.

Predictions regarding reflectivity are not so straightforward. It may be that incremental theorists attempt to gain control by thinking through various possible outcomes, presumably because it is they who believe they can enhance control through their efforts. The entity theorists, on the other hand, may see little to be gained through their efforts and thus not think through possible outcomes in a bid to enhance control. Alternatively, the entity theorists may be uncertain as to their ability to influence outcomes and therefore entertain all possible outcomes so as to prepare themselves for their occurrence. Given these competing possibilities, no predictions are made in relation to individuals views of intelligence and reflectivity.

3.2.5.4 Summary of implicit theories of intelligence 97

Direct evidence linking implicit theories of intelligence to self-handicapping and defensive pessimism is limited in the case of the former and apparently non-existent in the case of the latter. Thus, predictions regarding the role of implicit theorising about intelligence in the proposed process model are advanced tentatively. Because entity theorists focus on their ability they are more inclined to see success and failure in these terms rather than in terms of effort. Consistent with a self-worth motivation perspective, because threats to individuals ability constitute threats to their selfworth, entity theorists may be more inclined to alter the meaning and implications of failure and engage in self-worth protection strategies such as self-handicapping and defensive expectations. Again, there is no evidence to enable predictions regarding reflectivity and so this issue is exploratory. The implicit theory construct is incorporated into the expanding process model shown in Figure 3.5.

3.2.6 Attributions The final component of the proposed process model involves attributions for success and failure. The causes individuals attribute to events can determine how they behave on future occasions. According to Weiner, once a cause, or causes, are assigned, effective management may be possible and a prescription or guide for future action can be suggested (1985, p. 548). In the context of the present study, future actions are viewed in terms of self-protection strategies in the forms of self-handicapping and defensive pessimism.

98

Motivational and affective predictors

Self-protection strategies

Academic outcomes

Public selfconsciousness Active selfhandicapping Motivation orientation Esteemrelevant comp valuation Level and stability of self-concept Perceived control Self-pres selfhandicapping Persistence

Active defensive expectations

Selfregulation

Grades (end of 1997)

Self-pres defensive expectations Future academic plans Reflectivity

Implicit theories

Figure 3.5 The proposed process model incorporating the association between implicit theories of intelligence and self-protection strategies Note. Broken paths estimated at Time 2

The perceived cause of an outcome is hypothesised to vary primarily along three dimensions: locus, stability, and controllability (Weiner, Frieze, Kukla, Reed, Rest, & Rosenbaum, 1971; Weiner, 1985). Locus refers to the extent to which an event is internal or external (whether the cause is central to the person or whether it is more related to the environment or other external factors). Stable causes refer to those which are likely to persist over time. Unstable causes are those which are likely to be susceptible to change over time. Controllability reflects the extent to which a cause is within an individuals control.

99

3.2.6.1 Attributions, expectancy, and affect According to Weiner (1985), attributions influence performance by way of expectancies and affect. Expectancies refer to the beliefs one holds about causing success or avoiding failure in the future. The causes to which one attributes past success or failure can determine what these expectancies will be. Importantly, it is the stability and controllability dimensions of causality which most contribute to expectancies. Attributing failure to stable and uncontrollable factors such as ability leads to a decrease in expectancy of success (see also Covington, 1989). On the other hand, failure attributed to unstable and controllable factors such as effort leads to an increase in expectancy of success (see also Covington, 1989). It is these expectations which contribute to better performance and motivation in the future and help bring about success or enhance persistence in the face of failure.

The second way in which attributions impact on academic outcomes is through individuals affective reactions to these attributions. The individual, it is claimed, makes an attribution regarding the cause of an event and following this is an emotional reaction to the perceived cause (Weiner, 1985). In fact, Weiner (1985) outlines affective reactions distinct to each attribution. For example, internal success brings about pride and enhances self-esteem; internal failure brings about a decrease in self-esteem; internal controllable failure (e.g., lack of effort) brings about guilt; and internal uncontrollable failure (e.g., lack of ability) evokes shame (see also Covington, 1989; Covington & Omelich, 1984b). Amongst these attributions, however, ability seems to be pivotal because failure due to it evokes both shame and humiliation whereas failure due to effort evokes guilt and primarily impacts on humiliation via ability attributions (Covington & Omelich, 1984b).

Importantly, these affective reactions in conjunction with expectations influence future behaviour. Weiner (1994) later presented a model of achievement striving which accounted for findings using the affective component of the model: Individuals who attribute failure to a lack of ability (uncontrollable and internal) experience shame and embarrassment and this leads to a performance decrement. On the other hand, individuals who attribute failure to a lack of effort (controllable and internal) react with guilt and this prompts greater effort in the future and enhanced subsequent 100

performance. Covington and Omelich (1990, in Covington, 1997) further proposed that shame impacts negatively on achievement through the individuals expectancies because shame triggers doubts about ones abilities whenever academic tasks are encountered.

3.2.6.2 The role of ability attributions Covington and Omelich (1984b) proposed that the importance of effort in failure is relevant to affective responses primarily to the extent that effort has implications for ability and that it is ability which triggers the affective responses of shame and humiliation. Indeed, in response to Brown and Weiners (1984) query as to why shame and pride are more closely linked to ability rather than effort, Covington and Omelich (1984b) argue, consistent with a self-worth motivation perspective, that in competitive situations such as school and university, ability is the dominant cognition (p. 167).

Covington and Omelich (1984a) explored the full attributional model proposed by Weiner and found that shame following failure was mainly due to low ability attributions. Humiliation was also more strongly predicted by ability than effort while guilt was predicted more strongly by effort than ability. Most importantly the impact on performance and persistence from failure attributions largely derived from ability attributions via humiliation. Indeed this was earlier supported by Covington and Omelich (1979a) who examined the causal impact of attributions and found that, whilst the causal role of failure attributions was not strong, the attribution having the major impact on performance via affect was ability. Similarly, in a later study, Covington and Omelich (1985) found that failure due to low ability evoked most humiliation and most shame whereas failure due to effort evoked most guilt. Moreover, the effect of effort on humiliation and shame via ability was stronger than the direct effect of effort on humiliation and shame. Ames (1984) also found a strong relationship between ability attributions and affect and not much support for a relationship between effort attributions and affect. It may be, then, that because failure due to low ability evokes such negative affective reactions, individuals attributing in this way are more inclined to manoeuvre such that the implications of failure are not

101

so threatening to ability. Thus, individuals who see past failure as due to ability may be more inclined to self-handicap.

In terms of expectations, the findings are consistent with contentions by Covington (1989). Diener and Dweck (1980) found that helpless school students who attributed failure to ability tended to underestimate the number of problems successfully solved. These students also tended to see themselves as performing more poorly than mastery students. Moreover, they did not view success as predictive of future success and this was the case even after being successful on one set of problems. According to Diener and Dweck (1980), overall, the results show that if there is a way to devalue ones present performance, the helpless children are likely to make use of it (p. 950). They go on to argue that

perhaps the anticipation of failure prevents the helpless children from being even more adversely affected by failure when it does occur. That is, if, following successes, they allowed themselves to believe that they had high ability but were still prone to see failure as indicating a lack of ability, then the occurrence of failure might have even greater negative impact (p. 951).

In sum, it seems that ability attributions following failure are most predictive of ways individuals respond to academic scenarios and as such are proposed to be central to the present studys model of academic engagement and motivation. Because ability is stable and relatively uncontrollable, individuals who attribute in this way are in a number of respects rendered failure-prone and so may be more inclined to play it safe and self-protect. Moreover, individuals who see ability as the cause of failure are inclined to be most threatened in the event of failure because of the implications low ability has for self-worth. Accordingly, those who see outcomes as due to ability may be particularly vulnerable to threats to self-worth and thus more inclined to engage in self-protective strategies such as self-handicapping and defensive pessimism.

3.2.6.3 The role of external attributions

102

Students who attribute success to external factors such as luck, good teaching, or easy marking are also inclined to behave in ways that are consistent with the helpless profile (Abramson, Seligman, & Teasdale, 1978). These individuals perceive little control over their ability to maintain success primarily because outcomes seen as due to external factors are to a large extent seen as uncontrollable. Helpless learning profiles have been associated with maladaptive academic behaviours such as low effort, low persistence, poor self-efficacy, and low expectations (Diener & Dweck, 1980; Schunk, 1981). Moreover, in a review of self-worth protection in the classroom, Thompson (1994) argued that an important means of remediating strategies such as self-handicapping is to encourage students to make internal attributions for success.

External attributions for failure are also proposed to influence the extent to which individuals self-protect. In terms of self-handicapping, for example, Rhodewalt (1990) reports that a person who characteristically construes negative self-relevant outcomes as externally caused . . . might be the person most drawn to self-handicapping (p. 103). Indeed, Rhodewalt and Hill (1995; see also Murray & Warden, 1992) found that self-handicappers were particularly inclined to make external attributions for poor performance. Taken together, the evidence suggests that attributions associated with the tendency to self-protect include ability attributions for failure and external attributions for success and failure. Consistent with this evidence, these attributions are predicted to be associated with self-handicapping and defensive pessimism.

3.2.6.4 Summary of attributions Whilst the present study explores a variety of attributional dimensions, it is proposed that three in particular - ability attributions for failure, external attributions for success, and external attributions for failure - are relevant to self-protection. Ability attributions for failure and external attributions for success and failure lead individuals to believe there is little they can do to avoid failure or maintain success and so these individuals, fearing failure, behave in ways to mitigate its implications. Covington and Omelich (1979a) have shown that ability attributions are predominant amongst failure-prone students and as such are proposed to give rise to strategies which enable the student to alter the meaning of that failure (see also Weiner & Kukla, 1970). External attributions for success and failure are also proposed to be predominant 103

amongst failure-prone students because such attributions instil in students an uncertainty as to their ability to repeat success and avoid failure (Weiner & Kukla, 1970). This uncertainty is proposed to give rise to strategies designed to self-protect. According to Covington (1984a), it is difficult to imagine a more devastating pattern of attributions: blaming oneself for failure, yet taking little credit for success. Feeling that one is at the mercy of capricious forces beyond ones control is demoralizing, especially when failure threatens self-worth (p. 94). Accordingly, attributions are incorporated into the cumulative process model shown in Figure 3.6.

Motivational and affective predictors

Self-protection strategies

Academic outcomes

Public selfconsciousness Active selfhandicapping Persistence Esteemrelevant comp valuation Level and stability of self-concept Perceived control Self-pres selfhandicapping

Motivation orientation

Active defensive expectations

Selfregulation

Grades (end of 1997)

Self-pres defensive expectations Future academic plans Reflectivity

Implicit theories

Attributions

Figure 3.6 The proposed process model incorporating the association between causal attributions and self-protection strategies Note. Broken paths estimated at Time 2

104

3.3 Integrating the predictors into a higher order structure It is proposed, then, that inferring from self-worth motivation theory, a variety of factors can be incorporated into a model predicting self-handicapping and defensive pessimism. These factors include, public self-consciousness, motivation orientation, esteem-relevant competence valuation, implicit theories of intelligence, perceived control, level and stability of self-concept, and attributions. However, in adopting a self-worth motivation perspective to guide the selection of factors underpinning selfhandicapping and defensive pessimism, not only is there considerable conceptual overlap amongst the constructs, but it is anticipated that these factors also share much empirical variance. Given this, it may be useful to deal with the predictors in terms of clusters (viz. higher order factors) than as separate constructs (see Cohen & Cohen, 1983). Variable clustering in the form of higher order factors can also render the proposed model more parsimonious and enhance the focus of analyses.

In terms of the first cluster, it is evident that a number of factors have underpinning them a performance orientation: Ego-orientation refers to feelings of success derived from successful performance; esteem-relevant competence valuation refers to the dependence of students self-worth on performance; public self-consciousness denotes students concerns about how others view their performance; and avoidance-oriented performance refers to a motivation to perform arising from a fear of the negative consequences following poor performance. Essentially, these constructs reflect students concerns with performance and a motivation to perform.

A second cluster comprises external attributions for success and failure. Consistent with previous findings (Marsh, 1984; Marsh, Cairns, Relich, Barnes, & Debus, 1984), these two factors are expected to be highly correlated and are proposed to reflect an external attributional orientation such that students attributing success to external factors (e.g., luck or task difficulty/ease) are also predicted to be inclined to make external attributions for failure.

A third factor cluster reflects individuals views of intelligence. It is proposed that entity and incremental views of intelligence are inversely related in representing this 105

cluster. A fourth cluster is proposed to reflect students uncertainty as to their personal control over avoiding failure and maintaining success. It is anticipated that attributions of failure to ability and low perceived control are correlated more highly with each other than with other constructs and that these two factors reflect an uncertain personal control dimension. Similarly, individuals unstable in self-concept are proposed to have a shaky belief in their capacity and so are also uncertain regarding their control over future outcomes. Accordingly, unstable self-concept is proposed to represent another facet of uncertain personal control.

It is unclear as to how level of self-concept and task-orientation can be incorporated into these clusters. It may be that level of self-concept is negatively associated with uncertain personal control. It may also be that task-orientation is inversely associated with performance orientation. However, given that task- and ego-orientation have been found to be positively correlated (Duda & Nicholls, 1992; Harackiewicz et al., 1997; Skaalvik, 1997b), this is unlikely. Essentially, then, correlations amongst the first order factors will assist in the interpretation of self-concept and task-orientation and their role in the proposed higher order factor structure.

The present study, then, explores the possibility of a higher order factor structure to represent the predictors. It is proposed that this may be a fruitful perspective from which to view the model not only in terms of obviating potential statistical problems such as suppressor effects, but also because it enables a more parsimonious construction of factors giving rise to self-protection strategies and enhances the focus of the model. This higher order factor structure in the context of the proposed process model is shown in Figure 3.7.

106

HIGHER ORDER AFFECTIVE AND MOTIVATIONAL PREDICTORS

Public selfconsciousness

SELF-PROTECTION STRATEGIES

ACADEMIC OUTCOMES

Egoorientation Avoidance orientation Esteem-relevant comp valuation

Performance orientation

Task-orientation

Selfhandicapping

Self-regulation

External External causes causes of of success success External causes of failure Ability causes of failure

External orientation Reflectivity Persistence

Grades (end of 1997)

Uncertain personal control Defensive expectations Future plans

Perceived control

Instability self-concept

Self-concept level

Entity beliefs Implicit theories Incremental beliefs

Figure 3.7 Proposed higher order structure of affective and motivational predictors and its role in the Time 1 process model Note. Broken paths estimated at Time 2 107

3.4 The generality vs specificity of constructs Whereas some researchers construct generalised measures of academic affect (e.g., self-concept, attributions for success and failure, motivation, anxiety, locus of control) that are intended to broadly represent all academic subjects, others are interested in the development of students achievement-related motivations, beliefs, affects, and behaviors that are specific to particular content domains or course subjects (e.g., Eccles, Midgley, & Adler, 1984; Eccles, Wigfield, Harold, & Blumenfeld, 1993; Gottfried, 1982; Gottfried, Fleming, & Gottfried, 1994; Harter & Jackson, 1992; Marsh, 1988b, 1990, 1993a; Nolen & Haladyna, 1990). A further issue to consider, then, concerns the issue of domain specificity. As foreshadowed earlier, the present study assesses the issue of self-protection in the context of the academic domain. Within this general domain, two content areas are examined - maths and verbal - with a view to establishing the extent to which motivational and affective factors, selfprotection strategies, and academic outcomes operate distinctly from one domain to another.

Support for the domain specificity of academic affect is strongest in self-concept research. For example, Marsh, Byrne, and Shavelson found that correlations between math and English self-concepts based on each of three different instruments were close to zero. Marsh and Craven (1997; Marsh, 1990, 1993a) summarised a growing body of research showing that verbal and math self-concepts are nearly uncorrelated and that the effects of academic self-concept on subsequent outcomes (school grades, test scores, academic effort, persistence, coursework selection) are also very content specific. Based on this accumulated research, they questioned the usefulness of measures of general academic self-concept as a summary of self-concepts in specific school subjects.

Marsh (1986, 1990; Marsh, Byrne, & Shavelson, 1988) developed the Internal/External Frame of Reference Model (I/E Model) to explain the extreme content specificity of math and verbal self-concept. According to the I/E Model, academic self-concept in a particular subject is formed in relation to an external (social comparison) reference in which students compare their self-perceived 108

performances in a particular subject with the perceived performances of other students in the same subject and an internal (ipsative-like) reference in which students compare their performances in the particular subject with their own performances in other subjects. Hence students may have a favourable math self-concept if mathematics is their best subject even if they are not particularly good at mathematics. The external comparison process leads to a positive math-verbal correlation, the internal comparison process leads to a negative math-verbal correlation, and the combination of the two counter-balancing processes results in a near-zero correlation depending on the relative importance of each comparison process. Support for the I/E Model has been found in a large number of self-concept studies based on different self-concept instruments, achievement measures, nationalities, and, of particular relevance to the present investigation, ages from preadolescence to early adulthood (see reviews by Marsh, 1990, 1993a; Marsh & Craven, 1997). There appear, however, to be limits to the I/E Model effect. For example, whilst Skaalvik and Rankin (1995) have shown that the I/E Model holds for general measures of math and verbal selfconcept (such as that measured by Marshs SDQ), it does not hold for more specific self-efficacy-like math and verbal constructs (see also Skaalvik & Rankin, 1990; Sletta, Valas, Skaalvik, & Sobstad, 1996) nor is the near-zero correlation evident for students who perceive their achievement in the two areas to be relatively equal (Skaalvik & Rankin, 1992).

Marsh (1988b) also demonstrated good support for the generality of the I/E Model based on measures of anxiety and achievement in mathematics and English. Math and English anxieties, like the corresponding self-concept measures, were nearly uncorrelated. Speculating on the reason why the I/E Model - based on self-concept research - worked so well with anxiety responses, Marsh suggested that it may be that supposedly distinct affective constructs - anxiety, self-concept, self-efficacy, attributional dispositions, locus of control, intrinsic motivation, etc. - all measure largely the same construct. That is, whereas math and English affects are clearly distinguishable, the different affective constructs may not be (p. 147). He went on to note that tests of these speculations would require researchers to examine relations between math and verbal affect based on a wide variety of different affective constructs. 109

Educational researchers have also demonstrated the need to distinguish between math and verbal domains in a variety of other academic affective constructs. Marsh (1984, 1986; Marsh et al., 1984) demonstrated the domain specificity of attributions for academic success and failure, but found that it varied substantially depending on the particular attribution. Thus, for example, attributions to ability as the basis for academic success and failure were very domain specific, whereas attributions to effort and particularly external attributions (e.g., luck and task difficulty) showed greater generalisability across different school subjects. Vispoel and Austin (1995), using a critical incident approach to measuring attributions, also found that content specificity varied substantially for different attributions. They reported that internal attributions were very domain specific, whereas external attributions were not. Gottfried (1982) measured anxiety and intrinsic motivation in four school subjects (reading, math, social studies, and science) and concluded that the necessity for assessing both academic intrinsic motivation and anxiety within specific subject areas appears to be paramount (p. 213). Galloway, Leo, Rogers, & Armstrong. (1996) also found significant differences in mastery orientations of students between maths and English subjects. Motivational researchers (e.g., Eccles et al., 1993; Maehr, 1984; Maehr & Braskamp, 1986; Pintrich, 1988; Pintrich & Garcia, 1991; Pintrich, Smith, Garcia, & McKeachie, 1991; see also Skaalvik & Rankin, 1995) have also emphasised the domain specificity of constructs such as expectancy for success (defined in terms of perceived competency, anxiety, and self-concept) and task value (defined in terms of interest, usefulness, and challenge). Stipek and Gralinski (1996), however, reported that entity and incremental views of intelligence in math and social science did not result in content-specific factors, thus supporting Dwecks (1986) proposition that these two theories are theories of general intelligence, rather than specific to different domains.

In summary, the domain specificity in math and verbal self-concepts generalises to some other academic affects. For other constructs, however, math and verbal scales are at least moderately correlated or, in some cases, so highly correlated that it may be reasonable to subsume the math and verbal scales into a single construct. A better understanding of the relations between math and verbal scales for different constructs, 110

however, requires that the same or comparable populations of students complete parallel math and verbal scales constructed for a wide variety of different academic affective constructs and that these responses are analysed with appropriate confirmatory factor analysis models. The present study seeks to do this.

3.5 A need achievement perspective It is clear from the review of literature that students motivation can broadly vary along two dimensions: The motive to avoid failure and the motive to approach success. These motivations are best reflected in the need achievement theory presented by Atkinson (1957, 1958, 1964; see also McClelland, 1965). An issue foreshadowed here but expanded further in the final empirical chapter (Chapter 12) concerns this need achievement model and the way in which, self-handicapping, defensive expectations, and reflectivity can be represented in terms of the motives to approach success and avoid failure.

A useful framework under which to explore this issue is the quadripolar model of need achievement developed by Covington and Omelich (1991; see also Covington, 1992, 1997). This model represents need achievement in two-dimensional space such that students are dually mapped in terms of the motive to succeed and the motive to avoid failure. Covington and Omelich have further identified student typologies using this quadripolar model and it is later discussed how particular typologies are broadly congruent with reflectivity, defensive expectations, and self-handicapping. The synthesis of key strategies from a need achievement perspective is the final empirical issue dealt with in the thesis and is suggested to be a feasible means of integrating the diversity of findings generated in the study.

3.6 Chapter summary The present chapter has outlined the principles of the self-worth theory of motivation. It is proposed that this theory provides a conceptual base from which to select a variety of affective and motivational factors proposed to predict self-protection strategies. These factors have been examined not only in terms of how they directly or indirectly relate to a self-worth motivation but also in terms of how they are proposed to predict self-handicapping and defensive expectations. A series of cumulative 111

models has been presented which derives from the conceptual review of each factor. Also arising from this conceptual review is the proposition of a higher order factor structure underpinning these factors and this possibility is explored in the empirical components of the study. In addition to the possibility of a higher order structure, the issue of domain specificity is addressed. Whilst the academic arena is the general area within which the study is located, the extent of domain specificity (viz. maths vs verbal) within this arena is also explored. In sum, then, a variety of research issues are pursued in the present study not only in relation to the process model developed throughout the present chapter, but also in relation to supplementary issues (such as interaction effects and a test of the need achievement theory of motivation) addressed throughout the review. The specific nature of these research issues and related hypotheses are now discussed.

112

4. SUMMARY OF RESEARCH ISSUES AND SPECIFIC FOCI OF THE PRESENT STUDY

4.1 Quantitative analyses The central research issues revolve around the hypothesised process model (see Figure 3.7). This model is one in which a variety of motivational and affective antecedents are proposed to predict self-protection strategies in the forms of self-handicapping, defensive expectations, and reflectivity. In turn, self-handicapping, defensive expectations, and reflectivity are proposed to predict academic outcomes in the forms of self-regulation, persistence, future academic plans (reported at Time 1), and academic grades (reported at Time 2). This process model is assessed separately at two time points - students first and second years at university - as well as in a longitudinal form which connects the two process models across time.

4.1.1 Factors predicting the strategies In terms of the motivational and affective factors, the following predictions are advanced: Performance-orientation (comprising public self-consciousness, ego-orientation, avoidance-oriented performance, and esteem-relevant competence valuation) is positively associated with self-handicapping and defensive expectations; Task-orientation is negatively associated with self-handicapping and defensive expectations; An entity view of intelligence is positively associated with self-handicapping and defensive expectations whereas an incremental view is negatively associated with them; Uncertain personal control (comprising unstable self-concept, low control over future outcomes, and ability attributions for failure) is positively associated with self-handicapping and defensive expectations; An external attributional orientation is positively associated with self-handicapping and to a lesser extent, defensive expectations; and,

113

 Level of self-concept is negatively associated with defensive expectations (however, no prediction is made in relation to self-handicapping).

The research literature on factors that predict reflectivity appears to be non-existent and so analyses assessing relationships between affective and motivational predictors and reflectivity are exploratory.

4.1.2 Outcomes following from the strategies The second component of the process model is one in which self-handicapping, defensive expectations, and reflectivity predict various academic outcomes. At Time 1 these outcomes include self-regulation, persistence, future academic plans, and at Time 2 also include academic grades. In relation to these strategies, the following predictions are advanced: Self-handicapping is negatively associated with self-regulation, persistence, future plans, and grades; Defensive expectations is negatively associated with self-regulation, persistence, future plans, and grades; and, Reflectivity is positively associated with self-regulation, persistence, future plans, and grades.

4.1.3 Supplementary quantitative issues In addition to an examination of the proposed process model, other, more supplementary research issues are assessed separately. These include the domain specificity of constructs, the higher order factor structure underpinning the motivational and affective predictors, the distinction between defensive expectations and reflectivity (through tests of their interaction and the extent to which defensive expectations predict reflectivity), the relationship between active and self-presented dimensions of self-handicapping and defensive expectations, and a need achievement representation of self-handicapping, defensive expectations, and reflectivity.

114

4.2 Qualitative analyses The substantive issues are considered to also lend themselves quite well to qualitative methods. At both Times 1 and 2, selected students are interviewed with a view to not only illuminating key quantitative findings but also to extending the quantitative data. At Time 1, students high or low in self-handicapping or defensive expectations are interviewed. These interviews seek to shed further light on the key findings derived in the Time 1 quantitative process model. Time 2 qualitative analyses utilise the longitudinal nature of the data such that students evincing large shifts (upwards or downwards) in self-handicapping or defensive expectations are selected for in-depth interviews. These data are intended to explore concomitant shifts in the affective and motivational predictors and to identify, from students own perspectives, reasons why such shifts occur. Importantly, both qualitative studies are not intended to test the quantitative data and no hypotheses are made in relation to them. Rather, they are designed to illuminate the derived quantitative data, to access issues that are difficult to access using large-sample quantitative methods, and to ground the quantitative data in the context of students day-to-day academic lives.

115

5. METHOD

This chapter presents an overview of the institutions from which the samples were drawn, an outline of the survey procedure relating to the quantitative analyses, a description of the materials used in the questionnaire, and a brief orientation to the statistical procedures used to analyse the data. More specific details related to sample composition appear in the following chapter.

5.1 Sample and procedure Respondents were first year teacher education students from three universities in Sydney, Australia. Two of the institutions (University of Western Sydney, Nepean and University of Western Sydney, Macarthur) are network members of the one university (University of Western Sydney - UWS) located in Sydneys western suburbs and the third institution (University of Sydney) is located a few kilometres from Sydneys central business district. The University of Sydney is Australias oldest university and is a member of what is referred to as the big eight universities in Australia. Typically, this institution attracts amongst the highest performing school leavers. UWS is a much newer university, established in the late 1980s. These institutions offer undergraduate programs in Primary, Early Childhood, and Secondary teacher education and it was students in these programs who were surveyed.

Questionnaires (see Appendix A and Appendix D) were administered to students during lecture time. Students were briefly oriented to the broad aims of the study, but were not informed about the specific issues of interest to the researcher. Students were informed that the researcher was interested in how they go about their studies with a view to assisting educators in promoting student learning and motivation. The background questions on the instrument were worked through by the researcher with the group. Following this, the rating scale was explained to students and a few related example items were also worked through with the group. Students were then asked to complete the questionnaire on their own and to submit the completed form to the researcher at the end of the lecture time.

5.2 Materials 116

Aside from the demographic and background details, items on all subscales were responded to using a 7-point Likert-type rating scale (1 = Strongly disagree; 7 = Strongly agree). After reversal of appropriate items, high scores on items reflected more agreement to the respective item and subscale referents. Tables 5.1 (strategies), 5.2 (affective and motivational predictors), and 5.3 (academic outcomes) present subscale names and example items. Each of these subscales are described in detail below. Items were randomly interspersed throughout the questionnaire rather than presented subscale-by-subscale.

Table 5.1 Sample items from strategy subscales

Subscale
Active self-handicapping

Example item
I often fool around the night before a test or exam so I have an excuse if I dont do as well as I hoped

Self-presented self-handicapping

I tell others that I fool around the night before an exam more than I really do, so if I dont do as well as I had hoped, I can say that is the reason

Self-presented affective selfhandicapping

When an exam or assignment is coming up, I am inclined to tell others that Im more anxious than I really am, so if I dont do as well as I had hoped, they will think that is the reason

Defensive expectations

No matter how well I have done in the past, I go into academic situations expecting to do worse

Self-presented defensive expectations

No matter how well I have done in the past, when going into academic situations, I let others think that I expect to do worse than I really think Ill do

Reflectivity

I carefully consider all possible outcomes before tests and assignments

117

Table 5.2 Sample items from affective and motivational predictor subscales Subscale
Esteem-relevant competence valuation Esteem-relevant incompetence avoidance Public esteem-relevant competence valuation Public esteem-relevant incompetence avoidance Ability attributions for failure Ability attributions for success Effort attributions for failure Effort attributions for success External attributions for failure External attributions for success Academic public self-consciousness Ego-orientation Task/mastery orientation Avoidance-oriented performance Entity view of intelligence Incremental view of intelligence Low future control over success Low future control over avoiding failure Level of self-concept Stability of self-concept

Example item
Feeling I am competent is very important to how I feel about myself It is very important to how I feel about myself that I dont feel incompetent Others thinking that I am competent is very important to how I feel about myself It is very important to how I feel about myself that others dont think that Im incompetent Not doing so well in a course is because I lacked skill in that area The most important ingredient in getting good grades is my academic ability When I receive a poor grade, I usually feel that the main reason is that I havent studied enough Whenever I receive good grades, it is because I have studied hard Often I get poorer grades in courses because the teacher has failed to make them interesting I get good grades mostly because the course material was easy to learn I care a lot about how I present myself academically I feel really successful when I know more than other people I feel really successful when what I learn really makes sense Often the main reason I do my uni work is because I do not want to do poorly on tests and assignments There isnt much some students can do to make themselves smarter A student who works really hard could be one of the smartest in the class When I do well Im unsure as to how to repeat that success When I dont do well Im unsure about what to do to avoid that happening again I learn quickly in these subjects My opinion of myself tends to change a good deal instead of always remaining the same

118

Table 5.3 Sample items from academic outcome subscales Subscale

Self-regulation

Example item
Before taking an exam or quiz, I plan out how I will study the material

Persistence

If I have trouble understanding a problem or task, I keep going over it until I understand it

Future academic plans/intentions

I dont mind doing subjects in this area in my further education

5.2.1 Self-handicapping 5.2.1.1 Active self-handicapping The active self-handicapping subscale items (e.g., I often fool around the night before a test or exam so I have an excuse if I dont do as well as I hoped) were adapted primarily from the Academic Self-Handicapping Scale (Midgley et al., 1996) and, to a lesser extent, the Shortened Self-handicapping Scale (Strube, 1986). The complete adapted subscale is presented in Appendix B1 along with statistics pertaining to the psychometrics of the subscale which are further discussed below. For inclusion in the active self-handicapping subscale, items needed to reflect selfhandicapping as an active strategy aimed at establishing an obstacle before a performance scenario. The majority of items were drawn from the Midgley et al. (1996) scale which has been refined over the past few years to more explicitly address the active self-handicapping issue. One feature of the Midgley et al. scale is the trailer . . . so that if I dont do as well as I had hoped, I can say that is the reason which asks respondents about the intentions underlying the actual behaviour. This trailer was included so that the motivation for the self-handicapping behaviour was clear and broadly consistent with the most recent academic self-handicapping scale. Moreover, this trailer explicitly assesses the protective self-handicapping dimension. Several further items were included which were broadly consistent with the Midgley et al. conceptualisation and aimed at assessing the respondents active attempts to establish an alibi for failure before the fact. Only two items were drawn and adapted from

119

Strubes (1986) scale, because, upon closer scrutiny, it seemed as though many items in this scale did not reflect active self-handicapping. For example, an item in Strubes scale addresses the use of excuses after failure. Such an item does not reflect active self-handicapping which is an a priori strategy and not one which occurs after the event.

5.2.1.2 Self-presented self-handicapping Self-presented self-handicapping refers to reports of self-handicapping behaviours which either do not occur or are exaggerated. Thus the individual does not actively impede the chances of success but manages to negotiate the situation before the fact such that there appears to be an obstacle in place obviating success. The self-presented self-handicapping scale is parallel to the active self-handicapping scale presented above but emphasises the self-presented and/or public nature of the strategy. The items (e.g., I tell others that I fool around the night before an exam more than I really do, so if I dont do as well as I had hoped, I can say that is the reason) appear in Appendix B1.

5.2.1.3 Self-presented affective self-handicapping The third dimension of self-handicapping assessed in the Time 1 questionnaire relates to self-presented affective self-handicapping. There has been some discussion in the literature about the exaggeration of affective states that might serve as an excuse for upcoming failure (see Covington, 1984a, 1992; Snyder & Smith, 1982). To date, it appears that this has not been assessed through psychometric scales and thus represents a new operationalisation of self-handicapping. Affective states such as anxiety, depression, or bad moods, can be strategically reported so as to alter the implications of failure. The items (e.g., When an exam or assignment is coming up, I am inclined to tell others that Im more anxious than I really am, so if I dont do as well as I had hoped, they will think that is the reason) are presented in Appendix B1.

120

5.2.2 Defensive pessimism 5.2.2.1 Active defensive expectations The defensive expectations scale developed here reflects the concepts of Norem, Cantor, and colleagues who have contended that defensive pessimism is comprised of defensive expectations and reflectivity. The original defensive expectations scale required respondents to first rate how well they had performed on previous assessment tasks (a self-concept-like screener item) and then to report expectations for future performance. Those scoring higher on the screener and lower in terms of expectations were said to be reporting defensively negative expectations.

One concern with this approach is the assumption that only those who score relatively higher on the screener can be classified as having defensively negative expectations, whereas it is conceivable that students at any level of achievement can hold unrealistically negative expectations. Another concern with typically administered defensive pessimism scales is that they are used primarily to assign study participants to experimental groups and thus are not particularly conducive to correlational analyses. Because correlational techniques are the central focus of the present study, some adaptation of the existing defensive expectations measure was seen to be in order. This involved integrating the screener into each of the proposed expectation items in a way that assessed defensively negative expectations for students at all levels of achievement. Accordingly, the introduction to each item is: No matter how well I have done in the past . . .. These items have been drawn from two versions of the Defensive Pessimism Questionnaire devised by Norem and Cantor as well as the Life Orientation Test devised by Scheier, Carver, and Bridges (1994) which specifically addresses pessimism and negative expectations. Some rewording of the items was carried out that assisted readability and consistency. The adapted items (e.g., No matter how well I have done in the past, I go into academic situations expecting to do worse) appear in Appendix B2. Importantly, whilst this new operationalisation extends existing research into defensive expectations, it does represent a departure from previous operationalisations and so results to follow must be considered in this context.

5.2.2.2 Self-presented defensive expectations 121

The self-presented defensive expectations items are parallel to the active defensive expectations items but adapted to reflect individuals tendencies to report that they are feeling more negative about upcoming performances than they actually are. These items (e.g., No matter how well I have done in the past, when going into academic situations, I let others think that I expect to do worse than I really think Ill do) are also presented in Appendix B2.

5.2.2.3 Reflectivity The second dimension of defensive pessimism is reflectivity. This refers to the thinking-through process that accompanies the negative expectations. Through factor analysis, Norem (personal communication, December 12, 1996) has identified a reflectivity subscale comprised of reflectivity-related items (e.g., I carefully consider all possible outcomes before tests and assignments) which loaded most clearly on the one factor. This subscale is used in the present questionnaire as originally formulated (see Appendix B2).

5.2.3 Affective and motivational predictors 5.2.3.1 Esteem-relevant competence valuation Self-worth motivation theory suggests that individuals feelings about themselves are dependent on feeling competent (here referred to as esteem-relevant competence valuation) and avoiding feeling incompetent (protective esteem-relevant competence valuation). There may also be public and private dimensions to such esteem-relevant competence valuation. Private competence valuation refers to the need for individuals to feel competent within themselves (e.g., Feeling I am competent is very important to how I feel about myself) or to avoid feeling incompetent (e.g., It is very important to how I feel about myself that I dont feel incompetent). Public competence valuation refers to individuals need to be seen as competent (e.g., Others thinking that I am competent is very important to how I feel about myself) or avoid being seen as incompetent (e.g., It is very important to how I feel about myself that others dont think that Im incompetent). Thus there can be four dimensions to esteem-relevant competence valuation and subscales addressing these four dimensions are presented in Appendix B3.

122

5.2.3.2 Attributions Attributions refer to the perceived factors that have contributed to or caused previous outcomes. The attribution scale used in the present study is based on the Lefcourt, Von Baeyer, Ware, and Cox (1979) Multidimensional Multiattributional Causality Scale (MMCS). The MMCS can be broadly divided into internal and external dimensions that contribute to success or failure and within each of these two dimensions are a further two components. Within the internal dimension are ability and effort attributions. Within the external dimension are luck and context dimensions. Ability attributions refer to the extent to which individuals see their ability or competence as the cause of success (e.g., The most important ingredient in getting good grades is my academic ability) and failure (e.g., Not doing so well in a course is because I lacked skill in that area). Effort attributions refer to the extent to which individuals see their effort and persistence as the cause of success (e.g., Whenever I receive good grades, it is because I have studied hard) and failure (e.g., When I receive a poor grade, I usually feel that the main reason is that I havent studied enough). Luck attributions refer to the extent to which luck or chance factors are perceived to be the cause of success (e.g., Sometimes my success on exams depends on luck) and failure (e.g., Some of my lower grades have seemed to be due to bad luck). Context attributions are those for which factors such as teachers, marking standards, or content difficulty are seen to cause success (e.g., I get good grades mostly because the course material was easy to learn) or failure (e.g., Often I get poorer grades in courses because the teacher has failed to make them interesting). The Lefcourt et al. scale contains three items per subscale. It was considered that item pools for each subscale in the present study be comprised of a minimum of six items. Accordingly, a further three items were developed for each of the ability and effort subscales. The two external subscales (luck and context) were aggregated and thus comprised the requisite six items in total. The adapted subscales appear in Appendix B4.

123

5.2.3.3 Academic public self-consciousness Public self-consciousness has typically been assessed using two instruments (Fenigstein Scheier, & Buss, 1979; Scheier & Carver, 1985). It is the Scheier and Carver (1985) instrument (a revision of the Fenigstein et al. scale) which forms the basis of the academic public self-consciousness scale used in the present study. Public self-consciousness refers to the extent to which individuals are concerned about how they are evaluated and viewed by others. The public self-consciousness dimension of the Scheier and Carver instrument was adapted such that it reflected a concern on the part of students about how their academic performance is viewed by others. This adaptation involved transforming the original items to reflect concern with the academic image one projects to others and concern with how one is being viewed academically (e.g., I care a lot about how I present myself academically). The adapted items are presented in Appendix B5.

5.2.3.4 Motivation orientation Respondents motivation orientation was assessed using a shortened form of the Motivation Orientation Scale (Nicholls, 1989) together with another subscale referred to as avoidance-oriented performance. The Motivation Orientation Scale comprises ego-orientation and task-orientation. The avoidance-oriented performance subscale is adapted from an orientation outlined by Harter, Whitesell, and Kowalski (19928). The ego-orientation subscale contains items that reflect the extent to which students are motivated to outperform others in academic contexts (e.g., I feel really successful when I know more than other people). Task-orientation items assess the extent to which a student is motivated by mastery rather than competitive concerns (e.g., I feel really successful when what I learn really makes sense). The avoidance-oriented performance subscale reflects a tendency to be motivated to avoid poor performance and its negative implications (e.g., Often the main reason I do my uni work is because I do not want to do poorly on tests and assignments). All three subscales are presented in Appendix B6.

Subsequent to the present studys Time 1 data collection, this issue has been given detailed attention (Elliot & Church, 1997; Elliot & Harackiewicz, 1996; Middleton & Midgley, 1997; Midgley et al., 1998; Skaalvik, 1997a, 1997b).

124

5.2.3.5 Views of intelligence Views of intelligence can take two broad forms: entity and incremental. An entity view of intelligence refers to that which holds intelligence as something fixed and unchangeable (e.g., There isnt much some students can do to make themselves smarter). Six items in the entity subscale were adapted from Stipek and Gralinski (1996). Importantly, however, whereas the Stipek and Gralinski subscale reflected more of an attributional profile, the items included in the present questionnaire were adapted somewhat to more closely reflect the immutability of intelligence. Individuals with an incremental view of intelligence hold the belief that ones intelligence can change, particularly through the application or withdrawal of effort (e.g., A student who works really hard could be one of the smartest in the class). Stipek and Gralinski originally proposed three items to assess the incremental scale and these were supplemented with a further three items. The entire set of items appears in Appendix B7.

5.2.3.6 Perceptions of future control Items to assess respondents perceptions of control over future success and failure were adapted from Connells (1985) Multidimensional Measure of Childrens Perceptions of Control. Connells scale includes Unknown control, Powerful others control, and Internal control within both cognitive and social domains. It is the Unknown control scale in the cognitive dimension which was adapted for use in the present study and which was reworked to reflect perceptions of future control rather than the causal ascriptions for past outcomes. Also, success and failure items were separated to form two subscales and a further four items were added to each dimension. Thus, six items reflected low perceived control over future success (e.g., When I do well Im unsure as to how to repeat that success) and six items reflected perceptions of low control over future failure (e.g., When I dont do well Im unsure about what to do to avoid that happening again). All items are presented in Appendix B8.

125

5.2.3.7 Level of self-concept Level of self-concept was assessed using the Academic Self Description Questionnaire II (ASDQ; Marsh, 1988a). This instrument is designed to measure the

(multidimensional) self-concept of students. The full ASDQ II is comprised of 136 items that measure a variety of academic domains including maths and verbal domains which are the domains assessed in the present study. Importantly, the items (e.g., I learn quickly in these subjects) which assess maths and verbal self-concepts have the same stems and so these items were easily adapted from the original items for inclusion in the present questionnaire. The six self-concept items are presented in Appendix B9.

5.2.3.8 Self-concept stability Stability of self-concept was primarily assessed using the Rosenberg (1965) Stability of Self Scale. This scale addresses the extent to which an individuals self-concept changes over time (e.g., My opinion of myself tends to change a good deal instead of always remaining the same). Importantly, this variability in self-concept is more barometric, reflecting short-term shifts rather than changes in self-concept over the longer term. The present study also incorporated two items from the Self-concept Certainty Scale (Campbell et al., 1996). Thus, seven items comprised this scale (five of which represented the entire Rosenberg Scale and two derived from Campbell et al.) and these appear in Appendix B10.

5.2.4 Academic outcomes 5.2.4.1 Self-regulation Self-regulation items were drawn from the Cognitive Engagement Scale (Miller, Greene, Montalvo, Ravindran, & Nichols, 1996). These items address the extent to which students are involved in organising and planning their study, setting goals, and monitoring their progress (e.g., Before taking an exam or quiz, I plan out how I will study the material). The original scale comprised nine items and an item pertaining to planning for assignments was also added. Minor adaptation of the nine items was undertaken with a view to tightening the self-regulatory focus. The final ten items are included in Appendix B11.

126

5.2.4.2 Persistence The persistence subscale was also drawn from the Miller et al. (1996) Cognitive Engagement Scale and these items reflected the extent to which a student would spend time and invest further effort in the face of challenge and difficulty (e.g., If I have trouble understanding a problem or task, I keep going over it until I understand it). Some of the original items were modified primarily because they reflected helpseeking, and in some cases, actually implied cheating. Whilst the theme of the items was retained, they were adapted such that a sharper focus was placed upon the persistence component. The final form of the subscale is presented in Appendix B12.

5.2.4.3 Future academic intentions Students were asked to report on their intentions and willingness to engage in the target subject areas in the future. The future plans subscale was adapted from a Skaalvik and Rankin (1995) scale which comprised seven declarative statements about individuals willingness to engage in a given subject area in their further education or occupations (e.g., I dont mind doing subjects in this area in my further education). These items were modified so as to be more consistent with a higher education climate and to integrate with the two subject area domains in the questionnaire. The seven items comprising future plans are presented in Appendix B13.

5.3 Data analysis Data were analysed using SPSS for Windows and LISREL 7.2 (Joreskog & Sorbom, 1989a, 1989b). The majority of analyses entailed confirmatory factor analysis (CFA) and structural equation modelling (SEM) using LISREL which tests the fit between a sample covariance matrix and an hypothesised matrix. Maximum likelihood was the method of estimation used for the models. Because of the large number of parameters to be estimated in the central models and the large sample size required to enhance the reliability of these estimates, pairwise deletion of missing data was conducted. The raw data were used as input to PRELIS (Joreskog & Sorbom, 1988) and a covariance matrix was produced which was subsequently analysed using LISREL. In terms of goodness of fit indices, the Tucker Lewis Index (TLI) is emphasised as simulation studies have shown that it is relatively independent of sample size and also imposes an appropriate penalty for inclusion of additional variables in a given model (Marsh, 127

Balla, & Hau, 1996; McDonald & Marsh, 1990). Following Marsh et al. and McDonald and Marsh, the Relative Noncentrality Index (RNI) and Root Mean Square Error of Approximation (RMSEA) are also emphasised as measures of goodness of fit. TLI and RNI values at or above .90 and RMSEA values below .05 are typically considered to indicate an acceptable fit of the model to the data. It has recently been suggested, however, that the .90 criterion may tend to over-reject true models (Hu & Bentler, 1995). Accordingly, TLIs and RNIs in the high .80s are also considered acceptable.

5.3.1 Brief orientation to CFA and SEM According to Byrne (1994), in CFA approaches to factor analysis, the researcher draws on knowledge of the theoretical structure of the variables, proposes the factor structure a priori, and then statistically tests this hypothesised factor structure. In the CFAs performed in the present study, it was hypothesised that (a) each measured variable would have a non-zero loading on the factor it was designed to measure and a zero loading on all other factors, (b) the factors in all cases would be correlated, (c) and the error terms (referred to as uniquenesses) for each measured variable would be uncorrelated (unless stated otherwise). Moreover, using CFA procedures, higher order factors can also be generated in which, for example, a second order factor is hypothesised to underlie the first order factors. This is particularly relevant to the present study because as outlined at the conclusion of the literature review, higher order factors are hypothesised to underpin a variety of first order factors. Figure 5.1 shows a CFA structure in which there exists one higher order factor and three first order factors for which there are three indicators each. SEM refers to the causal relationships between the latent factors generated in the CFAs (see Figure 5.2). In SEM, the causal processes proposed in the study are typically estimated through a series of regression equations and these relationships are often represented pictorially. This model can be tested in a one-step (simultaneous) analysis and the goodness of fit indices can be assessed to determine the test of the fit between the model and the data.

128

Item 1

Item 2

Factor 1 (First order factor)

Item 3 Item 4

Item 5

Factor 2 (First order factor)

Higher/second order factor

Item 6 Item 7

Item 8

Factor 3 (First order factor)

Item 9

Figure 5.1 Representation of higher order factor structure

Item 1

Item 4

Item 2

Factor 1

Factor 2

Item 5

Item 3

Item 6

CFA/Measurement component Structural component Figure 5.2

CFA/Measurement component

CFA/Measurement component and structural component of SEM

129

6. TIME 1 QUANTITATIVE ANALYSES

Results in this chapter are presented in four sections. The first section (Section 6.1) presents results derived from analyses of data obtained from the first sample (Nepean) using the initial form of the instrument. The central objective of this component was to refine the content domain, reduce the number of items and subscales, and to provide confirmatory factor analysis (CFA) findings based on the reduced item and subscale pool. This refinement was seen to be important not only in terms of retaining items and subscales which were strongest from a measurement perspective but also to reduce the demands placed on respondents in subsequent administrations of the instrument. The second section (Section 6.2) of results presents findings of CFAs of data collected from a further two samples (Macarthur and Sydney) using the refined instrument. Tests for factor invariance across sex and institution were also conducted and on the basis of these results, data were pooled and CFA results for the entire sample of 584 students presented. Section three (Section 6.3) presents findings of higher order factor analyses as well as an assessment of the role of background variables in the context of these higher order factors. The final section (Section 6.4) incorporates the higher order factor structure into a full structural equation model which assesses the complete Time 1 process model. Also included in this final section are tests of (a) the interaction between defensive expectations and reflectivity, (b) the extent to which defensive expectations predict reflectivity, (c) the original operationalisation of defensive pessimism, and (d) the interaction of level and stability of self-concept.

6.1 Refining the instrument The first phase of data collection was the more exploratory phase of the study because the instrument at this time included quite a number of subscales that were intended to be refined following initial psychometric assessment. Accordingly, this component of data analysis involved determining which items and subscales to retain in subsequent questionnaire administration. This first involved exploring the utility of the maths/verbal domain distinction in the questionnaire. Secondly, item analyses were performed using a series of one factor congeneric models and poorly loading items 130

were dropped from subsequent analyses. Thirdly, having removed items on this basis, CFA was performed using all 29 subscales in the analysis. On the strength of this CFA, inter-subscale correlations were inspected and highly correlated subscales were eliminated from the instrument. Having removed items and refined the subscales, a final CFA for the first sample was conducted and factor loadings and factor correlations presented.

6.1.1 Method Respondents in the first phase of the study were 204 first year teacher education undergraduates from UWS Nepean. Most respondents (n=186, 91.2%) were female, 18 were male (8.8%). Most (n=170, 83.3%) were enrolled in a Primary education program and 34 (16.7%) were enrolled in an Early Childhood program. The mean age of respondents was 21.7 years (SD=5.9). For the majority of respondents (n=181), this semester was their first at university. The average Tertiary Entrance Rank (TER) was 58.90/100 (SD=16.02).

6.1.2 Refining the content domain The first task was to explore the degree to which math and verbal domains were distinct. As discussed in Chapter 3, little work has been conducted exploring the content specificity of a number of the constructs used in the present study. This issue was tested using the following approaches: (a) correlating raw subscale scores (the mean of the set of items in each subscale), (b) correlating raw subscale scores and also correcting for reliability, (c) performing two factor (math vs verbal) CFAs for each of the 29 subscales, and (d) performing the same CFAs in addition to correlating parallel uniquenesses which attenuates the upward bias in correlations between factors using parallel wording (see Marsh, 1993b; Marsh, Roche, Pajares, & Miller, 1997). Results of these operations are presented in Table 6.1

131

Table 6.1 Cronbachs alpha coefficients and correlations between math and verbal domains
Reliability Cronbachs : Cronbachs : maths verbal Ability - failure Effort - failure Ability - success Effort - success External - failure External - success Entity view Incremental view Ego-orientation Task-orientation Control future failure Control future success Avoidance orientation Pub self-conscious Private comp enhance Public comp enhance Private comp protect Public comp protect Level of self-concept Instability self-concept Active self-handicap Self-pres self-handicap Affective self-handicap Defensive expectations Self-pres def expectns Reflectivity Persistence Self-regulation Future plans
.75 .82 .77 .78 .66 .73 .85 .79 .91 .85 .88 .90 .86 .92 .89 .93 .79 .88 .89 .75 .93 .93 .88 .92 .92 .69 .76 .84 .91 .76 .80 .76 .82 .70 .76 .85 .78 .91 .85 .88 .91 .86 .93 .90 .93 .78 .88 .87 .75 .93 .92 .88 .91 .92 .67 .80 .85 .86

Raw
.82 .90 .81 .85 .88 .93 .95 .97 .96 .89 .96 .91 .97 .97 .97 .99 .97 .99 -.03 .81 .99 .99 .99 .93 .99 .94 .80 .93 .31

Raw corrected for reliability

1.09 1.11 1.06 1.06 1.29 1.25 1.12 1.24 1.05 1.05 1.09 1.22 1.13 1.05 1.08 1.06 1.22 1.13 -.03 1.08 1.06 1.07 1.13 1.02 1.08 1.38 1.02 1.10 .35

Correlations CFA
1.09 1.14 1.06 .98 1.33 1.30 1.13 1.23 1.09 1.09 1.09 1.01 1.14 1.05 1.09 1.07 .79 1.10 -.07 1.05 1.07 1.07 1.11 1.02 1.06 1.26 1.05 1.10 .36

CFA, parallel uniqueness correlated

.83 .90 .81 .82 .93 1.00 .99 .95 .99 .96 .97 .92 .99 .99 .98 .99 .98 .99 -.11 .78 .99 1.00 1.00 .95 1.00 .97 .77 .95 .28

132

Results presented in Table 6.1 indicate that for the most part the subscales are domain general in the sense that they are highly correlated. Given the format of the questionnaire, it was of concern that this may have resulted from some sort of method bias such that a response in the verbal column was identical to the response given in the math column. A closer inspection of the correlations in Table 6.1, however, reveals that there were some subscales for which this method bias was not apparent. Particular attention is directed to the self-concept subscale. Here the correlation is around zero and this conforms to a large body of research showing that there is little or no correlation between math and verbal self-concept (see Marsh, 1990, for a review). Thus, where there is a known domain specificity, correlations reflected this. Moreover, this provides strong evidence that the method-bias possibility can be rejected and on this basis, it seems that the other dimensions in the model can be argued to be domain general.

These analyses also hold statistical implications for research involving parallel questionnaire items (e.g., self-other agreement, longitudinal, or domain-specificity studies). In most cases, CFA applications assume that residual variance (uniqueness) associated with each observed variable is relatively independent of uniqueness of other observed variables. In the present case, however, because of the identical wording for both math and verbal domains, it is reasonable to assume that uniqueness associated with the parallel items will be correlated and if these correlations are not estimated in the model, the correlations between the two factors will be positively biased (much the same as scale score correlations corrected for reliability; see Marsh, 1993b; Marsh et al., 1997). Indeed, this is seen in Table 6.1 in which correlations between factors are, in a number of cases, not only extremely high, but improper. Accordingly, when the parallel uniqueness is estimated and the potential bias controlled, these correlations are attenuated to a proper value.

Another means by which specificity can be assessed is through analysis of mean differences between math and verbal constructs. Accordingly, a series of paired samples t-tests was performed. Means, standard deviations, and paired t values appear in Table 6.2. 133

Table 6.2 Means, SDs, and t values for math/verbal subscale pairs

Math M (SD) Ability - failure Effort - failure Ability - success Effort - success External - failure External - success Entity view Incremental view Ego-orientation Task-orientation Control future failure Control future success Avoidance orientation Pub self-conscious Private comp enhance Public comp enhance Private comp protect Public comp protect Level of self-concept Instability self-concept Active self-handicap Self-pres self-handicap Affective self-handicap Defensive expectations Self-pres def expectns Reflectivity Persistence Self-regulation Future plans
3.89 4.85 4.62 5.36 3.56 3.82 3.05 4.92 4.54 5.65 3.96 4.08 4.33 4.48 4.99 4.26 5.08 4.54 4.14 4.23 2.29 2.52 2.63 4.06 3.36 4.39 4.78 5.10 4.70 (0.97) (1.10) (0.80) (0.82) (0.85) (0.93) (1.10) (0.95) (1.20) (0.83) (1.20) (1.20) (1.10) (1.20) (1.10) (1.40) (1.10) (1.40) (1.20) (0.92) (0.94) (1.00) (1.10) (1.30) (1.20) (0.83) (0.92) (0.87) (1.30)

Verbal M (SD)
3.86 4.81 4.66 5.37 3.64 3.81 3.05 4.92 4.55 5.68 3.97 4.05 4.35 4.50 5.00 4.28 5.10 4.57 4.55 4.22 2.28 2.50 2.62 3.99 3.34 4.41 4.84 5.16 5.10 (0.99) (1.10) (0.80) (0.87) (0.90) (0.98) (1.10) (0.94) (1.20) (0.83) (1.20) (1.30) (1.10) (1.20) (1.10) (1.40) (1.10) (1.40) (1.10) (0.91) (0.94) (1.00) (1.10) (1.30) (1.30) (0.81) (0.94) (0.83) (0.98)

t (df)
.73 .98 1.11 .14 2.70 .38 .18 .01 .42 .91 .42 .85 .52 1.70 1.00 .52 .81 1.30 3.50 .13 1.60 1.80 .87 1.90 2.10 1.10 1.30 3.00 4.10 (203) (203) (203) (203) (203) (203) (203) (203) (203) (203) (203) (203) (203) (203) (203) (202) (203) (202) (202) (203) (201) (202) (201) (203) (202) (203) (203) (203) (202)

p
.47 .33 .27 .89 .01 .70 .86 .99 .68 .36 .67 .40 .60 .10 .32 .60 .42 .21 .01 .90 .11 .07 .39 .06 .03 .29 .20 .01 .01

Consistent with correlational findings, self-concept and future plans evinced the most significant mean differences between math and verbal subscales. Other significant mean differences occurred for external attributions for failure, self-presented defensive expectations, and self-regulation. These differences, however, were not substantial and for the most part were a function of the relatively large sample size and the large number of tests involved. It can be argued, then, that consistent with correlational findings, the bulk of the subscales are domain general.

In the light of the pervasive domain generality, it was considered that subsequent analyses incorporating both math and verbal domains would not prove substantially more enlightening than analyses in just one of these two domains. It was therefore

134

decided that subsequent analyses would focus on only one of these domains. For two reasons, the math domain was selected as the one to retain. First, there is a good deal of recent data relating to mathematics anxiety amongst university students (e.g., Bessant, 1995; Pajares & Urdan, 1996; Schneider & Nevid, 1993; Vance & Watson, 1994) and this may be particularly relevant to the present study of self-protection. Second, the author has previously examined constructs incorporated in the present study in relation to mathematics (see Martin & Debus, 1998) and so present analyses can be seen as somewhat ongoing from the previous study. All subsequent analyses, then, are based on responses to the mathematics component of the instrument.

6.1.3 Refining the number of items Having selected a content domain on which to focus, the next stage in data analysis was to work through the instrument subscale-by-subscale with a view to removing items that were not loading particularly well on hypothesised subscales. This process involved examining results from (a) a series of one-factor congeneric CFA models, (b) item-total correlations, and (c) Cronbachs alpha coefficients given the removal of each item from the subscale. These statistics are presented in Appendix B for each of the 29 subscales in the instrument. An illustrative example of these tables is presented in Table 6.3 in relation to the active self-handicapping scale. Also presented is a detailed rationale for retention and deletion of items in this scale which served as a template for retention and deletion of items in each of the other subscales.

135

Table 6.3 Item analysis for active self-handicapping scale Item

I let myself get run-down (dont look after myself) when assignments or exams are due so I have an excuse if I dont do as well as I hoped I tend to not try hard at assignments so I have an excuse if I dont do as well as I hoped I often fool around the night before a test or exam so I have an excuse if I dont do as well as I hoped I let my friends keep me from paying attention in tutorials or from doing my study or assignments so I have an excuse if I dont do as well as I hoped I tend to not study very hard before exams so I have an excuse if I dont do as well as I hoped I often become occupied on things other than study the night before an exam, so I have an excuse if I dont do as well as I hoped I purposely dont get enough sleep or rest before upcoming exams and assignments, so I have an excuse if I dont do as well as I hoped I dont attend all the classes in this area so I have an excuse if I dont do as well as I hoped *I tend to put assignments and study off until the last moment so I have an excuse if I dont do as well as I hoped *I often think about other things, allow myself to be distracted, or daydream when I try to study so I have an excuse if I dont do as well as I hoped *I often get involved in a lot of activities when exams or assignments are coming up so I have an excuse if I dont do as well as I hoped

Item-total correlation
.77 .75 .73 .65 .66 .79 .77 .76 .64 .58 .70

Alpha with item removed

.91 .91 .91 .91 .92 .90 .91 .91 .92 .93 .92

1 Factor CFA loading

.79 .77 .73 .70 .68 .81 .78 .79 .68 .59 .73

* Item removed following first CFA Note. Coefficient of determination .92; Cronbachs alpha .92; Chi square = 61.97, DF = 20; TLI = .94 RNI = .96

136

The active self-handicapping subscale comprised 11 items. Following a one factor congeneric CFA, three items were removed from the subscale. These items appear in Table 6.3 (denoted by *) along with their associated statistics. The first basis upon which these items were dropped from the final subscale was the enhanced reliability given their removal (whereas removal of most other items would attenuate the reliability). A second basis was the conceptual overlap between items. For example, while one of the items (I often get involved in a lot of activities when exams or assignments are coming up so I have an excuse if I dont do as well as I hoped) dropped from the subscale did have a reasonably high factor loading and item-total correlation, it was similar to another item in the subscale (I often become occupied on things other study the night before an exam, so I have an excuse if I dont do as well as I hoped) and therefore considered conceptually redundant. Moreover, the modification index (which provides information about the relationships between aspects of the model - such as between uniquenesses or between items - which were not freed to be estimated - Joreskog & Sorbom, 1989a, 1989b) for the uniqueness of these two items derived from the CFA output suggested a good degree of empirical overlap. A third rationale for deletion of these items (and items in other subscales) was based on the fact that this subscale was a strong one with a high alpha value and coefficient of determination (COD - an indicant of reliability - Joreskog & Sorbom, 1989a, 1989b) and so reducing the subscale would not compromise its psychometric integrity. Indeed, when these items were removed, the alpha and COD were still high.

A final basis upon which items were removed was through their cross loadings with other subscales. To generate cross loadings, CFAs were performed on (a) the set of predictors, (b) the set of strategies, and (c) the set of academic outcomes. In these three sets of analyses, modification indices were requested and the estimated change statistics were examined. No systematic rule of thumb was used in this phase of the process. Rather, when there was some indecision as to whether to drop or retain an item on the basis of item-total correlations or factor loadings, the cross loadings of that item identified in the modification indices were then assessed as the final criterion for retention or deletion. If the item had substantial cross loadings, then it was dropped from subsequent analyses. As it turned out, while no problematic cross 137

loadings were evident in relation to the self-handicapping subscale, other subscales did present salient cross loadings. For example, one item (I care a lot about how I present myself academically) was dropped from the public self-consciousness subscale because the modification indices indicated rather substantial cross loadings with avoidance-oriented performance (estimated change, .40) and public esteemrelevant competence valuation (estimated change, .39).

Subsequent to assessment of each subscale using these four criteria for retaining or deleting items, a second set of one factor CFAs was performed. The statistics generated in this second CFA for the active self-handicapping subscale are presented in the first part of Table 6.3. Statistics for the other subscales are presented in Appendix B and the rationale for exclusion of items generally follows that which was applied to the active self-handicapping subscale. In some cases, items were not excluded from subscales because either the COD and the Cronbachs alpha for the subscale were not so strong that items could be deleted or because the subscale contained relatively fewer items than other subscales (ideally each latent construct was comprised of at least six items that resulted in a reasonably reliable subscale).

6.1.4 Refining the number of subscales The next set of analyses involved deciding which subscales to retain in subsequent analyses. Of central focus was the empirical overlap of a number of conceptually-alike subscales. Indeed, a number subscales were expected to present high empirical overlap. These included (a) the three self-handicapping constructs, (b) the two defensive expectations constructs, (c) the four esteem-relevant competence valuation constructs, and (d) the two control over future success and failure constructs. Also, the instrument included six attributional dimensions and it was of interest to determine which of these dimensions could be retained and excluded in subsequent analyses.

Addressing these issues entailed CFA of the entire set of subscales in the model (29 subscales in all) with particular focus on the inter-subscale correlations in determining the degree of empirical overlap between constructs. For this and subsequent CFAs, the items in each subscale were aggregated into three to four (depending on the number of items in the subscale) item-pairs such that the first two items in the subscale were 138

assigned to the first pair, the next two to the second pair, and so on. In the context of the present study, this procedure is preferable to factor analysis on the full set of individual items because (a) the ratio of respondents to the number of variables is increased, (b) each variable should have a lower unique component and be more reliable, and (c) the factor loadings are less affected by the idiosyncratic wording of individual items (Marsh & ONiell, 1984). Accordingly, item parcels were computed (by generating the mean of two [in most cases] or three [in a few cases] items in a given subscale) and these item parcels were used as indicators in the CFA.

The 29 factor CFA yielded a chi square of 5687.32 (df=3421) with a Tucker Lewis Index (TLI) and Relative Noncentrality Index (RNI) of .81 and .84 respectively. The Root Mean Square Error of Approximation (RMSEA) was .057. Factor loadings are presented in Appendix C.

6.1.4.1 Assessing and refining self-handicapping and defensive expectations The first task of the subscale refinement process was to consider the central strategies (active self-handicapping, self-presented active self-handicapping, affective selfhandicapping, defensive expectations, and self-presented defensive expectations). This involved assessing the correlations (a) amongst the central strategies themselves and (b) between the central strategies and the other subscales. These correlations, generated by the 29 factor CFA, are presented in Table 6.4.

139

Table 6.4 Correlations among central strategy subscales and both predictor and outcome subscales
Active SH Self-presented SH Affective SH Active SH .92 Self-presented SH .88 .99 Self-pres affective SH .39 .46 .45 DE .56 .67 .68 Self-presented DE .17 .31 .34 Reflectivity .31 .36 .36 Ability - failure .13 .26 .28 Effort - failure .04 .14 .17 Ability - success -.23 -.11 -.03 Effort - success .50 .48 .53 External - failure .44 .51 .47 External - success .39 .36 .37 Entity view -.01 .06 .07 Incremental view .23 .36 .38 Ego-orientation -.28 -.18 -.09 Task-orientation .35 .42 .37 Avoidance Orientation -.21 -.18 -.14 Future plans .26 .34 .31 Control future failure .27 .34 .33 Control future success -.54 -.53 -.40 Persistence .24 .35 .35 Public self-consc .04 .15 .15 Private enhance s-wth .20 .28 .27 Public enhance s-wth .20 .29 .28 Public protect s-wth .05 .09 .12 Private protect s-wth -.05 -.06 -.02 Self-concept .27 .39 .38 Instab self-concept -.36 -.35 -.23 Self-regulation SH=Self-handicapping; DE=Defensive expectations Coefficients > .18 are significant at p<0.05 DE Self-presented DE

.76 .52 .60 .10 -.16 .02 .45 .34 .28 -.07 .27 .03 .49 -.26 .64 .68 -.37 .40 .32 .37 .43 .35 -.28 .52 -.09

.56 .50 .15 .04 .01 .52 .44 .37 -.02 .35 .04 .50 -.15 .51 .61 -.44 .50 .31 .43 .47 .29 -.06 .43 -.06

In terms of the correlations among the strategies, it appears that the area of greatest overlap is between self-presented self-handicapping and self-presented affective selfhandicapping (r=.99). Further inspection of Table 6.4 indicates that the pattern of correlations between these two self-presented self-handicapping subscales and the other subscales is very similar. On the strength of this, it seems that one of these two self-handicapping subscales could be dropped from subsequent analyses. Arguably the most tentative of the self-handicapping subscales is the affect-related subscale in that it reflected a more tangential dimension of self-handicapping. Of primary interest in the study are the actual behaviours in which individuals engage or self-present to

140

protect their self-worth. Accordingly, it was considered defensible to drop the affectrelated dimension of self-handicapping from subsequent analyses. 6.1.4.2 Assessing and refining control over future success and failure Correlations between the control-related measures and central strategies are presented in Table 6.5.

Table 6.5 Correlations among strategy subscales and perceived control over future success and failure subscales

Control over future failure Control future failure .93 Control future success .26 Active SH .34 Self-presented SH .64 DE .51 Self-presented DE SH=Self-handicapping DE=Defensive expectations Coefficients > .18 are significant at p<0.05

Control over future success

.27 .34 .68 .61

These results indicate that the correlation between control over future success and failure is high. Also, the pattern of relationships between both these subscales and the central measures is very similar. It therefore appears that the two subscales are not only conceptually alike but also empirically congruent and on these bases it was considered that one subscale would suffice to capture the hypothesised effects in the study. It was decided to retain the control over future failure subscale for subsequent analyses. The primary rationale for this decision was based on the theoretical perspective taken in the study, namely, that fear of failure is the driving force for selfprotection and that an uncertainty as to whether one can avoid future failure is likely to elicit such a fear. Notwithstanding this, an uncertainty as to whether one can maintain (or repeat) successful performance is also related to a fear of failure, but arguably not so directly. Nevertheless, it is important to note that both dimensions yield a similar empirical profile and on this basis, subsequent findings in relation to control over future failure are also broadly applicable to control over future success.

141

6.1.4.3 Assessing and refining esteem-relevant competence valuation Correlations amongst the competence valuation subscales as well as their correlations with central strategies are presented in Table 6.6. Table 6.6 Correlations among strategy subscales and esteem-relevant competence valuation subscales

Private enhancing Public enhancing Private enhancing .86 Public enhancing .85 .98 Public protective .98 .85 Private protective .04 .20 Active SH .15 .28 Self-presented SH .32 .37 DE .31 .43 Self-presented DE SH=Self-handicapping DE=Defensive expectations Coefficients > .18 are significant at p<0.05

Public protective

Private protective

.95 .20 .29 .43 .47

.05 .09 .35 .29

Results in Table 6.6 indicate that the private dimensions of esteem-relevant competence valuation are most highly correlated as are the two public dimensions. Associations between both private dimensions and the central subscales (selfhandicapping and defensive expectations) yield similar patterns, while associations between both public dimensions and central subscales also yield similar patterns of correlations. Moreover, the correlations among the competence valuation subscales indicate that the two enhancing dimensions are more distinct than the two protective dimensions. In view of these correlations, private and public enhancing esteemrelevant competence valuation subscales were retained in subsequent analyses.

6.1.4.4 Assessing and refining the attribution subscales It will be recalled that the entire set of attributional dimensions was included in the first administration of the questionnaire. In the review of literature it was argued that ability attributions for failure and external attributions for success and failure underpinned an inclination to self-protect. Thus, at a theoretical level, these three subscales were the most likely candidates for retention in subsequent analyses. It was decided that retention of additional subscales would be based on the empirical relationships amongst them and the central strategy subscales. Correlations between

142

the attributional dimensions and the central strategy subscales are presented in Table 6.7.

Table 6.7 Correlations among strategy subscales and attribution subscales

Ability Ability Effort success failure failure Ability - success .06 Ability - failure .23 .02 Effort - failure .34 .05 .52 Effort - success .23 .48 .17 External - failure .23 .44 .37 External - success .04 .31 .13 Active SH .14 .36 .26 Self-presented SH -.16 .60 .10 DE .04 .50 .15 Self-presented DE SH=Self-handicapping DE=Defensive expectations Coefficients > .18 are significant at p<0.05

Effort success

External failure

External success

-.03 .05 -.23 -.11 .02 .01

.78 .50 .48 .45 .52

.44 .51 .34 .44

Results in Table 6.7 suggest that the three attribution dimensions selected on a theoretical basis to be retained in subsequent analyses (ability failure, external success, and external failure) were amongst those most highly correlated with the central subscales and on this basis it was considered empirically defensible to retain them. Retention of additional subscales was to be considered in the light of compelling empirical associations that emerged from the correlational data. These data indicate that effort attributions for both success and failure and ability attributions for success were not substantially correlated with the central strategies as were the other three attribution subscales. On the strength of these findings, the three attribution dimensions retained for subsequent analyses were ability attributions for failure, external attributions for success, and external attributions for failure.

6.1.5 Summary of refinement analyses The preceding analyses were primarily aimed at reducing the item and subscale set. First, the content specificity of the constructs was examined and it was found that for the most part, the constructs are largely domain general and on this basis the mathematics-related items were selected to be the focus of subsequent analyses. Based on one factor congeneric CFA statistics, subscale reliabilities, item-total correlations,

143

and conceptual redundancy, numerous items from a variety of subscales were dropped. Correlations amongst conceptually-akin subscales and with central strategy subscales were then examined for evidence of empirical overlap and on the strength of these findings numerous subscales were dropped from subsequent analyses.

6.1.6 Confirmatory factor analysis of the refined instrument Having removed relatively weaker or empirically-redundant items and subscales, a CFA was performed on the remaining 22 subscales. Again, item parcels were used as indicators. No factor was comprised of less than three item parcels. This analysis yielded a chi square of 3543.33 (df=2114), a TLI of .84, an RNI of .86, and RMSEA of .058. Whilst these fit indices are not high, a final decision as to the acceptability of the hypothesised factor structure is suspended until the data for the whole sample (Nepean, Macarthur, and Sydney) are available. The factor loadings are presented in Table 6.8. The correlations between the central strategy measures and both predictor and outcome subscales are presented in Table 6.9. Correlations among predictor and outcome subscales are presented in Table 6.10.

144

Table 6.8 Factor structure of the final 22 factor model for Nepean (n=204)

Item parcel
Self-handicapping 1 Self-handicapping 2 Self-handicapping 3 Self-handicapping 4 Self-pres self-handicapping 1 Self-pres self-handicapping 2 Self-pres self-handicapping 3 Self-pres self-handicapping 4 Defensive expectation 1 Defensive expectation 2 Defensive expectation 3 Defensive expectation 4 Self-pres defensive expect 1 Self-pres defensive expect 2 Self-pres defensive expect 3 Self-pres defensive expect 4 Reflectivity 1 Reflectivity 2 Reflectivity 3 Private esteem-rel comp val 1 Private esteem-rel comp val 2 Private esteem-rel comp val 3 Public esteem-rel comp val 1 Public esteem-rel comp val 2 Public esteem-rel comp val 3 Ability - failure 1 Ability - failure 2 Ability - failure 3 External - failure 1 External - failure 2 External - failure 3 External - success 1 External - success 2 External - success 3 Ego-orientation 1

Factor loading
.78 .86 .84 .82 .82 .81 .86 .79 .87 .88 .80 .70 .79 .84 .90 .88 .56 .74 .76 .87 .89 .74 .89 .89 .89 .71 .69 .71 .44 .63 .66 .54 .77 .67 .86

Item parcel
Ego-orientation 2 Ego-orientation 3 Public self-conscious 1 Public self-conscious 2 Public self-conscious 3 Task-orientation 1 Task-orientation 2 Task-orientation 3 Avoidance orientation 1 Avoidance orientation 2 Avoidance orientation 3 Entity view 1 Entity view 2 Entity view 3 Increment view 1 Increment view 2 Increment view 3 Control future fail 1 Control future fail 2 Control future fail 3 Self-concept 1 Self-concept 2 Self-concept 3 Self-concept stability 1 Self-concept stability 2 Self-concept stability 3 Self-regulation 1 Self-regulation 2 Self-regulation 3 Persistence 1 Persistence 2 Persistence 3 Future plans 1 Future plans 2 Future plans 3

Factor loading
.91 .86 .87 .90 .91 .76 .76 .80 .83 .84 .76 .78 .86 .83 .71 .78 .74 .80 .77 .85 .85 .89 .78 .74 .75 .72 .70 .80 .75 .62 .73 .75 .77 .87 .90

145

Table 6.9 Correlations among central strategies and both predictor and outcome subscales for Nepean (n=204)

Active SH

Self-presented SH

DE

Self-presented DE

Reflectivity

Active SH .93 Self-presented SH .39 .46 DE .56 .67 Self-presented DE .15 .29 Reflectivity .30 .36 Ability - failure .51 .48 External - failure .44 .51 External - success .39 .37 Entity view -.02 .05 Incremental view .23 .36 Ego-orientation -.28 -.18 Task-orientation .34 .42 Avoidance-orientation -.21 -.19 Future plans .27 .34 Control future failure -.54 -.53 Persistence .24 .35 Public self-consciousness .04 .14 Private enhancing s-wth .19 .28 Public enhancing s-wth -.05 -.07 Self-concept .26 .38 Instability self-concept -.36 -.36 Self-regulation SH=Self-handicapping; DE=Defensive expectations Coefficients > .17 are significant at p<0.05

.76 .50 .60 .45 .33 .28 -.07 .27 .03 .49 -.26 .64 -.37 .40 .32 .37 -.30 .52 -.09

.55 .50 .52 .43 .37 -.02 .35 .04 .50 -.15 .51 -.44 .50 .31 .43 -.08 .43 -.06

.63 .39 .39 .14 .29 .38 .54 .53 -.02 .54 -.16 .70 .62 .63 -.05 .52 .35

146

Table 6.10 Correlations among predictors and outcomes for Nepean (n=204)

1 2 3 1.Ability - failure 48 2.External - failure 44 78 3.External - success 36 36 38 4.Entity view 02 14 33 5.Incremental view 27 42 37 6.Ego-orientation 24 10 03 7.Task-orientation 40 45 52 8.Avoid orientation -34 -12 -10 9.Future plans 61 48 40 10.Control fut fail -46 -39 -58 11.Persistence 54 33 33 12.Pub self-cons 43 31 29 13.Private s-wth 46 33 10 14.Public s-wth -45 08 -01 15.Self-concept 44 30 47 16.Instability s-c -06 -15 12 17.Self-regulation Decimal omitted Coefficients > .17 are significant at p<0.05

10

11

12

13

14

15

16

-37 09 -21 25 -09 30 -26 06 01 10 06 -02 -20

09 31 13 10 -01 01 19 34 27 16 24 22

28 46 -08 21 -31 61 69 66 09 33 -04

16 19 10 28 43 65 43 10 26 45

-12 37 -46 67 56 70 -09 41 -10

-20 50 -09 -09 -08 69 -23 20

-32 38 32 32 -19 44 -15

-26 -11 -26 32 -30 60

78 94 -07 52 14

86 -01 46 25

-01 47 12

-21 08

06

147

6.2 Measurement properties for the entire sample

6.2.1 Method 6.2.1.1 UWS Macarthur and University of Sydney The refined instrument (see Appendix D) was administered to students from a further two institutions. One of these institutions, UWS Macarthur, is a member of the network from which the first sample (UWS Nepean) was drawn. UWS Macarthur was established at the same time as UWS Nepean and is located in Sydneys south-western suburbs. Respondents from Macarthur were 189 first year teacher education students. Approximately 84% (n=169) were female and 16% were male (n=20). The mean age of the sample was 21.4 years (SD=5.9). Most students (71%, n=134) were undertaking a general Primary school program, whilst the others (29%, n=55) were undertaking an Early Childhood program. The mean TER was 61.29/100 (SD=15.33).

The third institution, the University of Sydney, is situated close to Sydneys central business district, is the oldest in the country, and typically attracts amongst the highest performing school leavers. Respondents from Sydney were 191 first year teacher education students. Approximately 73% (n=140) were female and 27% were male (n=51). The mean age of the sample was 19 years (SD=2.4). Most students (n=123, 64.4%) were enrolled in a Secondary teacher education program, while the remainder (n=68, 35.6%) was enrolled in a Primary teacher education program. The mean TER was 78.34/100 (SD=10.16).

6.2.1.2 Total sample characteristics In total, 584 respondents in their first year at university were drawn from three institutions (Nepean, Macarthur, and Sydney) in Sydney, Australia. Approximately 85% (n=495) were female and 15% (n=89) were male. The mean age was 20 years (SD=5.2). The mean TER was 66.79/100 (SD=16.5). Most students (63.7%, n=372) were enrolled in a Primary teaching program, 15.2% (n=89) in an Early Childhood program, and 21.1% (n=123) in a Secondary program.

148

6.2.1.3 Procedure The procedure followed that outlined in Section 5.1 of the preceding chapter.

6.2.1.4 Materials The refined instrument was adapted from the instrument administered in the first study (see Appendix A) and appears in Appendix D. Essentially this adaptation involved dropping the verbal component of the first instrument and removing some subscales and some items, the rationale for which appears in Section 6.1. The wording of items retained in the refined instrument was unchanged and the same rating scale (1=Strongly disagree to 7=Strongly agree) was used.

6.2.2 Confirmatory factor analyses for Macarthur and Sydney The first analysis conducted involved an assessment of the factor structure for the Macarthur sample. An identical 22 factor structure to that tested for Nepean was examined using CFA. A chi square of 3665.56 (df=2114) was obtained, with a TLI of .81, RNI of .84, and RMSEA of .06. The factor loadings are presented in Table 6.11. Testing the same 22 factor structure for Sydney yielded a chi square value of 3465.13 (df=2114), with a TLI of .84, RNI of .86, and RMSEA of .058. Again, whilst these fit indices are not high, a final decision as to the acceptability of the hypothesised factor structure is suspended until the data for the whole sample (n=584) are assessed simultaneously. The factor loadings for Sydney are parenthesised in Table 6.11. For each of the two samples, the correlations between the central strategy measures and both the predictor and outcome subscales are presented in Table 6.12. Correlations amongst the predictor and outcome subscales are presented in Table 6.13.

149

Table 6.11 Factor loadings for the 22 factor model: Macarthur and Sydney*

Item parcel
Self-handicapping 1 Self-handicapping 2 Self-handicapping 3 Self-handicapping 4 Self-pres self-handicapping 1 Self-pres self-handicapping 2 Self-pres self-handicapping 3 Self-pres self-handicapping 4 Defensive expectation 1 Defensive expectation 2 Defensive expectation 3 Defensive expectation 4 Self-pres defensive expect 1 Self-pres defensive expect 2 Self-pres defensive expect 3 Self-pres defensive expect 4 Reflectivity 1 Reflectivity 2 Reflectivity 3 Private esteem-rel comp val 1 Private esteem-rel comp val 2 Private esteem-rel comp val 3 Public esteem-rel comp val 1 Public esteem-rel comp val 2 Public esteem-rel comp val 3 Ability - failure 1 Ability - failure 2 Ability - failure 3 External - failure 1 External - failure 2 External - failure 3 External - success 1 External - success 2 External - success 3 Ego-orientation 1

Factor loading
.79 .79 .90 .80 .64 .79 .91 .85 .80 .87 .88 .87 .70 .80 .90 .92 .62 .84 .65 .85 .90 .71 .79 .88 .86 .61 .60 .80 .67 .62 .70 .31 .66 .70 .79 (.82) (.84) (.86) (.81) (.73) (.76) (.89) (.88) (.83) (.92) (.90) (.93) (.82) (.85) (.95) (.92) (.53) (.66) (.73) (.87) (.92) (.80) (.82) (.89) (.89) (.75) (.69) (.80) (.64) (.63) (.68) (.47) (.78) (.62) (.84)

Item parcel
Ego-orientation 2 Ego-orientation 3 Public self-conscious 1 Public self-conscious 2 Public self-conscious 3 Task-orientation 1 Task-orientation 2 Task-orientation 3 Avoidance orientation 1 Avoidance orientation 2 Avoidance orientation 3 Entity view 1 Entity view 2 Entity view 3 Increment view 1 Increment view 2 Increment view 3 Control future fail 1 Control future fail 2 Control future fail 3 Self-concept 1 Self-concept 2 Self-concept 3 Self-concept stability 1 Self-concept stability 2 Self-concept stability 3 Self-regulation 1 Self-regulation 2 Self-regulation 3 Persistence 1 Persistence 2 Persistence 3 Future plans 1 Future plans 2 Future plans 3

Factor loading
.83 .88 .77 .89 .87 .78 .81 .86 .69 .80 .82 .61 .71 .87 .71 .66 .91 .86 .92 .82 .83 .87 .86 .74 .86 .86 .70 .88 .83 .72 .70 .85 .80 .89 .90 (.90) (.85) (.75) (.93) (.87) (.58) (.75) (.83) (.59) (.81) (.80) (.81) (.79) (.78) (.62) (.67) (.71) (.88) (.87) (.79) (.85) (.90) (.90) (.80) (.89) (.79) (.71) (.73) (.79) (.72) (.74) (.86) (.82) (.90) (.89)

* Sydney results parenthesised

150

Table 6.12 Correlations among central strategies and both predictor and outcome subscales: Macarthur and Sydney*

Active SH Self-presented SH DE Self-presented DE Active SH .88 (.91) Self-presented SH .45 (.38) .50 (.46) Defensive expectation .56 (.48) .72 (.60) .76 (.74) Self-pres def expectn -.01 (.15) -.11 (.27) .21 (.39) .24 (.28) Reflectivity .30 (.39) .33 (.44) .68 (.52) .46 (.47) Ability - failure .37 (.43) .38 (.51) .48 (.37) .49 (.45) External - failure .36 (.42) .42 (.43) .63 (.49) .50 (.44) External - success .16 (.40) .19 (.29) .16 (-.01) .13 (.10) Entity view -.04 (-.02) -.08 (.04) -.22 (.16) -.09 (.18) Incremental view .32 (.15) .41 (.24) .33 (.17) .49 (.15) Ego-orientation -.38 (-.14) -.32 (-.05) -.19 (.12) -.12 (.09) Task-orientation .45 (.24) .47 (.37) .47 (.48) .53 (.50) Avoid-oriented performance -.10 (-.06) -.14 (-.05) -.20 (-.17) -.21 (-.09) Future plans .39 (.41) .35 (.44) .65 (.57) .44 (.44) Control future failure -.51 (-.48) -.39 (-.43) -.37 (-.27) -.26 (-.19) Persistence .18 (.11) .23 (.20) .32 (.38) .45 (.36) Public self-consciousness -.14 (.07) -.11 (.13) .15 (.22) .25 (.25) Private enhancing s-wth .20 (.20) .23 (.26) .30 (.37) .42 (.39) Public enhancing s-wth -.05 (-.11) -.12 (-.14) -.22 (-.20) -.10 (-.11) Self-concept .17 (.22) .18 (.27) .50 (.41) .31 (.31) Instability self-concept -.38 (-.44) -.31 (-.34) -.20 (-.24) -.10 (-.25) Self-regulation * Sydney results parenthesised SH=Self-handicapping; DE=Defensive expectation For Macarthur, coefficients > .17 are significant at p<0.05; For Sydney, coefficients > .18 are significant at p<0.05

Reflectivity

.41 .29 .27 .01 .33 .33 .36 .32 .03 .19 .24 .45 .45 .32 .03 .26 .48 (.36) (.57) (.31) (.03) (.05) (.40) (.59) (.66) (.06) (.38) (.08) (.58) (.53) (.53) (.07) (.54) (.41)

151

Table 6.13 Correlations among predictors and outcomes for Macarthur and Sydney*

1 1.Ability - failure 2.External - failure 3.External - success 4.Entity view 5.Incremental view 6.Ego-orientation 7.Task-orientation 8.Avoid orient perf 9.Future plans 10.Control fut fail 11.Persistence 12.Pub self-cons 13.Private s-wth 14.Public s-wth 15.Self-concept 16.Instability s-c 17.Self-regulation
58 (42) 62 (42) 39 (27) -22 (-04) 28 (24) -10 (12) 49 (54) -34 (-29) 71 (70) -30 (-21) 37 (41) 17 (29) 28 (40) -57 (-46) 48 (36) -03 (-01)

2
83 (66) 32 (29) -06 (05) 32 (44) -15 (19) 47 (56) -21 (-09) 55 (52) -40 (-24) 20 (39) 14 (37) 24 (46) -06 (-11) 41 (26) -18 (-09)

10

11

12

13

14

15

16

21 (12) 02 (17) 51 (36) -14 (03) 57 (53) -16 (-19) 53 (35) -51 (-44) 26 (33) 09 (31) 30 (41) -29 (-18) 48 (43) -20 (-32) -69 (-72) 19 (17) -34 (-21) 16 (07) -25 (-09) 12 (08) -23 (-19) 05 (08) -10 (-11) 10 (03) -17 (05) 13 (01) -16 (-14) 11 (14) 42 (26) 05 (17) 42 (-01) -09 (09) 22 (-09) 08 (07) 19 (24) 05 (22) 29 (-10) 08 (23) 22 (-01) 05 (19) 41 (47) 04 (03) 22 (18) -33 (-11) 61 (51) 38 (60) 69 (67) 04 (10) 12 (25) -12 (06) -02 (26) .34 (12) -05 (08) 51 (38) 24 (49) 59 (58) 18 (45) 10 (15) 03 (18) 48 (39) -06 (-09) 47 (50) -25 (-22) 51 (73) 31 (54) 42 (64) -09 (-03) 28 (49) -08 (02) -14 (-28) 20 (30) 01 (-05) 16 (-02) -01 (-04) 55 (65) 06 (-16) 12 (22) -35 (-37) 22 (36) 12 (23) 16 (36) -27 (-36) 43 (36) -20 (-20) -05 (-01) 18 (01) -11 (-13) 27 (22) -19 (-27) 74 (60) 70 (71) 94 (83) 02 (08) 19 (37) 05 (11) 70 (90) 06 (03) 19 (21) 21 (15) 04 (02) 17 (33) -04 (06) -08 (-11) 14 (11) -09 (-03)

* Sydney results parenthesised Decimal omitted For Macarthur, coefficients > .17 are significant at p<0.05; For Sydney, coefficients > .18 are significant at p<0.05

152

6.2.3 Testing for factor invariance across the three institutions and across sex CFA and SEM require large sample sizes and so ideally data are best analysed on a pooled basis rather than group-by-group. One important justification for pooling data is to demonstrate invariance in the factor structure of the entire instrument across variables for which it might be possible that the factor structure varies. Two such variables were considered important to test: institution and sex. Testing for factor invariance essentially involves comparing a number of models in which aspects of the factor structure are systematically held invariant across groups and assessing fit indices when elements of these structures are constrained. Relatively invariant fit indices are indicative of invariant factor structure. The present analyses, then, examined the comparative fit indices for five models across the three institutions and sex. The first model holds the number of factors invariant (FACT=IN); the second holds the factor loadings invariant (LOAD=IN); the third holds both factor loadings and uniquenesses invariant (LOAD=IN UN=IN); the fourth holds the factor loadings, the factor variances and correlations invariant (LOAD=IN CORR=IN); and the fifth holds the factor loadings, the uniquenesses, and the correlations invariant (LOAD=IN UN=IN CORR=IN). Statistics generated from these five models across the three institutions are presented in Table 6.14.

Table 6.14 Invariance tests across the three institutions

Model FACT=IN LOAD=IN LOAD=IN UN=IN LOAD=IN CORR=IN LOAD=IN UN=IN CORR=IN CHISQ 10674.03 10844.62 11262.52 11580.63 12019.91 DF 6342 6438 6578 6944 7084 CHISQ/DF 1.68 1.68 1.71 1.69 1.69 TLI .830 .830 .823 .834 .827 RNI .852 .849 .839 .841 .831

IN=Invariant FACT=Number of factors LOAD=Factor loadings UN=Uniquenesses CORR=Correlations and variances

153

Results in Table 6.14 indicate that when successive elements of the factor structure are held invariant across institutions, the fit indices are quite comparable and this suggests that pooling the three institutional groups in subsequent analyses is acceptable. The results of tests for factor invariance across sex are presented in Table 6.15. These fit indices also indicate that the factor structure is much the same across males as it is for females. On the basis of these results, it is justifiable to also pool the data across sex.

Table 6.15 Invariance tests across sex

Model FACT=IN LOAD=IN LOAD=IN UN=IN LOAD=IN CORR=IN LOAD=IN UN=IN CORR=IN CHISQ 5898.88 5944.60 6280.42 6214.55 6544.59 DF 4228 4276 4529 4346 4599 CHISQ/DF 1.39 1.39 1.43 1.39 1.42 TLI .926 .927 .919 .927 .920 RNI .935 .935 .927 .932 .924

IN=Invariant FACT= Number of factors LOAD=Factor loadings UN=Uniquenesses CORR=Correlations and variances

Accordingly, a final CFA using all 584 respondents was carried out (for completeness the results of a CFA for combined Nepean and Macarthur data are presented in Appendix E). The data fit the model well (chi square=4111.63; df=2114; TLI=.91; RNI=.93; RMSEA=.04). The factor loadings are presented in Table 6.16. The correlations between the central strategy measures and both the predictor and outcome subscales are presented in Table 6.17. Correlations amongst the predictor and outcome subscales are presented in Table 6.18.

154

Table 6.16 Factor structure of the 22 factor model for the entire sample (n=584)

Item parcel
Self-handicapping 1 Self-handicapping 2 Self-handicapping 3 Self-handicapping 4 Self-pres self-handicapping 1 Self-pres self-handicapping 2 Self-pres self-handicapping 3 Self-pres self-handicapping 4 Defensive expectation 1 Defensive expectation 2 Defensive expectation 3 Defensive expectation 4 Self-pres defensive expect 1 Self-pres defensive expect 2 Self-pres defensive expect 3 Self-pres defensive expect 4 Reflectivity 1 Reflectivity 2 Reflectivity 3 Private esteem-rel comp val 1 Private esteem-rel comp val 2 Private esteem-rel comp val 3 Public esteem-rel comp val 1 Public esteem-rel comp val 2 Public esteem-rel comp val 3 Ability - failure 1 Ability - failure 2 Ability - failure 3 External - failure 1 External - failure 2 External - failure 3 External - success 1 External - success 2 External - success 3 Ego-orientation 1

Factor loading
.80 .84 .87 .82 .74 .79 .88 .84 .83 .89 .86 .83 .77 .83 .92 .90 .56 .73 .72 .86 .91 .76 .84 .88 .88 .70 .67 .77 .56 .64 .69 .46 .73 .67 .84

Item parcel
Ego-orientation 2 Ego-orientation 3 Public self-conscious 1 Public self-conscious 2 Public self-conscious 3 Task-orientation 1 Task-orientation 2 Task-orientation 3 Avoidance orientation 1 Avoidance orientation 2 Avoidance orientation 3 Entity view 1 Entity view 2 Entity view 3 Increment view 1 Increment view 2 Increment view 3 Control future fail 1 Control future fail 2 Control future fail 3 Self-concept 1 Self-concept 2 Self-concept 3 Self-concept stability 1 Self-concept stability 2 Self-concept stability 3 Self-regulation 1 Self-regulation 2 Self-regulation 3 Persistence 1 Persistence 2 Persistence 3 Future plans 1 Future plans 2 Future plans 3

Factor loading
.88 .86 .81 .90 .89 .71 .78 .82 .72 .79 .80 .77 .80 .81 .68 .71 .77 .84 .85 .82 .85 .89 .85 .75 .84 .79 .70 .81 .79 .68 .74 .82 .82 .90 .91

155

Table 6.17 Correlations among strategies and both predictors and outcomes for entire sample (n=584)

Active SH

Self-presented SH

DE

Self-presented DE

Reflectivity

Active SH .90 Self-presented SH .37 .46 DE .53 .66 Self-presented DE .10 .22 Reflectivity .32 .37 Ability - failure .43 .46 External - failure .42 .46 External - success .33 .28 Entity view -.01 .01 Incremental view .20 .32 Ego-orientation -.27 -.18 Task-orientation .34 .43 Avoid-oriented perform -.18 -.13 Future plans .32 .37 Control future failure -.52 -.46 Persistence .17 .27 Public self-consciousness -.01 .07 Private enhancing s-wth .20 .27 Public enhancing s-wth -.08 -.10 Self-concept .20 .27 Instability self-concept -.41 -.35 Self-regulation SH=Self-handicapping; DE=Defensive expectations Coefficients > .10 are significant at p<0.05

.73 .37 .58 .44 .46 .13 -.02 .26 .01 .48 -.14 .61 -.31 .37 .24 .34 -.22 .48 -.17

.36 .47 .48 .45 .20 .03 .36 .01 .50 -.15 .45 -.30 .43 .27 .41 -.10 .34 -.14

46 42 31 07 23 37 50 49 04 39 06 59 54 50 03 44 42

156

Table 6.18 Correlations among predictors and outcomes for the entire sample (n=584)

1 2 3 1.Ability - failure 48 2.External - failure 44 75 3.External - success 34 33 25 4.Entity view -07 05 20 5.Incremental view 25 40 40 6.Ego-orientation 11 04 -04 7.Task-orientation 47 49 56 8.Avoid-oriented -25 -10 -17 9.Future plans 68 50 38 10.Control fut fail -29 -33 -52 11.Persistence 44 31 31 12.Pub self-cons 31 28 24 13.Private s-wth 38 34 35 14.Public s-wth -47 -02 -13 15.Self-concept 42 33 41 16.Instability s-c -02 -14 -25 17.Self-regulation Decimal omitted Coefficients > .10 are significant at p<0.05

10

11

12

13

14

15

16

-55 14 -24 17 -11 17 -22 06 -06 08 01 04 -17

22 33 12 15 -01 05 13 26 19 12 18 15

17 44 04 21 -23 58 57 66 10 22 -03

11 22 06 39 39 61 36 13 16 45

-09 44 -35 64 47 60 -07 39 -08

-14 38 -03 02 -06 62 -05 21

-30 33 24 28 -26 43 -17

-11 01 -18 28 -22 65

75 91 01 36 10

83 03 28 20

01 31 04

-11 12

-01

157

6.3 Higher order structure of the data and the role of respondents background characteristics

6.3.1 The higher order nature of the factors As discussed in the review of literature, groups of variables selected for analysis in the present study are in many respects conceptually akin. This is not surprising given that the overarching theory of self-worth motivation was invoked to assist in the selection of variables. It was therefore considered that analyses could be rendered more parsimonious and focused if conducted on the basis of variable clusters which are both conceptually and empirically related. This is an important step from an empirical perspective in the sense that considering variables in clusters accounts for a good deal of collinearity and conceptual redundancy which could be problematic for subsequent model solutions (see Bollen, 1989; Cohen & Cohen, 1983; Tabachnick & Fidell, 1996 for discussions of the problems of collinearity amongst predictor variables). With advances in SEM software such as LISREL, variable clusters can be quite easily and efficiently dealt with in the form of higher order factors. The formation of higher order factors is achieved through conceptual considerations such as assessing which variables are thematically alike, and empirical considerations in the form of correlations amongst subscales derived from the first order CFA.

6.3.1.1 Considering self-handicapping and defensive pessimism in terms of higher order factors Five first order factors comprise the strategies in the present study: active selfhandicapping, self-presented self-handicapping, defensive expectations, self-presented defensive expectations, and reflectivity. However, correlations presented in Table 6.17 suggest that there is a good deal of empirical overlap between the five strategies. Although representing the strategies in higher order factors was not originally intended, given the empirical overlap, dealing with them in this way is quite feasible. Accordingly, a number of models were proposed (see Figure 6.1) which represented the variety of ways in which the strategies can be operationalised as a group. The first model (Model 1) retains the five strategies as five first order factors and correlates each of these factors. Model 2 proposes one overarching higher order strategy factor.

158

Active SH Self-pres SH Active DE Self-pres DE Reflectivity

Active SH Self-pres SH Active DE Self-pres DE Reflectivity Strats HO factor

Active SH Self-pres SH SH HO factor

Active SH Active DE Active HO factor

Reflectivity Active DE Self-pres DE DE HO factor Self-pres SH Self-pres DE

Reflectivity

Self-pres HO factor

Model 1

Model 2

Model 3

Model 4

Active SH Self-pres SH Active DE Self-pres DE Reflectivity Self-pres HO factor Active HO factor

Active SH Active SH Self-pres SH Active DE Self-pres DE Reflectivity DP HO factor SH HO factor Self-pres SH Active DE Self-pres DE Reflectivity

HO Strat factor

Model 5

Model 6

Model 7

Figure 6.1 Seven models used to explore the multidimensionality of strategies

Note. Exogenous factors are inter-correlated in all models

159

Model 3 proposes one higher order self-handicapping factor comprised of active and selfpresented self-handicapping, one higher order defensive expectations factor comprised of active and self-presented defensive expectations, and retains reflectivity as a first order factor. Model 4 proposes one higher order active strategy factor comprised of active selfhandicapping and defensive expectations, one higher order self-presented strategy comprised of self-presented self-handicapping and defensive expectations, and retains reflectivity as a first order strategy. Model 5 is the same as Model 4 except that reflectivity is included in the higher order active strategy. Model 6 is the same as Model 3 except that reflectivity is included in the higher order defensive pessimism factor. Finally, Model 7 is the same as Model 2 except that reflectivity is retained as a first order factor. The fit statistics for each of these models are presented in Table 6.19.

Table 6.19 Fit statistics for the seven strategy models Chi square Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
391.00 677.11 409.04 642.50 650.51 409.44 677.11

df
142 147 145 145 146 146 147

RNI
.97 .93 .97 .94 .94 .97 .93

TLI
.96 .92 .96 .93 .93 .96 .92

RMSEA
.055 .079 .056 .077 .077 .056 .079

Results in Table 6.19 show that three models, Models 1, 3, and 6, represent a good fit to the data. For two reasons it was considered that Model 1 was not the best model to incorporate into subsequent analyses. First, the correlations generated in the 22 factor CFA (see Table 6.17) between self-presented and active self-handicapping (r=.90) and defensive expectations (r=.73) are quite high and to retain them as separate constructs could present problems of collinearity (e.g., suppressor effects - see Bollen, 1989; Tabachnick & Fidell, 1996). Second, the pattern of correlations in the 22 factor CFA (see Table 6.17) between both active and self-presented strategies and other constructs was quite similar and did not provide evidence of a great deal of discriminant validity. On these bases, Model 1 was not pursued and Models 3 and 6 were retained as possible candidates.

160

In deciding between Models 3 and 6, the pattern of correlations generated in the 22 factor CFA between the strategies and other constructs (see Table 6.17) was inspected again and it was apparent that reflectivity correlated with other constructs in quite a different manner than the other four strategies. More importantly, it operated quite differently from defensive expectations, the construct conceptually aligned with it (see Norem & Illingworth, 1993). Given this, it was not considered defensible to incorporate it into a higher order construct with defensive expectations. Accordingly, the factor structure to be tested using CFA was one that proposed two higher order factors, (active and self-presented) self-handicapping and (active and self-presented) defensive expectations, and one first order factor, reflectivity (Model 3). The factor loadings and factor correlations appear in Table 6.20. Self-handicapping and defensive expectations correlated quite highly (r=.62) while the correlation between reflectivity and defensive expectations was lower than expected (r=.37). This relatively low correlation between defensive expectations and reflectivity indicates that the two are relatively independent constructs. Further tests of their independence are the focus of analyses in Section 6.4.2 below.

Table 6.20 Strategy factor loadings and correlations based on two factor higher order CFA Higher order CFA factor loadings
First order CFA factors Active SH Self-presented SH Active DE Self-presented DE Reflectivity SH
.89 1.00 .77 .97 1.00

DE

Reflectivity

Correlations
SH DE Reflectivity
1.00 .62 .21 1.00 .37 1.00

SH=Self-handicapping; DE=Defensive expectations Correlation coefficients > .11 are significant at p<0.05

161

6.3.1.2 A six factor model of the affective and motivational predictors As outlined in the review of literature, four higher order predictor factors are proposed, whilst correlations derived in the 22 factor CFA (see Table 6.18) related to both selfconcept level and task-orientation would determine their position in this higher order structure. These correlations indicated that relative to other predictor constructs, selfconcept and task-orientation were not systematically related to other sets of subscales. Thus, arising from these empirical considerations, self-concept level and taskorientation were retained as separate first order factors. Thus, four higher order factors and two first order factors are proposed to represent the affective and motivational predictors. The first higher order factor represents a performance orientation and is comprised of ego-orientation, public self-consciousness, private and public esteemrelevant competence valuation, and avoidance-oriented performance. Conceptually, these subscales revolve around performance in the sense that students are concerned with how they are seen to be performing, feel successful by outperforming others, see themselves in terms of their ability to demonstrate competence, and are motivated to perform to avoid the negative implications of poor performance. Based on intercorrelations and subscale-content considerations, the second cluster of subscales is comprised of the two external attribution subscales and this higher order factor represents an external attributional orientation. The third cluster is proposed to be comprised of control over future failure, ability attributions for failure, and unstable self-concept. These three represent uncertain personal control. A fourth cluster supported by the correlations comprises entity and incremental beliefs about intelligence and represents a views of intelligence higher order factor.

This factor structure was examined using CFA and yielded an acceptable fit, TLI=.91, RNI=.91, RMSEA=.05 (chi square=1994.54; df=793). The factor loadings and correlations are presented in Table 6.21.

162

Table 6.21 Predictor factor loadings and correlations based on six factor higher order CFA

Higher order factor loadings

First order CFA factors Public comp valuation Public self-conscious Private comp valuation Ego-orientation Avoidance orientation Control future failure Ability failure attrib Unstable self-concept Entity view Incremental view External success attrib External failure attrib Task-orientation Self-concept Perform orientn
.97 .93 .84 .66 .65 .72 .93 .51 -.74 .75 .83 .90 1.00 1.00

Uncertain control

Intell views

External attribn

Taskorientn

Selfconcept

Correlations
Performance orient Uncertain control Intelligence views External attributions Task-orientation Self-concept
1.00 .48 .07 .43 .42 .02 1.00 -.24 .64 .12 -.43 1.00 -.16 .38 .07 1.00 .01 -.08 1.00 .14 1.00

Correlation coefficients > .10 are significant at p<0.05

6.3.2 Assessing the role of background variables Before assessing the structural relations in the hypothesised Time 1 process model, it was considered important to address the potential influence of students background characteristics. Thus the final component of the measurement phase of the study was to incorporate background variables into the full confirmatory factor model. In order to

163

test the extent to which background variables were associated with the central constructs, a CFA was conducted of the first and higher order factors and the background variables and the correlations between these variables assessed. The primary criterion used to incorporate any background variables into subsequent analyses was whether there were systematic and compelling correlations between a given background variable and other constructs in the model.

6.3.2.1 The background variables In total, thirteen background variables were incorporated into the CFA. Sex (0 female, 1 male), age (in years), number of semesters enrolled at university, and final school result (TER) were entered as originally scored. Two orthogonal contrasts were generated for university (-1, -1, 2 to compare Sydney with both Macarthur and Nepean; -1, 1, 0 to compare Macarthur and Nepean) and also for degree (-1, -1, 2 to compare secondary with both primary and early childhood; -1, 1, 0 to compare primary with early childhood). Five polynomial orthogonal contrasts were generated to compare six levels of final school year math course (-5, -3, -1, 1, 3, 5; 5, -1, -4, -4, -1, 5; -3, 5, 2, -2, -5, 3; 2, -5, 3, 3, -5, 2; 0, 3, -6, 6, -3, 0; where, for example, the first contrast assesses the linear effect of final year maths course level). All background variables were incorporated into the CFA as single item measures with loadings fixed to unit value and their uniquenesses fixed to zero.

6.3.2.2 The results of the CFA incorporating background variables The fit of the full CFA was acceptable, RNI=.90, TLI=.89, RMSEA=.04 (chi square=6097.14, df=3021). Given that the loadings for the background variables were fixed to unit value and loadings of other factors are essentially the same as those that were presented earlier, the only results presented here - and which are of central interest - are the correlations amongst the factors. These are shown in Table 6.22.

Correlations between background variables and other constructs indicate that there are three relatively stronger associations which could be considered to be freed in the Time 1 process model. The first is between the first course contrast and future plans:

164

Students enrolled in a Secondary program report less willingness than students in Early Childhood or Primary programs to study maths in the future. The second is between the first university contrast and future plans: Students at Sydney are less inclined to pursue mathematics in the future than students at Nepean and Macarthur. The third is between the linear math course contrast (Math course 1) and maths self-concept : The higher the level of maths students undertook in their final school year, the higher their maths self-concept. Importantly, however, the criterion used as the basis for incorporating background variables into the complete academic process model is that there be a set of systematic relationships between a given background variable and other constructs. On this count, no background variables are particularly compelling candidates. Accordingly, no background variables are incorporated into the Time 1 process model which is now assessed.

6.4 Examining the Time 1 process model and assessing supplementary issues

6.4.1 Time 1 process model The central Time 1 process model freed the beta paths between (a) the four higher and two first order predictor factors and the higher order self-handicapping, higher order defensive expectations, and first order reflectivity strategies, and (b) the three strategies and the three Time 1 outcomes, self-regulation, persistence, and future plans. The six predictors were free to correlate, as were the residual variances among the three strategies and among the three outcomes. Modification indices derived from the first application of this model identified a large estimated change between self-concept and both future plans and persistence and so these parameters were freed to be estimated as well. To enhance model parsimony, structural parameters not significant at the p<0.10 level or better were dropped from the model. In trimming the model in this way, implicit beliefs about intelligence were excluded from a subsequent estimation. As it turned out, all paths retained in this subsequent estimation were significant at p<0.05.

165

Table 6.22 Correlations among background variables and other first and higher order constructs in the model
1 1. Sex 2. Age 3. Semester 4. Final mark 5. EC & Prim vs Sec 6. EC vs Primary 7. UWS vs Sydney 8. Nepean vs Macarth 9. Math course 1 10. Math course 2 11. Math course 3 12. Math course 4 13. Math course 5 Mastery orientation Self-concept Performance orientation External orientation Uncertain control Intelligence Reflectivity Self-handicapping Defensive expectations Self-regulation Persistence Future plans
.01 .03 .02 .31 .05 .22 -.03 .06 -.03 -.03 -.01 .03 -.14 .02 -.03 .08 -.09 .02 -.08 .15 .03 -.15 -.24 -.08

2
.09 .06 -.13 .06 -.21 .03 -.12 .23 -.13 -.02 .06 .07 -.05 -.06 -.14 -.01 -.02 -.04 -.14 -.05 .11 .15 .08

10

11

12

13

.24 .19 .11 .21 -.41 .08 .05 -.10 -.05 .06 .02 -.01 -.01 -.04 -.12 -.03 -.06 .01 -.07 -.04 -.03 -.20

.31 .04 .55 -.06 .20 .24 -.27 -.23 .11 -.01 .06 .09 -.05 -.17 -.02 .03 -.01 .03 -.02 -.03 -.18

.34 .74 -.02 .06 .07 -.09 -.03 .07 -.07 -.13 .03 .06 -.10 .01 -.06 .16 .02 -.11 -.21 -.33

.12 -.14 -.02 -.02 .05 .01 -.05 -.03 -.11 -.09 -.04 -.03 -.03 -.13 .01 -.02 -.07 -.02 -.23

-.02 .14 .10 -.14 -.10 .08 -.09 -.09 .03 .06 -.16 -.01 -.03 .11 .01 -.12 -.21 -.35

.01 .01 .01 -.01 .02 .01 .03 .05 .05 .04 .02 -.01 -.07 .02 .01 .01 .22

-.19 -.55 -.48 .38 .01 .31 .09 .01 -.13 .04 -.01 .01 .04 -.03 .05 .10

-.34 -.44 -.11 .02 .06 -.01 -.05 -.08 .00 -.04 -.06 -.02 -.01 .02 .03

.57 -.57 .01 -.23 -.05 .10 .16 -.03 .10 .05 .06 .01 -.04 -.13

.21 .01 -.28 -.05 .01 .12 -.03 .04 -.01 -.02 .01 -.04 -.15

.01 .06 .02 -.08 -.09 .01 -.06 -.03 -.05 -.01 .02 .02

166

The remaining structural relations are presented in Figure 6.2. The data yielded an acceptable fit to the model, TLI=.90, RNI=.91, RMSEA=.046 (chi square=4271.03, df=1905). Performance orientation positively predicted the three strategies. External attributional orientation predicted self-handicapping and to a lesser extent predicted defensive expectations. Uncertain personal control positively predicted defensive expectations in quite a marked fashion, moderately positively predicted reflectivity, and to a lesser extent positively predicted self-handicapping. Task-orientation negatively predicted self-handicapping and defensive expectations, and positively predicted reflectivity. Self-concept to a small (but significant) degree predicted reflectivity and persistence and to a large degree predicted future academic plans. In terms of the relationships between strategies and outcomes, self-handicapping and defensive expectations negatively predicted self-regulation and persistence, while reflectivity quite strongly positively predicted self-regulation and to a lesser extent positively predicted persistence. The three strategies did not significantly predict future academic plans for which the only significant predictor was self-concept.

6.4.2 Further testing of defensive pessimism As outlined in the review of literature, no clear explication of how defensive expectations and reflectivity work together to constitute defensive pessimism has been provided to date. Most of the earlier research assessed only defensive expectations (e.g., Norem & Cantor, 1986a, 1986b). Later research (Norem & Illingworth, 1993) suggested that defensive pessimism is represented by the interaction of defensive expectations and reflectivity (with individuals high on both representing the defensive pessimist). A more recent suggestion is that defensive expectations relax students sufficiently to enable them to engage in reflectivity which enhances subsequent performance (Norem, personal communication, March 26, 1997; March 7, 1998). To explore these two possibilities, the present study examined models in which (a) defensive expectations and reflectivity interact and (b) defensive expectations predict reflectivity.

167

Public selfconsciousness

.93 .66 .65 Performance orientation .14 .21 .20 -.32 .35 -.15 .39 -.41 .57 .85 .86 -.24 .16 .87 .76 Uncertain personal control .16 .87 Selfconcept .62 Self-pres defensive expectations .46 -.17 .52 Defensive expectations .83 Active defensive expectations .20 Persistence External orientation Reflectivity .21 .20 Selfhandicapping -.33 Self-regulation Active self-handicapping .93 .97 Self-pres self-handicapping

Egoorientation Avoidance orientation Private comp valuation Public comp valuation External External causes causes of of success success External causes of failure Ability causes of failure Control future failure Instability self-concept

.84

.97

Task-orientation

.56

Future plans

Figure 6.2 Significant structural relations in Time 1 process model Note. All paths significant at p<0.05

168

Another theoretical issue to be considered in supplementary analyses concerns the pool of students for which it is possible to be defensively pessimistic. To date, defensive pessimists have been defined as those who score in the top third of a selfconcept-like screener and who also report negative expectations. The present study operationalised defensive expectations such that it was possible for students at any level of self-concept or previous performance to be defensively expectant. Thus, for comparability, it is considered important to operationalise the defensive expectations construct along the lines of previous research and assess its role in the context of this operationalisation.

6.4.2.1 Interaction of defensive expectations and reflectivity Whilst the findings presented thus far support the hypothesised distinctiveness of defensive expectations and reflectivity (e.g., see the quite different predictive paths between the two strategies and the academic outcomes in Figure 6.2), a further test as to their inter-relationship and the way they operate requires that their interaction be studied: It may be that their interactive effect over and above their main effects provides better support for the defensive pessimism construct as theoretically advanced by researchers in the area.

Dealing with interaction effects in structural equation modelling is a complex business and to date there have been a variety of approaches to modelling them. The present approach is one recommended by Ping (1996a, 1996b). The technical issues involved in this approach are presented in detail in Appendix F1. Following Pings methodology, the latent interaction term was incorporated into the Time 1 process model as a strategy alongside the other three strategies (self-handicapping, defensive expectations, and reflectivity). When the latent interaction was examined in this way, it was not found to be significantly predicted by any of the affective and motivational constructs nor did it significantly predict any of the three academic outcomes. Thus, in terms of the Time 1 data, defensive expectations and reflectivity seem to operate primarily as main effects and do not interact.

169

6.4.2.2 A model in which defensive expectations predict reflectivity As noted above, another conceptualisation of defensive pessimism is one in which defensive expectations relax students sufficiently to engage in reflectivity (Norem, personal communication, March 26, 1997, March 7, 1998). To test this issue, the final trimmed model was estimated again but rather than correlating (the residuals of) defensive expectations and reflectivity, the defensive expectations construct was freed to predict reflectivity. Results from this model (chi square=4270.87, df=1906) indicated that defensive expectations did not significantly predict reflectivity (=.07). Rather, the bulk of the variance in reflectivity was explained by the affective and motivational predictors as is hypothesised in the present study. Having carried out this analysis, however, it is important to recognise that the model in which defensive expectations and reflectivity are correlated (the central model presented in Figure 6.2) is statistically equivalent to a model in which defensive expectations predict reflectivity. Thus, the process model hypothesised in this study is quite consistent with recent suggestions by Norem. The purpose for estimating the model in which defensive expectations predict reflectivity (even though it is statistically equivalent to the central process model), is to explicitly emphasise the predictive path, rather than establish the relationship through correlated residuals.

6.4.2.3 Interaction of level of self-concept and defensive expectations The present study is unique because the nature of questionnaire items was such that defensive expectations were assessed for students at all academic levels, whereas Norem and colleagues have typically administered a self-concept-like screener (Ive generally done pretty well in academic situations in the past), taken those scoring in the top third of that item, and designated those also scoring high in negative expectations as defensive pessimists. To date, then, defensive pessimism has been assessed primarily in terms of students who rate themselves in the top third of the selfconcept-like screener. Accordingly, previous findings in relation to defensive pessimism have been restricted to a minority of students. Thus the present studys operationalisation extends the treatment of defensive pessimism through its consideration of students at all levels of previous academic performance. However, whilst this might be conceptually defensible, an empirical justification for extending

170

defensive expectations to encompass all students (and not just the top third) is required.

Arguably the ideal means of justifying the present operationalisation is to directly test defensive pessimism along the lines of that originally proposed. Classifying defensive pessimists as those who score in the top third of a self-concept-like screener and high in negative expectations implies an interaction between the screener and negative expectations. Whilst the present study did not administer the self-concept-like screener used in much of the defensive pessimism research, one item in the selfconcept scale (I have always done well in these subjects) is parallel and so this item was used to assess defensive pessimism along the lines of that typically operationalised. The interaction between this self-concept item and defensive expectations was then incorporated in a structural model in which the self-concept item, the defensive expectation items, and the interaction between the two dimensions were predictors of self-regulation, persistence, and future academic plans. As the selfconcept screener was a single item, its uniqueness was conservatively fixed at .2.

Results indicated that the interaction of the self-concept item and defensive expectations did not significantly predict any academic outcomes. (Moreover, when the effect of this interaction on grades was later tested, results were not statistically significant.) As a further test of the traditional defensive pessimism operationalisation, the full Time 1 process model was estimated for only those students scoring in the top third of the target self-concept item. Consistent with the process model findings derived from the whole sample, results derived from the process model restricted to students scoring in the top third of the self-concept item indicated that defensive expectations remains a negative predictor of both self-regulation and persistence. It appears, then, that defensive expectation effects are not restricted to students scoring relatively higher in self-concept.

171

6.4.3 A test of the interaction of level and stability of self-concept It was also proposed in the review of literature that findings regarding stability of selfconcept may differ as a function of the level of individuals self-concept (see Kernis et al., 1989, 1992, 1993). It was tentatively proposed that individuals high and unstable in self-concept are more likely to engage in self-protective strategies such as selfhandicapping and defensive expectations, however, no firm predictions were made in relation to this matter. To test this issue, it was proposed that the interaction between the two be investigated. Accordingly, the latent interaction between the two was operationalised in much the same way as defensive expectations and reflectivity presented above (see Appendix F1 for technical details). The latent interaction was incorporated into the Time 1 process model as a predictor alongside the other affective and motivational predictors in the model. Results indicated that this interaction did not significantly predict any of the strategies (including the interaction of defensive expectations and reflectivity). Based on Time 1 data, then, it is concluded that level and stability of self-concept operate primarily as main effects.

6.5 Summary of Time 1 results A number of key findings emerged in the first phase of quantitative analyses. The first concerned the domain/content specificity of the majority of constructs in the model. Math and verbal domains were, for the most part, highly correlated suggesting that the majority of constructs under study - particularly the strategies - are domain general. Secondly, the content specificity data showed that using CFA procedures, rather than standard raw score correlations, was important because it enabled parallel uniquenesses to be correlated thereby attenuating the upward bias in correlations between math and verbal constructs.

Having identified a content focus for subsequent analyses (mathematics), a series of CFAs identified items and subscales that could be dropped from subsequent analyses. For the remaining 22 scales, the factorial invariance across sex and institutions was examined. Fit indices derived from these analyses indicated that the factor structure across sex and institutions was markedly similar and on this basis the data were pooled and subsequent analyses were performed using the entire (n=584) sample.

172

Given the conceptual overlap between a number of subscales and consistent with Cohen and Cohen (1983), subscale clusters were considered a feasible means of representing the large number of constructs in the study. Essentially this involved modelling higher (second) order factors that underlay first order constructs. These analyses comprised two parts. First, the strategies were assessed for their empirical distinctiveness and the presence of possible higher order factors was explored. Second, the affective and motivational predictors were examined on a similar basis. In terms of the strategies, numerous models were explored and it was found that the best representation of the strategies was one that proposed a higher order self-handicapping strategy comprised of active and self-presented self-handicapping, a higher order defensive expectation strategy comprised of active and self-presented defensive expectations, and retained reflectivity as a first order strategy. In terms of the fourteen affective and motivational predictors, a six factor model was considered the ideal structure for further analyses. This structure comprised higher order performance orientation, external attributional orientation, intelligence beliefs, and uncertain personal control factors and first order task-orientation and self-concept factors.

Having ascertained acceptable measurement properties, factorial invariance, and compelling predictor and strategy higher order clustering, the full Time 1 process model was then examined using structural equation modelling. Performance orientation positively predicted the three strategies. Task-orientation negatively predicted self-handicapping and defensive expectations and positively predicted reflectivity. An external attributional orientation positively predicted self-

handicapping and to a lesser extent, defensive expectations. Uncertain personal control quite strongly positively predicted defensive expectations and reflectivity and to a lesser extent positively predicted self-handicapping. Self-handicapping and defensive expectations negatively predicted self-regulation and persistence, while reflectivity positively predicted both these outcomes. Level of self-concept strongly positively predicted future academic intentions and to a lesser extent, reflectivity and persistence.

Further tests of theory related to defensive pessimism incorporated the interaction between defensive expectations and reflectivity and tested a model in which defensive 173

expectations predict reflectivity. Consistent with present contentions, defensive expectations and reflectivity seemed to operate as quite independent constructs. Similarly, no significant interaction between a self-concept item and defensive expectations was found. This suggested that the effects of defensive expectations are relatively independent of level of self-concept and supported the present operationalisation of defensive expectations as a construct not simply confined to students whose previous performance is high.

174

7. EXPLORING THE TIME 1 PROCESS MODEL FROM A QUALITATIVE PERSPECTIVE

7.1 Qualitative perspectives in a quantitative domain There has, in recent years, been growing recognition that quantitative and qualitative research methods can complement each other in ways that both consolidate findings and shed light on issues that were previously seen to be best addressed by one and not the other. Whilst the emphasis of the present study is on the quantitative data, it is proposed that a qualitative approach to the issues can add to these quantitative data in important ways. Gill (1996, see also Abott-Chapman, 1993; Bogdan & Biklen, 1982; Campbell, 1975; Cook & Campbell, 1979; Cronbach, 1975; Firestone, 1987; Gage, 1989; Glesne & Peshkin, 1992; Howe, 1988; Patton, 1987, 1990) argues that the adoption of a mixed methods approach will often prove more productive than a single line of inquiry. The multiple discourses of educational research may then be utilised in ways which strengthen the research endeavour even as they inevitably broaden the research enterprise (p. 44). According to Brewer and Hunter (1989), each approach is not without its weaknesses and there is a place for one to compensate for the weaknesses of the other. They concluded that a multimethod approach to research is important and that the aim of such research is to attack a research problem with an arsenal of methods that have nonoverlapping weaknesses in addition to their complementary strengths (p. 17).

It is considered that the substantive issues in the present study lend themselves quite well to this multimethod approach. Whilst the quantitative data can indicate the nature and strength of relationships between the constructs, there can be a place for a more in-depth analysis of such data from respondents perspectives. Self-handicapping and defensive pessimism are strategies that may manifest themselves for a variety of subtle or not so subtle reasons and it is not until individual respondents personal perspectives are examined that the richness of this information can be used to better understand the constructs and their relationships in the substantive area. Indeed, there may be ways in which self-handicapping and defensive pessimism occur and reasons for these that have not been considered in the questionnaire and which only qualitative methods can identify. 175

The role of qualitative research in the context of a quantitative report can have three broad aims (Patton, 1987; Strauss & Corbin, 1990). Firstly, it can seek to test the results generated in the quantitative analysis. Secondly, it can shed light on the quantitative findings, giving voice and life to the results, and provide an insight into the personal perspectives of respondents, particularly as they relate to the key quantitative findings. Thirdly, qualitative research can introduce and provide insight into issues that have not been identified by the quantitative methods and perhaps move the research in potentially fruitful new directions. The present study adopts the second and third aims. As Patton (1987) argues,

when a large-scale survey has revealed certain marked and significant response patterns, it is often helpful to fill out the meaning of those patterns through in-depth study using qualitative methods. The quantitative data identify areas of focus; the qualitative data give substance to those areas of focus . . . Qualitative data can put flesh on the bones of quantitative results, bringing the results to life through in-depth elaborations (p. 38).

Qualitative research permits the study of people, their interpretations of reality, and the contexts in which these interpretations are made. Essentially it enables the researcher to explore respondents views of the world. It is this human experience which the present study attempts to uncover from the perspectives of the participants (see also Filstead, 1970; Guba & Lincoln, 1981; Stake, 1975). Specifically, the aim of this component of the research is to go beyond the quantitative data and capture the personal perspectives of self-handicappers and defensive pessimists. Selfhandicapping and defensive pessimism, it is argued, are ideal constructs to study in a qualitative fashion primarily because the extent of their deployment is largely mediated by the way individuals view the world and their part in it and so personal perspectives about these phenomena can be particularly valuable in understanding them.

In sum, then, it is argued that incorporating a qualitative component in the present 176

study is not only defensible but important. The use of qualitative data in the present study is for the primary purposes of (a) exploring the quantitative findings from students own perspectives and (b) gaining new insights that might be provided by respondents and which suggest a need for further consideration. Importantly, the intention of the qualitative component of the thesis is not to test the quantitative findings. Rather, it is to focus on the key areas identified in the quantitative analyses with a view to locating them within the context of students own academic lives and personal perspectives. 7.2 Research issues The issues to be addressed in this component of the study closely follow the model generated by the Time 1 quantitative study (see Figure 7.1) and which revolve around self-handicapping and defensive expectations. This first involves exploring the nature of self-handicapping and defensive expectations. Secondly, the affective and motivational constructs that predict self-handicapping and defensive expectations are each examined. Thirdly, the academic outcomes that follow from self-handicapping and defensive expectations are explored. These research issues are fully outlined below and are presented in Table 7.1.

While the quantitatively-driven process model locates affective and motivational factors as the first part of the model, the qualitative analysis begins with an examination of the issues of self-handicapping and defensive expectations (the midlevel components of the model). This is because the project is centrally concerned with these two constructs and it is considered that by presenting these results first, an appropriate backdrop is provided against which students reports about the affective and motivational predictors and academic outcomes can be set.

177

AFFECTIVE AND MOTIVATIONAL PREDICTORS

SELF-PROTECTIVE STRATEGIES

ACADEMIC OUTCOMES

1.1.1 Public selfconsciousness 1.1.2 Egoorientation 1.1.3 Avoidance orientation 1.1.4 Private comp valuation 1.1.5 Public comp valuation 1.4 Mastery orientation 1.2.1 External External causes causes of of success success 1.2.2 External causes of failure 1.2.2 Ability causes of failure 2.1 Selfhandicapping

1.1 Performance orientation

3.1 Selfregulation

- 1.2 External orientation 2.2 Defensive expectations 3.2 Persistence

1.3.1 Control future failure

1.3 Uncertain personal control

Figure 7.1 Model identified through quantitative analysis

1.3.2 Instability self-concept

All paths positive except where otherwise indicated NUDIST nodes also indicated - see Appendix I

178

Table 7.1 Research issues to be addressed in Time 1 qualitative analysis

Component of model
Strategies

Constructs comprising components of model

Self-handicapping

Research issues to be addressed

Students views on how self-handicapping occurs in their own lives; Students beliefs about why they might engage in self-handicapping; Perceived advantages and disadvantages of self-handicapping; Perceived public/self-presented aspects of self-handicapping.

Defensive expectations

Students views on how defensive expectations operate in their own lives; Students beliefs about why they might hold defensive expectations; Perceived advantages and disadvantages of defensive expectations; Perceived public/self-presented aspects of defensive expectations.

Predictors

Performance orientation

Students concern with how others view their academic performance; The extent to which outperforming others provides feelings of success; The extent to which students are motivated to perform out of fear of poor performance; The extent to which students self-esteem is dependent on feeling and appearing competent.

External attributional orientation Uncertain personal control

The extent to which students attribute success and failure to external factors. The extent to which students attribute failure to ability; Their perceived control over avoiding doing poorly in the future; Stability and consistency of students self-concept.

Task-orientation Outcomes Self-regulation Persistence

The extent to which mastery provides students with feelings of success and satisfaction. The self-regulatory behaviours in which students engage. The extent to which students are inclined to persist at tasks in the face of challenge.

179

7.3 Method 7.3.1 Sample Included in the Time 1 questionnaire was an invitation for students to record their names and telephone numbers if they were willing to be interviewed at a later time. In total, 134 students volunteered this information. A list of these students was generated on which students were sorted in ascending order on the basis of their responses to the self-handicapping subscale and defensive expectations subscale. Further sorting was done by sex, age, and institution. Respondents were selected using purposive sampling such that the sample was intended to match approximately, age, sex, and institutional composition of the larger sample as well as ensuring that those sampled represented the four target academic profiles (low and high self-handicapping and low and high defensive expectations). Twenty-four students were selected and agreed to a follow-up interview. The mean age of respondents was 21 years 10 months. Six respondents were male and 18 were female. Eight respondents were drawn from each of the three institutions in such a way that two were high self-handicappers, two were low selfhandicappers, two were high in defensive expectations, and two were low in defensive expectations. Sample composition appears in Table 7.2.

Table 7.2 Time 1 interview participants*

University 1 High self-handicapper (HISH) Low self-handicapper (LOSH) High defensive pessimist (HIDE) Low defensive pessimist (LODE)
Trudy (HISH) Christine (HISH) Reg (LOSH) Gina (LOSH) Gwen (HIDE) Cassie (HIDE) Tony (LODE) Amanda (LODE)

University 2
Peter (HISH) Carol (HISH) Sharon (LOSH) Lucy (LOSH) Marie (HIDE) Robert (HIDE) Bernadette (LODE) Tasha (LODE)

University 3
Sophie (HISH) Rachel (HISH) Angela (LOSH) Tania (LOSH) Joe (HIDE) Dianne (HIDE) Brendan (LODE) Lynne (LODE)

* Pseudonyms used and University not specified to ensure confidentiality

180

7.3.2 Procedure Quiet offices on each of the three campuses were used as the settings in which to conduct the interviews. Respondents were welcomed and informed that the researcher was interested in complementing the quantitative data with in-depth interviews that would provide some insights into the personal perspectives of students. Respondents were not informed about the basis upon which they were selected. Rather, they were told that the researcher wanted to talk to a broad range of students and that they comprised one component in that range. Respondents were then asked if audio taping the interview was acceptable to them and were informed that the interview data were confidential and that pseudonyms would be used when reporting extracts from the interviews. Respondents were informed that the three universities ethics committees required that informed consent be provided to the researcher and were asked to sign a consent form (see Appendix G). Once this was signed the interview proceeded. The interview was based on a semi-structured interview schedule (see Appendix H). The interview began with some questions about what subjects respondents were studying, their thoughts about the course and university life in general. When some rapport had been established through the introductory questions, the issues centrally relevant to the study were explored. The average duration of each interview was approximately 30 minutes (ranging between 20 and 45 minutes). The interview was transcribed verbatim and the transcripts were used for the data analysis.

7.3.3 Data analysis 7.3.3.1 Formation of categories The manner in which the data are coded and categorised (see Appendix I) directly influences the way in which the data are analysed and so data categorisation is one of the most important aspects of qualitative research. Categorisation can dictate the way themes are generated, the way these themes are integrated in subsequent analyses, and ultimately the conclusions drawn from the study. One criticism of qualitative research is that researchers often do not document how their categories are formed and so do not leave the important audit trail for the reader (Lincoln & Guba, 1985). In suggesting ways by which the categorisation process can be more transparent, Constas (1992) has proposed that researchers fully document their category development by 181

outlining the components of categorisation and the temporal aspects of category formation. He divides the components of categorisation into various elements, three of which - origination, verification, and temporal designation - are relevant here.

Origination refers to where the responsibility or authority of categorisation resides. Constas lists five elements of origination of which three are invoked here: (a) Participants: Category components are in part based on participants responses; (b) Investigative: Other category components are derived from the researchers intellectual interests and theoretical viewpoint; and (c) Literature: Category components are in part based on established (primarily quantitative) research findings. Verification refers to the strategies used to justify the creation of categories, of which, two are used here: (a) Rational: Categories have face validity and the appearance of logical connectedness; and (b) Referential: Established research findings have been used to justify the category generation. In terms of temporal designation, most of the categories and component issues were determined a priori in the sense that they are based on the extant quantitative literature. Notwithstanding this, there is an element of iterative processing in that additional categories are formed in response to participants reports which, on the basis of the researchers interpretation, went beyond the predetermined category structure.

7.3.3.2 Software used to analyse qualitative data Data were analysed using the NUDIST (Non-numerical Unstructured Data Indexing, Searching, and Theorising) software program (Richards, Richards, McGalliard, & Sharrock, 1993). NUDIST is designed to integrate a number of documents (in this case, interview transcripts) for analysis; index (using [pre]determined categories - see Appendix I) components of text (in this case interview questions and responses); search for text; search indexing; and reorganise indexing systematically in a manner that can both harness and generate theory. Essentially, NUDIST assists in organising coded data into nodes and node patterns that constitute the fabric of data analysis and theorising findings.

7.4 Qualifications relevant to the data analysis

182

One difficulty encountered during the interviews which does have an important bearing on the presentation of findings concerns the amount and nature of the information provided by respondents. A few examples illustrate some of the difficulties encountered along these lines: Respondents provided one word answers to questions and did not comment further when invited to do so; respondents did not relate to the question and pursuing the issue seemed to confuse them; respondents seemed to misinterpret the issue under focus and attempts at clarification were unsuccessful; respondents had partly answered a later question while responding to a former and so directing the later question to them again to obtain a direct response was considered inappropriate; respondents were unsure as to what was being asked of them and clarification was unsuccessful. For these reasons, when presenting findings, it is often the case that the entire groups responses are not provided. Instead, the focus is usually upon data derived from respondents who have interpreted the questions as they were intended or have provided responses which provide rich, illustrative qualitative information, represent challenging insights, point the field in new directions, or present ideas for further consideration.

Second, it is emphasised that the qualitative data are not designed to test the quantitative findings. The position held here is that selecting a small proportion of students from a large sample cannot be expected to yield qualitative data directly parallel to quantitative findings. Indeed, the weaker the quantitative finding, the lower the chances are that responses from students selected for a follow-up qualitative study will reflect that finding. The qualitative data presented here, then, are used in a way that sheds light on a given quantitative finding and illustrates how such a quantitative relationship manifests itself in students lives from students perspectives. Nevertheless, where systematic inconsistencies occur, these are reported, as are findings which run contrary to the quantitative findings but which may point the field in new directions.

A third point to consider is that the data to follow, particularly those pertaining to the salience of the affective and motivational constructs incorporated in the process model are not intended to reflect cause and effect relationships. Rather, the purpose of the Time 1 interview data is to clarify exactly how these constructs accompany the 183

identified self-handicapping and defensive expectations. In a sense, then, the interview data are more intended to map the terrain of factors associated with selfhandicapping and defensive expectations than to explain these strategies. Explaining these strategies was more the purpose of the quantitative model in which a set of factors was explicitly modelled to predict self-handicapping and defensive expectations. Notwithstanding this, in a number of respects it is clear that some affective and motivational constructs or key events in students lives do antecede selfhandicapping and defensive expectations and can be contended to give rise to these strategies in the process of students defensive manoeuvring.

Finally, as the interview data to follow show, the profiles of self-handicappers and defensively expectant students are remarkably similar in terms of the underlying affective and motivational constructs. Perhaps this is not surprising given that they are both proposed to be strategies subsumed under a shared conceptual framework in the form of self-worth motivation theory. Thus, in many parts of this chapter, students high in self-handicapping or defensive expectations are considered together because there is often a parallel salience of affective and motivational constructs operating in their academic lives. Whilst this is convenient from an organisational and presentational perspective, it is primarily a consequence of the parallel nature of the data. This is not to suggest that the strategies themselves are the same. The interview data presented below show how students can be clearly differentiated in terms of their strategic use of self-handicapping and defensive expectations 7.5 Results The results revolve around the research issues identified above (see Table 7.1). As discussed, the strategies, self-handicapping and defensive expectations, are presented first because these are the central constructs in the study and it is considered that presentation of these data first provides a backdrop against which to explore the affective and motivational predictors and academic outcomes.

184

7.5.1 Academic strategies

7.5.1.1 Self-handicapping 7.5.1.1.1 Self-handicapping in students academic lives The self-handicapping behaviour reported by respondents selected as high in selfhandicapping ranged in extremity and illustrated the diverse ways in which it occurs in students day-to-day lives. When asked how he went about assignments and study for exams, Peter (HISH), who is barely passing his course, indicated that his selfhandicapping9 behaviour seemed to be beyond his control:

Its just like if I have an assignment due, I seem to just want to clean. I want to make my room clean, or I just watch TV or go out. Its just something that happens. You know youve got to do something but you get off track and go somewhere.

When asked what he would do during exam week, Peter (HISH) replied that exam week was party week. In fact at the end of the interview, when Peter (HISH) was asked if he would like to add anything about the way he went about his studies, he returned to the issue of self-handicapping. What his comments underscored was the way in which he seemed to use his untidiness as an excuse for why he does not do so well. His comments also illustrate the extent to which he feels little or no control over the state of his study area.

Sophie (HISH) actually had a special term for her self-handicapping. When asked whether she engaged in such behaviour she replied:

Pointless time wasting? Yeah. Ill say, Ive got study to do, well I really need to clean my wardrobe. Yeah, I relate to that. Im the queen of pointless time wasting.

At no stage were the terms self-handicapping, defensive expectations, or defensive pessimism used by the interviewer in the interviews. Descriptions of the behaviour rather than their labels were used.

185

Sophie (HISH) reported that when exams or assignments were approaching she would engage in activities as diverse as cleaning her wardrobe, tidying under her bed, cleaning the fish tank, and cleaning the garage. Sophie (HISH) would even visit her out-of-town grandmother who, curiously, only received visits when Sophies (HISH) exams were looming.

Carol (HISH) would be inclined to watch television and leave her study until very late in the night. Christine (HISH) was much the same. She reported that she had been having difficulty adjusting to academic demands and finds herself becoming occupied on other things, leaving study to the very last minute. Moreover, despite the fact that she would promise herself not to do this in the future, each time an exam or assignment approached, she seemed to do the same, no matter how hard I try. Rachel (HISH) said that when an exam approached she would engage in activities like alphabetising her books and videos.

When put that some students seem to place obstacles in the way of success, the students who were selected as scoring low on the self-handicapping scale generally reported that they did not identify with such behaviour. Tania (LOSH) was well aware of the potential distractions surrounding her while studying at home and so that these would not interfere with her studies, she went to the library to study. Reg (LOSH) said that he could not go out to nightclubs and parties and expect to stay focused on his studies and so he stayed at home. Gina (LOSH) reported that she was so inclined to get on top of her assignments and study that she had no social life. She said that she declined invitations to go out for fear that it would interfere with her study.

7.5.1.1.2 Reasons for self-handicapping and perceived (dis)advantages A range of reasons was given by the high self-handicappers as to why they engaged in such behaviour. Rachel (HISH) reported that if she failed she could usually identify something that interfered with her success such as going and getting pissed every night. Carol (HISH) suggested that her behaviour could be used as an explanation for poor performance. Similarly, Christine (HISH) was in little doubt that her behaviour could be used as an excuse if she did not do so well. She said that it was

186

always important to have an alibi in the form of an excuse. When Trudy (HISH) was asked why she self-handicapped, she also identified the importance of an excuse:

If I leave it [study] to the last minute, then Ive got an excuse if I didnt do well. Then I can say to myself, You left it to the last minute. Its your fault. Anything is better than saying that you didnt do well because you dont have the innate capability to do well. Any excuse is better than, Youre just not smart enough to do it. I know that I should be putting effort in all the time, but then Ive got the excuse if I dont go well.

Sophie (HISH) said that she engaged in self-handicapping behaviour because by not trying then the whole things not an issue, implying that a lack of effort may rob the event of its importance and with the importance of the event at a minimum, the threat of failure is not so acute. Peter (HISH) reported that there were no advantages to selfhandicapping, but then he went on to identify the fact that it was a comfort to him when he did not do so well knowing that he had an excuse available. On the other hand, Christine (HISH) said that she worked better when she left things to the last minute. She went further by suggesting that if she did do well under these circumstances she would feel more successful and that this was some evidence that I cant be that bad at it. Christine (HISH) also returned to her self-handicapping behaviour when asked at the end of the interview if she had anything to add:

CHRISTINE (HISH). Yeah, especially the bit about leaving things to the last minute . . . cause if you do well when you havent put much time into it, you think, Oh yeah! AM. And if you dont do so well? CHRISTINE (HISH). Then youve got an excuse . . . Its just easier to cope with if you think you havent put as much work into it. AM. Whats easier to cope with? CHRISTINE (HISH). From feeling like a failure because youre not good at it. Its easier to say, I failed cause I didnt put enough work into it than, I failed because Im not good at it. 187

When the low self-handicappers were asked why some students might self-handicap, they, for the most part, did not seem aware of the protective or enhancing possibilities that self-handicapping presented. For example, Angela (LOSH) reported that now that they were out of school with no-one to push them, some students were enjoying their new-found freedom and that this is why they self-handicapped. Reg (LOSH) reported that he did not really know why students would self-handicap and speculated that those students wanted to have a break from study and then tackle the assignment or exam fresh.

7.5.1.1.3 The perceived public and/or self-presented dimensions of selfhandicapping When asked about the extent to which others would be made aware of the handicaps set in place before exams or assignments, the self-handicappers indicated that their strategy was publicly oriented. This, however, was the case to varying degrees. Carol (HISH) reported that she would only tell people if she were asked about it while Sophie (HISH) reported that her family would know about it. Christine (HISH) said that she let people know about her self-imposed handicaps and that everyone else in her peer group did much the same thing. Peter (HISH) also let people know about his behaviour for apparently both protective and enhancing reasons: When he does not do so well, people tend to attribute it to the communicated obstacles, but when he does well, they are appropriately impressed: [If I fail] theyll be sympathetic and say, You shouldnt have gone out. Or theyll say, Oh you went out, sorry. Or theyll say, How did you do that! if you got a good mark. Trudy (HISH) actually exaggerated her self-handicapping to others: Now I lie about it. I lied the other day about it. When I was asked why I crammed, I said that I had to work the last three days and I didnt.

188

7.5.1.1.4 Summary of self-handicapping A summary of self-handicapping findings is presented in Table 7.3. The qualitative data in relation to high self-handicappers illustrated the diversity of ways in which students can strategically manoeuvre to protect themselves in performance scenarios. A number of students self-handicapped because it provided them with an alibi in the event of academic failure. On the other hand, students selected as being low in selfhandicapping negotiated their studies such that the chances of success were maximised rather than compromised. These students were also somewhat oblivious to the reasons why students would choose to self-handicap. The data also indicated the ways in which public and self-presented motivations underpinned high selfhandicappers tendency to self-handicap.

7.5.1.2 Defensive expectations It will be recalled that defensive pessimism has typically been conceptualised as comprising two components: defensive expectations and reflectivity. Of these two, the focal construct in the literature has been defensive expectations. Accordingly, students selected for the defensive pessimism component of the Time 1 qualitative study were selected on the basis of their defensive expectation scores and not their reflectivity scores.

7.5.1.2.1 Defensive expectations in students academic lives Students selected as being high in defensive expectations reported that no matter how well they had performed in previous exams and assignments, they expected to perform more poorly in the future. For example, despite the fact that she had never failed anything, Dianne (HIDE) reported that she always had the feeling she was going to fail. Robert (HIDE) cited an instance when a lecturer announced that a few students had not performed well in an assignment, and despite having passesd everything to date, he proceeded to spend the next 24 hours worrying that he was one of them (it turned out that he was not). He said thinking like this was like walking a tightrope between knowing that he has passed everything before but expecting that he will do more poorly in the future.

189

Table 7.3 Summary of findings in relation to Time 1 self-handicapping interview data

The issue
Perceived experience of self-handicapping

Summary of findings
High and low self-handicappers were clearly differentiated in terms of the strategies they use when going about their studies: All self-handicappers engaged in behaviour that impeded success whereas low self-handicappers made their study a priority and were focused on maximising their chances of success.

Perceived reasons for and advantages of selfhandicapping

Most high self-handicappers saw self-handicapping as providing an excuse in the event of poor performance. Low selfhandicappers, for the most part, were unsure as to why one would engage in self-handicapping behaviour.

Perceived public and/or self-presented dimensions of self-handicapping

There was some evidence for the public and self-presented dimensions to self-handicapping, but this was the case to varying degrees. Notwithstanding this, all self-handicappers seemed to be prepared to report the handicap if asked about their studies.

190

Cassie (HIDE) also reported that despite having performed well in the past she thought she would do worse in the next assignment or exam. She reported that each time she begins an assignment or studies for an exam, she starts from scratch in her expectations and feelings about how she will perform. She said that she would slowly build her confidence as she felt she was mastering the subject matter, but that no matter how confident she eventually feels about it, the next assessment task is met with the same pessimism.

Students selected as being low in defensive expectations reported that they were generally optimistic about future performance in assignments and exams. In contrast to the students high in defensive expectations, when asked how they would perform in the future given previous successful performance, these students replied that they would be optimistic. As Bernadette (LODE) reported, basically when Im on a roll, I feel good. I think, Well Ive got this mark and I can do it again.

7.5.1.2.2 Reasons for defensive expectations and perceived (dis)advantages Gwen (HIDE) cited cultural reasons for why she was defensively expectant. According to her, there was much pressure on her to perform well and that having lower expectations was a way of dealing with the fear of not meeting these expectations: Coming from an Asian10 family, if you get a good mark, theyll expect you to do the same thing again. Also, she has found in the past that when she is optimistic, these positive expectations are not fulfilled: In the past when Ive looked at things optimistically, something bad has happened and now I question myself all the time.

Joe (HIDE) reported that by holding unrealistically negative expectations and entertaining the possibility of not doing so well, he does not feel let down if those expectations are confirmed. Indeed, this disappointment seemed to be what most defensively expectant students were particularly concerned to protect themselves from. As Robert (HIDE) reported:

I try to be pessimistic cause that way I think the falls less when you

191

do actually come-a-cropper. . . I think if you border slightly on the pessimistic, then if I do better than I expected, then its a pleasant surprise, and if I do worse than expected, then its less of a fall. You just try to minimise those falls.

Marie (HIDE) saw mixed blessings in her defensive expectations. While she recognised that it probably was not so good for her self-esteem, in other respects it reduced the potential blow to the self-esteem by not thinking that youre going to reach the moon when you can only get as far as the clouds. Dianne (HIDE) added that if it turns out that she does perform well, then the advantage of having thought negatively is that she is pleasantly surprised. When Marie (HIDE) was asked at the end of the interview whether she had anything to add, she reported that her negativity was learnt at home: My parents have always said, Dont set your goals too high because youll only get disappointed . . . Theyre always careful not to raise my hopes so I dont get disappointed.

According to Cassie (HIDE), Robert (HIDE), and Joe (HIDE), thinking more negatively is also a good way of motivating them into action. Joes (HIDE), negative expectations motivated him into action by making him feel more stressed: If youre negative, you stress more and a bit of stress is good.

7.5.1.2.3 The perceived public and/or self-presented dimensions of defensive expectations Cassie (HIDE) reported that she lets her friends know that she is thinking negatively about upcoming performances because they tend to disagree with her, leaving her feeling more positive about herself. Gwen (HIDE) indicated much the same. When she tells her friends that she is feeling negative about an upcoming performance, her friends tend to offer her support and encouragement. Marie (HIDE) said that she also communicated her negative expectations to others and did so to avoid disappointing them. According to her, by communicating low expectations to others she hopes that they in turn have lower expectations of her. Marie (HIDE) also reported that if others hold negative expectations about how she will perform, they are less likely to be
10

The country is not specified to ensure anonymity.

192

disappointed in her if she does not do well. Gwen (HIDE) too, held this belief. She seemed to regulate her familys expectations (which she felt were onerous and culturally derived) of her by communicating to them her own negative expectations. A public and self-presentational motivation also underlay Joes (HIDE) motivation to tell others about his negative expectations. In fact, Joe (HIDE) said that he would exaggerate to others the extent to which he felt negative about his performance. After an exam he would tell his friends that he was going to fail even if he privately felt otherwise. According to him, he did this for socially-prescribed reasons of modesty, so that his friends did not say, Oh youre full of yourself.

7.5.1.2.4 Summary of defensive expectations Table 7.4 presents a summary of findings in relation to defensive expectations. There was a clear differentiation between expectations held by students high in defensive expectations and those low in defensive expectations. No matter how well, students high in defensive expectations had performed previously, they tended to be pessimistic about upcoming performances. For a number of students this was protective in the sense that in the event of failure, they would not be disappointed. Some students also indicated that their defensive expectations tended to motivate them. There also seemed to be some public element underpinning defensive expectations in that students high in defensive expectations communicated their expectations to others to either seek reassurance or to regulate others expectations of them. Additionally, a selfpresentational dimension underpinned some accounts in the sense that students would exaggerate their negative expectations to others for both protective and enhancing purposes.

193

Table 7.4 Summary of findings in relation to Time 1 defensive expectations interview data

The issue
Perceived experience of defensive expectations

Summary of findings
Reports about the experience of defensive expectations provided clear differentiation between students high and low in defensive expectations. All students high in defensive expectations reported that despite performing well in the past, they were pessimistic about their future performance. All students low in defensive expectations, on the other hand, reported that if they performed well in the past, they would expect similar if not better performance in the future.

Perceived reasons for and advantages of defensive expectations

Students high in defensive expectations reported that holding unrealistically low expectations minimised the risk of disappointment in the event of poor performance. Defensive expectations in some cases were also seen as a way of motivating oneself.

Perceived public and/or self-presented dimensions of defensive expectations

Some students high in defensive expectations were inclined to let others know about their negative expectations, and one actually exaggerated the negativity of his expectations. Some communicated their negative expectations to give their audience the opportunity to disagree with their negative expectations and to bolster their low confidence. Others did so to regulate others expectations of them and to avoid their disapproval.

194

7.5.2 Affective and motivational predictors Not only is the Time 1 qualitative study concerned with exploring self-handicapping and defensive expectations from students perspectives, it is also aimed at illuminating key quantitative findings derived from the Time 1 quantitative process model. Essentially this involves exploring students perceptions about the salience of factors underpinning performance orientation, external attributional orientation, uncertain personal control, and task-orientation in their academic lives and the idiosyncratic ways in which these factors are played out as these students go about their studies.

7.5.2.1 Performance orientation Performance orientation was found in the quantitative study to be positively associated with self-handicapping and defensive expectations. Performance orientation is comprised of esteem-relevant competence valuation, academic public selfconsciousness, ego-orientation, and avoidance-oriented performance.

7.5.2.1.1 Esteem-relevant competence valuation Esteem-relevant competence valuation refers to the extent to which individuals feelings about themselves are dependent on feeling competent (private esteemrelevant competence valuation) or being seen as competent (public esteem-relevant competence valuation). Most self-handicappers feelings about themselves were, to varying degrees, dependent on feeling competent and being seen as competent. For example, Carol (HISH), who had not performed particularly well in the previous semester saw competence as representative of her worth and for this reason experiencing and demonstrating competence was very important to her: I know people who dont go to uni and people see them as less of a person than people who do go to uni. Like, its stupid, but thats the way the world works. Christines (HISH) feelings about herself were also dependent on others seeing her as competent. In fact, she feels that her academic performance is dependent on others view of her: Because if everyone thinks that youre dumb and not doing well then youre probably not going to do as well.

Low self-handicappers feelings about themselves were not so bound up with feeling or demonstrating competence. Gina (LOSH) was not overly concerned with her 195

competence image because she was confident in herself and others views of her competence would not alter this. Lucy (LOSH) did not see competence as particularly important to her self-worth because, thats just one part of the person - the academic part. Tania (LOSH) was not concerned about her competence image because, Im doing what I want to be doing and thats the most important thing. Other LOSHs, whilst recognising the importance of competence, did not see it in terms of their selfesteem. Angela (LOSH) said that she would like to be seen as competent because others would know she was working hard and striving to improve herself. Similarly, while Sharon (LOSH) reported that she would like her peers to see her as competent, this was because competence reflected well on her application to her studies.

For most of those high in defensive expectations, feeling competent and being seen as competent were quite important. Cassie (HIDE) reported that while doing her course she has felt incompetent much of the time and has decided to return to nursing (her previous vocation) part-time because she knows she is good at it and through it will derive feelings of competence: I want to do something that I can say, Im good at this. It makes me feel good about myself. While Robert (HIDE) reported that he was not concerned with how his competence was viewed, he did not want to be seen as a goof-off. Maries (HIDE) feelings about herself were also dependent on her competence image. Perhaps this was not surprising given that she felt that others thought less of people who were not academically competent: The world views competent people as in a higher category than, say, people who dont go to university.

Students low in defensive expectations, whilst acknowledging that competence did have the potential to influence their views of themselves, indicated that competence did not have the esteem-relevant implications that it seemed to for students high in defensive expectations. For example, Lynne (LODE) said that while she liked to put her best foot forward academically, her feelings about herself were not determined by this and she did not dwell on it. Others, according to her, can think what they like. Tasha (LODE) too reported that she did not dwell on competence nor used it as the basis for her feelings about herself. Similarly, Amanda (LODE) reported that

196

incompetence was not a threat to her feelings about herself. According to her, Id probably be cross with myself, but I dont think I was less of a person for it.

7.5.2.1.2 Ego-orientation Ego-orientation refers to the extent to which individuals gain satisfaction and feel successful in outperforming others. High self-handicappers generally reported feeling successful outperforming others. Carol (HISH) and Christine (HISH) indicated that outperforming others would make them feel more successful than mastery of the material and that this was primarily because outperformance was something visible to other people whereas mastery was not. Trudy (HISH) reported that outperforming others makes her feel as though, Im not as down there as I thought - at least for a little while. She said that she always likes to know how close to the top of the class she is and when asked whether mastery or outperformance would make her feel most successful, she replied, Oh, the top of the class. Because thats how youre measured in society.

Some low self-handicappers also reported feeling successful when they outperformed others, but in a slightly different sense from the high self-handicappers. Gina (LOSH) was illustrative of this. While she was more interested in outperforming others, she did not conform to the stereotypical ego-oriented student in that she held some regard for those she outperformed: When asked how she felt when she outperformed others she replied, I feel proud, but I also feel sorry for whoever fails. Also, outperforming others was not enjoyed for the sake of beating others, but in confirming her positive self-concept: Its sort of the proof of the pudding . . . That Im right in being confident in my abilities. Angela (LOSH) also reported that outperforming others gave her satisfaction but considered that this was superficial success.

Two students were illustrative of how students high in defensive expectations come to place value on outperforming others. While Gwen (HIDE) reported that outperformance would make her feel most successful, she made the point that outperformance would mean more to her not because she beat others per se, but because, in true defensively expectant fashion, the relief of performing well would be great. And while Joe (HIDE) recognised that mastery was a more laudable goal, 197

outperformance would make him feel more successful primarily because university is a competitive environment and that beating others is consistent with this climate. He also reported that beating others is the best way to gauge how he is doing at university: If youre doing better than others then you know youre doing well and thats how you gauge yourself.

While reports by students low in defensive expectations also implied an egoorientation, this was qualified. For example, Bernadette (LODE) reported that she did enjoy outperforming others, but it was not so much to beat others, as to satisfy personal standards: Its more a sense of doing better for myself. Interestingly for Bernadette (LODE), an ego-orientation also held implications for mastery. She reported that she preferred to outperform others because thats indicative of being able to understand. While it was important to Lynne (LODE) to outperform others, this was because it makes her feel good about herself and not because others have not performed as well as her. Amanda (LODE) made the point that outperforming others is not cause for her to feel successful and reported that she saw competition as competition with herself and not with others: If I can grasp something by stretching my mind or worked really hard on something then Ive made progression in myself and Im proud of myself more so than being able to do something better than someone else.

7.5.2.1.3 Academic public self-consciousness Academic public self-consciousness refers to the extent to which students are concerned with how their academic performance is viewed by others. Trudy (HISH) was illustrative of self-handicappers high in public self-consciousness. She had withdrawn from university before and was in great fear of failing again because of what people would think of her if she did. Her family in particular is concerned with how she is performing and this, she says, contributes to her concern about how her performance is viewed. The extent of her public self-consciousness is reflected in the fact that when she receives results back in class she is so concerned with how others will see her react, she has actually asked to be excused from class so she can look at her result.

198

Low self-handicappers were less concerned with how they were viewed by others. For example, Tania (LOSH), who enjoys her course and is performing well, reported that she was not so much concerned with how she was viewed by others, but with her own personal standards and goals. According to her, I know what Im aiming for and other people might not be aiming for that same thing. If I get what Im aiming for then thats what matters. Similarly, Lucy (LOSH) reported that youve got to do it for yourself. The extent to which Gina (LOSH) felt others views were important to her was dependent on how they confirmed her positive view of herself: Its good to know that they agree with my view.

Defensively expectant students reported some concern with how they were viewed by others. For example, Gwen (HIDE) reported that how she was seen by others was important to her. Her responses also underscored a cultural influence: Respect from others, according to her, is important in her Asian culture and it is this need for respect that underlies the concern with how she is viewed by others. Other defensively expectant students, whilst not dominated by a concern about how they were viewed by others, did recognise that this concern was present in their lives. For example, Joe (HIDE) said that he would not want people to know if he failed university, because Id feel like I was a bit of a reject. Marie (HIDE) said she was concerned about how her old school friends would see her choice of course and her performance in that course. Her school friends had selected more scientific courses and she feared that they saw her choice of teacher education as a bludge and that she must not have achieved a very high Tertiary Entrance Rank to have elected to undertake this course.

Whilst some students low in defensive expectations were concerned with how they were seen by others, this concern was qualified. Amanda (LODE) and Brendan (LODE), whilst recognising that it would be nice that others saw them as performing well, reported that this did not dominate their thinking. Lynne (LODE) reported that other people could think what they liked about her (but did note that she would prefer that they did not think she was less than what she was). Bernadette (LODE) was more concerned about how her lecturers saw her academic performance because they would be more inclined to help if they saw her as being prepared to work hard.

199

7.5.2.1.4 Avoidance-oriented performance Avoidance-oriented performance refers to the extent to which students are motivated to perform in their studies out of a need to avoid the negative consequences of poor performance. Most of the students high in self-handicapping reported motivation to do their work more out of a desire to avoid the negative consequences of poor performance than a desire to strive for success. Two students were illustrative. Peter (HISH) was motivated to study because he did not want to repeat his course and because he did not want his family to think he was dumb. Trudy (HISH), who had already withdrawn from university once before, was motivated to avoid failure because the shame of withdrawing again would be extremely difficult for her to bear. A number of low self-handicappers, on the other hand, were inclined to be motivated more towards attaining success than to avoid poor performance and its negative consequences. For example, Gina (LOSH) said that she was more oriented towards success, because I dont think I will fail, while Tania (LOSH), who has been performing quite well and enjoys her course, was oriented towards success because this was indicative of her mastery of the course.

Most students high in defensive expectations were inclined to be motivated to avoid poor performance and its negative consequences. The reason Marie (HIDE) studied was to just get through. Robert (HIDE) actually saw disadvantages in striving for success: I know one person who pushed for the high marks and didnt get them and questioned their ability to undertake the course and why they took the course. Dianne (HIDE) acknowledged that fear that others would not see her as successful in the event of failure was a motivating force.

No students low in defensive expectations reported being motivated primarily out of a fear of poor performance and its negative consequences. Brendan, Amanda, and Lynne (LODEs), who have all performed well at university thus far, reported that they enjoyed the satisfaction that came with success and that it was this that motivated them in their studies. Interestingly, however, while Tasha (LODE) typically strived for success, she was now beginning to fear failure because she has not performed at university as well as she expected: Because Im not getting such good results, Im starting to think that I have to work hard to avoid failing. 200

7.5.2.1.5 Summary of performance orientation A summary of findings in relation to the performance orientation constructs is presented in Table 7.5. The findings for students high in self-handicapping and defensive expectations were broadly similar, as were the findings for students low in self-handicapping and defensive expectations. Students high in self-handicapping and defensive expectations were concerned with how others viewed their academic performance, however, it seemed that students high in defensive expectations were not so dominated by these concerns. Students high in self-handicapping and defensive expectations were more failure-oriented while students low in self-handicapping and defensive expectations were more success-oriented. For the most part, students high in self-handicapping and defensive expectations reported that their feelings about themselves were dependent on feeling and demonstrating competence. Generally, these students feelings of success also depended on outperforming others. On the other hand, for students low in defensive expectations and self-handicapping, their self-esteem was not so dependent on feeling or demonstrating competence. Moreover, whilst they acknowledged that outperforming others satisfied them, the extent to which this was the case was qualified.

7.5.2.2 External attributional orientation In the quantitative study, an external attributional orientation was positively associated with both self-handicapping and defensive expectations. The external attributional orientation construct is comprised of attributions for success and failure to external factors. In the case of success, cause is typically attributed to good luck, good teaching, or easy grading. In the case of failure, cause is typically attributed to bad luck, poor teaching, or difficult grading.

201

Table 7.5 Summary of performance orientation findings for students high and low in self-handicapping and defensive expectations

Self-handicapping
Public self-consciousness HISHs were concerned with how they were viewed by others, whereas LOSHs were more concerned with adhering to and striving towards personal standards

Defensive expectations
HIDEs reported a concern with others view of them but were not dominated by it. LODEs were also concerned with others view of them but this concern was qualified.

Avoidance-oriented performance

Most HISHs were motivated to perform out of a desire to avoid poor performance and its negative consequences rather than a desire to strive for success. LOSHs tended to be more success-oriented.

HIDEs were more motivated to avoid negative consequences of failure than they were success-oriented. One even cited the risks of successorientation. LODEs were more likely to strive for success, but one student who had performed more poorly than expected was now somewhat motivated to avoid poor performance and its negative consequences.

Esteem-relevant competence valuation

HISHs esteem was dependent on private and public competence, but to varying degrees. LOSHs feelings about themselves were less dependent on private and public competence.

HIDEs feelings about themselves were dependent on private and public competence. LODEs reported a preference for a high competence image but their self-esteem was not dependent on it.

Ego-orientation

HISHs feelings of success were dependent on outperforming others. LOSHs also reported satisfaction in outperforming others, but this was qualified.

HIDEs

feelings

of

success

were

somewhat

dependent

on

outperforming others. While some LODEs enjoyed outperforming others, this was qualified.

202

Interestingly, no self-handicappers attributed their previous success to external factors. Instead, most of the high self-handicappers tended to attribute their success to their ability. For example, Sophies (HISH) success was a combination of her ability and her cunning: Natural ability and a natural knowledge of the subject and I was basically able to bullshit my way through it. Carol (HISH) and Christine (HISH) believed success was due to a combination of hard work and ability, while Trudy (HISH) saw her success as coming through hard work. In terms of external attributions for failure, Carol (HISH) was the only self-handicapper who attributed failure to external reasons, citing bad luck as the reason. The others mainly attributed their failure to insufficient effort and in some cases, low ability. For example, Trudy (HISH) saw failure as due to a combination of low effort and low ability in the sense that she had neither been insightful enough nor taken the time to ascertain what the examiner expected of her.

The majority of students high in defensive expectations attributed their past success to internal factors. Cassie (HIDE) for example, reported that she did well when she thought highly of herself. According to her, if Im confident in myself, I can put things down on paper or talk. Robert (HIDE) attributed his success to his ability. In respect to failure, he cited instances where he has invested the effort, but because he has not clicked into it, has not performed well. Other students high in defensive expectations saw success and failure as due to a mix of internal and external factors. For example, Dianne (HIDE) reported that it would come down to the fact that the teaching in that subject would be clear and interesting and probably I put a bit more effort into it. In terms of attributions for previous poor performance, no students high in defensive expectations attributed poor performance to external factors. The majority (Cassie, Dianne, Joe, Marie) of students high in defensive expectations attributed their past poor performances to insufficient effort.

Thus interview data relating to attributions for success and failure (see Table 7.6 for a summary) do not reflect patterns parallel to those that emerged in the quantitative data. Rather than attributing failure to external factors, self-handicappers and students high in defensive expectations tended to attribute failure mainly to effort and partly to ability. This may in part be due to the differential nature of the quantitative and 203

qualitative methods: The nature of the questionnaire was such that respondents were able to endorse both ability and external attributions for failure. In the qualitative component, however, respondents were asked for the main reason for past performances and so were really only given the opportunity to provide just one reason for failure. Alternatively, the interview situation may have prompted some selfserving bias such that students high in self-handicapping and defensive expectations were unwilling to cite low ability as the reason for past failure and were somewhat more inclined to attribute past success to ability. These and other interpretations are discussed more fully in the following chapter.

7.5.2.3 Uncertain personal control Uncertain personal control is comprised of attributions for failure to ability (addressed above), perceived control over ones ability to avoid doing poorly in the future, and instability of self-concept. Uncertain personal control was found in the quantitative analyses to be positively associated with self-handicapping and defensive expectations . 7.5.2.3.1 Control over avoiding future failure Generally, high self-handicappers reported that they felt little control over their ability to avoid poor performance in the future. For example, Christines (HISH) experience is that hard work does not necessarily produce a good result and so when thinking about upcoming exams or assignments, she says: You never know whats going to happen . . . you can work a lot harder at it, but still I find it so hard to judge how theyre going to be marked and what they expect. Peter (HISH) also felt little control over his ability to avoid future poor performance: I dont know whats going to happen. Trudy (HISH) reported experience with inconsistent marking on

assignments that she feels were equal in quality and so feels she has little control over her ability to avoid poor performance in the future.

No students low in self-handicapping reported diminished control over their ability to avoid doing poorly in the future. Angela (LOSH) reported that through effort she enhances her control: I know personally, to avoid doing poorly most of the time - 85 percent of the time - I will put a lot of work in and then I can control it. Gina (LOSH) reported that she knew exactly what she needed to do in order to avoid doing 204

poorly in the future and her comments also implied effort. When asked how much control she felt, she replied: I think a lot as long as I go back to my tutor or lecturer and ask what I should have done and review the resources I used and find out what I left out.

While HIDEs generally did not outrightly report low control, they did imply that the extent to which they could avoid doing poorly in the future was dependent on factors beyond their control. For example, Dianne (HIDE) and Marie (HIDE) reported that their control was dependent on the extent to which the lecturers were able to point out clearly where they had gone wrong. Robert (HIDE) acknowledged that success and failure were ultimately up to him, however, he still doubted his ability to perform in the future. Moreover, while Joe (HIDE) reported that he felt he did have control over his ability to avoid doing poorly in the future, whether or not I use that control is another matter.

In contrast to the students high in defensive expectations, students low in defensive expectations reported relatively high control over their capacity to avoid future failure. Bernadette (LODE) reported that she was confident in her ability to avoid future failure and knew what steps to take to enhance her control. Lynne (LODE) reported that because she was dedicated and worked hard, she had control over avoiding poor performance in the future. Tony (LODE) indicated that by applying himself before an assignment is due he enhances his control over future outcomes.

7.5.2.3.2 Instability of self-concept No high self-handicappers reported that their view of themselves as students was particularly consistent. For the most part, their opinion of themselves as students tended to fluctuate. Rachel (HISH) was illustrative:

With essays, Ill have my mind on it and then Ill start to do it and then it just goes out the window and Ill screw it up and then Ill get down in the dumps. And then something else will happen and it [her view of herself] will change again.

205

No students low in self-handicapping reported an unstable self-concept. They tended to be more consistent in their self-concept. For example, according to Gina, if I get results back and theyre not good, Im not shattered. I just think I havent put enough into it, Ive got to do more. Im still confident.

As with the high self-handicappers, no students high in defensive expectations reported that their view of themselves was particularly consistent. Cassies (HIDE) opinion of herself as a student ranged from really good to really stupid. The extent to which Gwens (HIDE) opinion of herself fluctuated would depend on the subject and how much she enjoyed that subject. In subjects she enjoyed she was inclined to think highly of herself, whereas she thought less of herself in subjects she disliked.

Students low in defensive expectations communicated a more consistent and stable self-concept. In parts, however, this stability was dependent on a variety of factors. For example, Lynne (LODE) reported that her self-concept varied from subject to subject and also in response to the grades she received: I guess when you get results back, they change your opinion. You might hand something in and not think its that good, but you might get a good mark and I think, Oh, I do have an OK point of view.

7.5.2.3.3 Summary of uncertain personal control A summary of the interview data pertaining to uncertain personal control is presented in Table 7.6. These data indicate that most high self-handicappers felt little control over their ability to avoid future failure and illustrated the diverse ways in which this lack of control played out in their academic lives. While students high in defensive expectations also perceived low control, it was not as marked as that of HISHs and the evidence was not without some qualifications. Students high in self-handicapping and defensive expectations also reported that their view of themselves was not particularly consistent and that it was likely to fluctuate in response to a variety of events in their academic lives.

7.5.2.4 Task-orientation

206

Task-orientation refers to the extent to which students feel successful and gain satisfaction through mastery, learning, and solving problems. In the Time 1 quantitative study, mastery orientation was found to negatively predict both selfhandicapping and defensive expectations.

Self-handicappers were asked which of mastery or outperformance would make them feel more successful. Trudy (HISH) was actually dismissive of a task-orientation: According to her, no-ones going to care if you learn something new - maybe you can solve world peace by it or something . . . It doesnt matter that you learnt great things getting those points - thats not what unis about. While Sophie (HISH) indicated that solving a challenging problem was rewarding, she was mindful of how successful performance follows mastery. Peter and Rachel (HISHs), on the other hand, were more clearly task-oriented. Peter (HISH) reported that gaining new skills was important to him because when you get a new skill, it stays with you, while Rachel (HISH) reported that I like to know why everything works the way it does.

By contrast, all low self-handicappers endorsed mastery and recognised the contribution of mastery to their academic lives. According to Lucy (LOSH), mastery is something in yourself that you can feel good about rather than outperforming others and bagging them out. Reg (LOSH) reported gaining great satisfaction in just doing or knowing something that youve never done before and getting the hang of it and being able to do something well. Tania (LOSH) saw the long term benefits of mastery: It doesnt matter that you went better than everyone in a first year test. Its best if you can learn something that you use in the future to transfer knowledge to students.

While students high in defensive expectations endorsed mastery, in many parts, this endorsement was qualified. For example, Marie (HIDE) reported that mastery was more important to her because one needed mastery in order to beat others. Cassie (HIDE) and Dianne (HIDE) reported that mastery was important to them but more because it was difficult to compare their performance with that of others doing different subjects. Similarly, Cassie (HIDE) reported that because everyone studied

207

different subjects at university, competition with them was difficult. She therefore tended to focus on mastery.

Students low in defensive expectations were more unambiguously concerned with mastery. Three students were illustrative. Bernadette (LODE) reported that she loved challenge and the feeling that she has learnt something new. According to Brendan (LODE), university is not a place of competition, its more about learning. Amanda (LODE) agreed: If I can grasp something by stretching my mind or work really hard on something, then Ive made progression in myself and Im proud of myself, more so than being able to do something better than someone else.

The findings relating to task-orientation (summarised in Table 7.6) are broadly similar for students high in self-handicapping and defensive expectations: Both groups tended not to value mastery as much as outperforming others. Notwithstanding this, students high in defensive expectations were a little more likely to value mastery than selfhandicappers. Students low in self-handicapping and defensive expectations were a good deal more likely to value mastery, some citing the longer-term importance of mastery and the fact that university is about learning rather than competition.

7.5.3 Academic outcomes The final component of the model involves the academic outcomes that follow from the strategies, self-handicapping and defensive expectations. Two academic outcomes that were found in the Time 1 quantitative analysis to be predicted by these strategies are self-regulation and persistence.

208

Table 7.6 Summary of external attributional orientation, uncertain personal control, and task-orientation for students high and low in selfhandicapping and defensive expectations Self-handicapping
External attributional orientation Contrary to quantitative findings, HISHs did not attribute success or failure to external factors. Rather, HISHs attributed success to ability and effort, while LOSHs attributed success mainly to effort. Uncertain personal control - Ability failure attributions Few HISHs tended to attribute failure to ability - most attributed failure to insufficient effort. Few HIDEs attributed failure to a lack of ability. The majority attributed failure to insufficient effort.

Defensive expectations
Few HIDEs attributed success and failure to external factors. Most reported a mix of external and internal factors. For LODEs, success and failure were only partly due to external factors (see ability attributions for failure below).

- Low control over future failure

Generally, HISHs reported low control over their ability to avoid failure in the future. No LOSHs reported low control over their ability to avoid future failure.

HIDEs, whilst not entirely doubtful of their ability to avoid future failure, did imply that the extent to which they could avoid failure was not within their control. No LODEs reported low control.

- Unstable self-concept

HISHs self-concept tended to fluctuate, but not to a marked degree. LOSHs self-concept seemed more consistent.

No HIDEs reported that their view of themselves was particularly consistent but in some respects this was qualified. Whilst LODEs reported a more stable self-concept, under some circumstances it would fluctuate.

While HISHs expressed a preference for performance, mastery Task-orientation was important to them. LOSHs were more likely to endorse taskorientation than ego-orientation.

HIDEs reported valuing mastery, but this was qualified. LODEs were unambiguously task-oriented.

209

7.5.3.1 Self-regulation The qualitative data indicated that self-handicappers were not inclined to be selfregulators. Three students were illustrative. Rachel (HISH) found that while she might set goals, she would never stick to them and she reported that she now never plans her approach, does not set goals, nor monitors her progress: If I set goals, I never reach them. I get distracted. If I set a specific goal, I know the way I work; Ill end up looking at something completely different. Peter (HISH) went to the library only a few days before an assignment was due and he did not make any plans about how to go about the assignment. Nor would he set any short term goals. Rather, the goal was just to finish the assignment. When asked if he had any study plans for the week before exams, he said that he did not because that was party week. Sophie (HISH) has found that the timetables she organises for study week are just another example of how she goes about wasting time before exams. According to her, planning a timetable is a brilliant time-waster.

By contrast, all the students low in self-handicapping reported self-regulatory behaviour. Gina (LOSH) constantly revised her assignments, retyping them, and checking that she has addressed all the relevant issues. Lucy (LOSH) breaks the questions into smaller components, carefully addresses each point, and monitors how she is addressing each point. Sharon (LOSH) also develops a plan about how she is going to go about her assignment and keeps track of where she is up to and whether she is working within the time limits she has set. Reg (LOSH) was also organised about how he went about his exams and assignments, setting time limits, monitoring his progress, and setting goals. Tania (LOSH) was much the same in that she carefully planns her approach to her essays, sets goals she wants to work towards, and constantly revisits the question to ensure she is on track.

Students high in defensive expectations tended not to self-regulate or were selfregulators to only a minimal extent. Dianne (HIDE) had given up on attempts to plan or organise her study: Ive tried to set goals and say, Ill have so much done by this time, but it never happens so I dont do it anymore - theres just no point. Because Joe (HIDE) has found that he never sticks to his plans, he would set plans that were so loose that there was little risk of not adhering to them. He also found that this 210

approach was usually of no practical use and that the few days before the assignment or exams were due was when he attended to them. Other students high in defensive expectations were not self-regulators, but indicated that were it not for outside commitments they would be. Gwen (HIDE), for example, found that because of her heavy family commitments she could not really set goals or plan her study. Robert (HIDE), who was juggling two jobs and study, reported that he would like to plan and organise his study, but in practice, he was only able to do his study and assignments as they were due without much forethought, planning, or goal-setting.

Respondents low in defensive expectations tended to be self-regulators. Amanda (LODE) and Tasha (LODE) keep notes on how they plan to go about their assignments. Amanda (LODE) reported that she always monitors how she is addressing the essay or exam question. Similarly, Bernadette (LODE) reported that she was very systematic about her study and assignments, dividing the question or task into a set of minor tasks or issues, and carefully addressing each. Lynne (LODE) indicated that she was very organised about her approach to her study: I organise a little timetable for myself. I have every half hour accounted for and Id set my alarm so I wouldnt go over time.

7.5.3.2 Persistence For the most part, high self-handicappers were not inclined to persist at challenging tasks. Carol (HISH), Christine (HISH), and Peter (HISH) were open about the fact that they would be more inclined to abandon challenging tasks. Christine (HISH) said she got very frustrated if she could not solve a problem immediately. Peter (HISH) said that if he did not solve a problem in ten to fifteen minutes, he would ask for help. Sophie (HISH) reported that she was unlikely to persist with things which were of no consequence to her future: Id give up a lot easier than most people, just because Id think, This is futile. Im not going to lock myself in a room 24 hours a day, seven days a week trying to master something that wont really have a huge bearing on my future. Despite the fact that Rachel (HISH) saw herself as a persister, she reported that her persistence was dependent on the extent to which she was interested in the task or issue at hand.

211

The students low in self-handicapping, on the other hand, tended to be persisters. Gina (LOSH) reported that, Ill always work until I get it. Other low self-handicappers reported that persistence was important for their understanding of subsequent course content. When Angela (LOSH) was asked to what extent she would work at challenging tasks or problems, she replied: Id work it out. You have to. It was probably given to you for a reason. Maybe if you dont get that bit then you wont get the next bit. Tania (LOSH) said that persistence pays off for her because she finds she enjoys arriving at a solution and the knowledge that she gains through it.

Students high in defensive expectations tended to be low persisters. Marie (HIDE) reported that she would jump straight in and write the first thing that comes to mind. Other students low in defensive expectations were not high persisters and the extent to which they persisted depended on a variety of factors. For example, Joe (HIDE) and Dianne (HIDE) said that their persistence depended on how much they were interested in the issue at hand. According to Dianne (HIDE), if it interested me, then I can sit at something for hours trying to figure it out. But if something bores me, I just write the first thing that comes to mind.

On the other hand, students low in defensive expectations were clearly more inclined to persist in the face of challenge. Two students were illustrative. Amanda (LODE) and Bernadette (LODE) tended to work at challenging tasks for some time, leave the problem for a while, then come back to it with a fresh perspective. Also, while away from the problem, Bernadette (LODE) would read up on the target issue.

7.5.3.3 Summary of outcomes The interview data (see Table 7.7 for a summary) showed that students high in selfhandicapping and defensive expectations were not inclined to be self-regulators nor would they persist in the face of challenge. The data also extended current understanding of the consequences of self-handicapping and defensive expectations by showing how self-handicappers and students high in defensive expectations come to abandon self-regulatory behaviour and persistence and opt for more superficial, short term, and limited engagement at academic tasks. Important to note, however, is that whilst self-handicappers and defensively expectant students share much in terms of 212

lower self-regulation and persistence, these shared characteristics are qualified in the sense that the negative consequences of defensive manoeuvring did not appear to be so marked for the defensively expectant students as they were for the self-handicappers. For example, whilst self-handicappers were clearly not inclined to persist or selfregulate, some students high in defensive expectations would persist to the extent that the task interested them or would self-regulate if their work or family commitments were not so pressing.

7.6 Summary of chapter It was considered that the phenomena under study lent themselves quite well to qualitative analysis. The Time 1 qualitative study explored students personal insights into self-handicapping and defensive expectations, extending current understanding of defensive manoeuvring in general and also the nature of self-handicapping and defensive expectations more specifically. The Time 1 qualitative data also indicated the idiosyncratic ways in which students engage these strategies, the salience of these strategies in students lives, the strategic nature of self-handicapping and defensive expectations in the context of self-worth protection, the various factors that accompany students defensive posturing, and the academic outcomes that are associated with the strategies.

Notwithstanding the variety of ways in which self-handicapping and defensive expectations are similar (the most notable similarity being their function in protecting self-worth), the data illuminated key ways - albeit subtle - in which the two strategies are distinct (for example, in contrast to self-handicapping, defensive expectations did not appear to be so inimical to academic outcomes such as self-regulation and persistence).

213

Table 7.7 Summary of self-regulation and persistence findings for students high and low in self-handicapping and defensive expectations

Self-handicapping
Self-regulation HISHs reported that they tended not to engage in self-regulatory behaviour. For them, there was little evidence of planning, setting goals, or monitoring progress. All LOSHs, on the other hand, reported self-regulatory behaviour.

Defensive expectations
Most HIDEs were not self-regulators or engaged in self-regulatory behaviour to only a limited extent. HIDEs tended to be self-regulators.

Persistence

HISHs were not inclined to persist in the face of challenge. LOSHs tended to persist at challenging tasks and commented on the importance of persistence in the context of their overall learning.

HIDEs were not particularly inclined to persist but some did persist to the extent that the task interested them. LODEs were more inclined to persist at challenging tasks.

214

In sum, then, the Time 1 qualitative data have been important in (a) illuminating the Time 1 quantitative findings, (b) underscoring the marked similarities between selfhandicapping and defensively expectant students, but (c) also profiling the ways in which students high in self-handicapping and defensive expectations do diverge, and (d) superimposing on the central quantitative process model the very human face that constitutes the fabric of the present study.

215

8. DISCUSSION OF TIME 1 RESULTS

The central issues explored at Time 1 concern the affective and motivational factors that predict the use of self-handicapping, defensive expectations, and reflectivity and the impact of these three strategies on academic outcomes such as self-regulation and persistence. In the review of literature it was proposed that self-handicapping and defensive expectations (and to a lesser degree, reflectivity) are strategies primarily designed to protect individuals self-worth in anticipation of poor performance and thus the broad conceptual orientation adopted is one drawing on self-worth motivation theory. Indeed, the affective and motivational factors proposed to underlie the three strategies are argued to be factors that, to varying degrees, underpin the motive to protect the self.

The data clarified the relative salience of affective and motivational factors in predicting self-handicapping, defensive expectations, and reflectivity. The strongest predictor of self-handicapping was an external attributional orientation and to a lesser, but significant extent, uncertain personal control, low mastery orientation, and high performance orientation. The strongest predictor of defensive expectations was uncertain personal control, followed by an external attributional orientation, low taskorientation, and high performance orientation. The strongest predictor of reflectivity was also uncertain personal control, followed by high task-orientation, a performance orientation, and to a lesser extent, high self-concept. Neither self-concept nor beliefs about the nature of intelligence significantly predicted self-handicapping or defensive expectations. Both self-handicapping and defensive expectations negatively predicted self-regulation and persistence, while reflectivity positively predicted these outcomes. In addition to clarifying relationships proposed in the central model, data also provided insights into a number of other issues concerning (a) the nature of the relationship (and, indeed, the distinction) between reflectivity and defensive expectations, (b) the higher order factor structure of self-handicapping and defensive expectations comprising both active and self-presented dimensions, and (c) the domain specificity of the relevant constructs.

8.1 What factors predict the use of the three strategies? 216

8.1.1 Task-orientation The quantitative data showed that task-orientation is inversely associated with selfhandicapping and defensive expectations and positively associated with reflectivity. Task-orientation is an important construct in the model primarily because it reflects an adaptive orientation and thus provides an important contrast to the other somewhat less adaptive affective and motivational predictors. In the review of literature it was proposed that task-oriented individuals have a mastery focus and tend to see success as deriving from effort rather than from ability (Duda et al., 1992; Middleton & Midgley, 1997; Skaalvik, 1997b; Treasure & Roberts, 1994; Walling & Duda, 1995). Consequently, failure tends not to be attributed to low ability and so from a self-worth motivation perspective, task-oriented individuals are not particularly vulnerable to threats to their self-worth. It can be further proposed, then, that because of this, taskoriented students are not particularly inclined to engage in self-protective strategies in anticipation of poor performance. According to Plant and Ryan (1985), taskorientation does not carry the self-worth implications that ego-orientation does. Taskorientation, according to them, is different from ego-orientation because the taskoriented individuals motivation to perform an activity is derived from its intrinsic properties rather than an investment of self-esteem (p. 440).

Research to date, however, has provided a rather mixed set of findings pertaining to task-related factors. While Midgley and Urdan (1995) showed that school task focus and task goals were negatively correlated with self-handicapping scores, in more recent research Urdan et al. (1998) found no significant relationship between class task goals and self-handicapping. The present findings, then, may lend clarity to the area in supporting the findings of Midgley and Urdan (1995). Notwithstanding this, it is recognised that climate goals (e.g., class-level or school-level task goals) are different from individual motivation orientations (such as that assessed in this study).

The quantitative and qualitative task-related data also make a unique direct contribution to the area of defensive pessimism. Both sets of results showed that taskoriented students are less inclined to hold defensive expectations. Again, it is contended that because task-oriented students see academic outcomes more in terms 217

of effort than ability, they are rendered less vulnerable in upcoming performance situations from a self-worth motivation perspective. Also, they do not feel compelled to cushion the blow of failure through defensive expectations because they see failure more as diagnostic feedback than reflecting poorly on their self-worth. That mastery does not underpin defensive expectations is congruent with the fact that defensively expectant students are focused on performance outcomes and how they expect to fare in upcoming achievement tasks. Moreover, whilst previous research has shown that defensively expectant students expend effort which leads to successful performance in the short term, the present data show that this effort expenditure does not arise from an underlying task-orientation. Rather, the effort expenditure is proposed to arise from an underlying anxiety in a bid to enhance control (Norem & Cantor, 1986a, 1986b). Indeed, to explore this further, anxiety is incorporated into the study at Time 2. Thus the effort expenditure of defensively expectant students is not grounded in an adaptive task-orientation, but rather, is grounded in a more failure-avoidant orientation. This may be successful in the short term, but in the longer term is argued to be detrimental to performance.

The quantitative data showed that individuals high in task-orientation were also more inclined to engage in reflectivity. This may not be surprising given the facts that reflectivity is very much a self-regulatory strategy (as is discussed more fully below) and that a task-orientation is consistent with self-regulation in the sense that selfregulation is very much effort-oriented. Moreover, reflectivity is about gaining control and control is enhanced through the application of effort - a theme consistent with task-orientation. The finding that task-orientation positively predicted reflectivity also provides some insight into how defensive expectations and reflectivity are distinct. While the two are proposed to be self-protective in the sense that they are to varying degrees underpinned by a motive to avoid failure, one is a relatively more adaptive means of addressing failure avoidance while the other may be more

counterproductive. Specifically, while reflectivity is grounded in a proactive, effortoriented motivation orientation (high task-orientation), defensive expectations are inversely predicted by this proactive, effort-oriented dimension. The distinction between defensive expectations and reflectivity is revisited later in this chapter and expanded in the general discussion in Chapter 13. 218

Taken together, the findings in relation to task-orientation represent an important positive focus in the model. Pivotal to an understanding of the relationship between task-orientation and the three strategies are the facts that task-orientation is an effortoriented construct and that academic outcomes are seen to be very much effort-related. Poor performance for task-oriented students is therefore less threatening because it tends not to hold ability-related implications for self-worth (see also Covington & Omelich, 1979b). Consequently, task-oriented students are less likely to engage in defensive manoeuvring in the forms of self-handicapping and defensive expectations which, in the present study were found to be inimical to effort-related constructs such as self-regulation and persistence.

8.1.2 External attributional orientation An external attributional orientation was a key factor underpinning the need to protect the self in the form of self-handicapping, and to a lesser extent, defensive expectations. This finding runs counter to Weiners (1985) claim that external attributions for positive or negative outcomes do not influence feelings about the self (p. 560 - it is recognised, however, that the present operationalisation is different from Weiners operationalisation of attributions and also that the present data showed that an external attributional orientation was almost unrelated to level of self-concept). The quantitative data suggest that attributions to external causes do evoke implications for the self-worth to the extent that individuals who perceive that external factors strongly contribute to their failure or success are also inclined to self-handicap. Indeed, it can be said, then, that external attributions influence feelings about the self to such a degree that the individual engages in strategies found in the present study to be quite counterproductive in an educational sense. This is also consistent with contentions by Covington (1984a) that an external attributional orientation is demoralizing and is likely to impact on individuals inclination to protect the self-worth.

Not so expected was the fact that an external attributional orientation significantly predicted defensive expectations when previous work has shown that defensive pessimists are inclined to accept responsibility for outcomes (Cantor & Norem, 1989; Showers, 1988; but note the Showers, 1988, finding that defensive pessimists are 219

higher in external locus of control). Given that defensive pessimism is about gaining control (Norem & Cantor, 1986a, 1986b), the present data may reflect defensively expectant students response to perceptions of low control that are grounded in the belief that academic outcomes are due to external causes. Alternatively, it may be that perceiving success and failure to be beyond ones control raises the prospect of academic failure and setting defensive expectations is a cushioning strategy in anticipation of such failure. This latter proposition is more consistent with the selfworth motivation perspective adopted in the present study.

The qualitative data, however, did not parallel the quantitative findings. Few interviewees high in self-handicapping or defensive expectations attributed success and failure to external outcomes. Rather, they tended to attribute failure to insufficient effort and success to ability. This may be because these attributions are the least adaptive, implying helplessness, and students primarily concerned with self-protection may not be inclined to concede to their application, particularly in a face-to-face interview (see Juvonen & Murdock, 1993). The discrepancy between quantitative and qualitative findings may also partly be an artefact of the methodology: The nature of the questionnaire made it possible to provide high ratings to both external and internal causes of academic outcomes, whereas interviewees were asked to report the main reason for their past successes and failures. Thus interviewees high in selfhandicapping or defensive expectations may have endorsed a number of theoreticallycompeting attributions in the questionnaire, but when asked in the qualitative study to identify the main reason for their success and failure, opted for internal rather than external attributions. In view of this, the focus of Time 2 interviews was broadened to include questions about both internal and external attributions rather than the most salient one.

The qualitative findings were also somewhat consistent with research demonstrating (a) self-serving bias in attributions in which students attribute success more to ability than to other factors and (b) self-protection in which students attribute failure to insufficient effort (e.g., see Skaalvik, 1990). These self-serving and self-protective tendencies are proposed to be particularly salient for individuals in an interview situation in which it is not uncommon for respondents to manage the impression they 220

convey to an interviewer (Juvonen & Murdock, 1993). In the context of the present study, students (particularly those high in self-handicapping and defensive expectations) tended to attribute failure to causes that minimised the implications such failure held for their ability and attributed success to factors that maximised the favourable ability-related implications of performance. On the other hand, when responding to a questionnaire administered on anonymous and group bases, students were not so concerned with attributing in self-serving and self-protective ways and were clearly prepared to concede that past failure, for example, was at least partly due to ability (see Section 8.1.3.3).

8.1.3 Uncertain personal control Uncertain personal control (comprising low control over future failure, ability attributions for failure, and unstable self-concept) was the strongest predictor of defensive expectations and reflectivity, and to a lesser extent, self-handicapping.

8.1.3.1 Low control over avoiding future failure The data in relation to low control over future failure and defensive expectations are important because to date their association has not been empirically examined directly. Central to the theory of defensive pessimism is that defensive pessimists feel out of control and their defensive expectations and reflectivity are a means of playing through a variety of outcomes which enhances their cognitive and affective control over such outcomes. Indeed, the fact that the broader construct, uncertain personal control, was the strongest predictor of defensive expectations and reflectivity provides support for this component of defensive pessimism theory. The data are also consistent with less centrally-related findings involving control and expectations. For example, Pyszczynski (1982) reported that uncertainty about an outcome elicits a motivation to protect the self which can be achieved by underestimating ones chances of obtaining the outcome. Also, Schunk (1983) reported that to avoid the implications of failure, students low in perceived control set unrealistically low expectancies for upcoming performances (see also Albersnagel et al., 1986). Similarly, Diener and Dweck (1980) demonstrated that helpless children (who perceive uncertain control over avoiding future failure) were more likely to underestimate their performance in upcoming tasks. Taken together, these studies, in conjunction with the present data, 221

show that low control over avoiding future failure is an important contributor to selfprotection generally and defensive expectations more specifically.

The interview data showed that low control over avoiding future failure was often due to self-handicappers perceptions that examiners were inconsistent in assessment of their work. As a consequence, these students saw little relationship between their effort and subsequent academic outcomes. This is consistent with previous research demonstrating that individuals who receive reinforcement they perceive to be independent of their effort perceive low control over their ability to achieve desired outcomes and avoid undesired ones (Perry & Dickens, 1984; Skinner et al., 1985). This type of feedback has been referred to as non-contingent feedback. Noncontingent feedback communicates to students that success and failure are not due to their personal agency and leads to a belief that control over their ability to avoid failure in the future is uncertain. The present data, particularly the interview data, add to the body of research which shows that low perceived control over future outcomes often evoked through non-contingent feedback - gives rise to self-handicapping (Berglas, 1987; Berglas & Jones, 1978; Higgins & Harris, 1988; Perry et al., 1986; Perry & Dickens, 1984; Rhodewalt & Davison, 1986; Riggs, 1992).

8.1.3.2 Instability of self-concept The quantitative data support the hypothesised association between an unstable selfconcept and both defensive expectations and self-handicapping. The qualitative findings were parallel in that respondents high in self-handicapping and defensive expectations reported that their self-concept was not particularly consistent from one time to another. Individuals whose self-concept is not stable, it is contended, hold a shaky belief in their competence, confidence, and skill, and as a consequence their perceived capacity to meet upcoming tasks is also shaky (see Arkin & Oleson, 1998). On such occasions, the perceived likelihood of failure is increased and in response to this, it is argued, students are inclined to take preemptive protective action to mitigate the implications of this failure.

Consistent with this rationale, Harris and Snyder (1986; see also Riggs, 1992) report that individuals unstable in self-esteem are plagued by doubts of performing 222

successfully in an esteem-threatening, evaluative situation (p. 451). In response to these doubts, individuals are inclined to self-handicap prior to a given task such that if the failure they suspect is impending does in fact occur, they have an alibi (Kimble, Funk, & DaPolito, 1990; McFarlin & Blascovich, 1981; Schneider & Turkat, 1975). The present data support this proposition.

In terms of self-concept stability and defensive expectations, the research literature is slim. In view of this, interpreting the present findings rests on the more general relationship between self-concept instability and self-protective behaviour. Kernis and Waschull (1995) have reported that self-worth is always on the line for individuals unstable in self-concept. According to them, self-esteem instability is accompanied by a heightened concern about feelings of self-worth, which permeates individuals involvement in everyday experiences (p. 103). In terms of the present study, a disproportionate concern with the self-worth seems to underpin unrealistic negativity. Indeed, Kernis et al. (1993, in Kernis & Waschull, 1995) found that individuals unstable in self-esteem, relative to individuals with a more stable self-esteem, perceived negative events to last longer and viewed these events more unfavourably.

It seems then, that it may not only be ones level of self-esteem that is critical in determining the extent to which one self-protects, but also the stability of these selfappraisals. This is consistent with Berglass (1987) contentions in relation to selfhandicapping and with speculation by Norem (personal communication, October 30, 1996) in relation to defensive pessimism. Moreover, whilst some have suggested that this instability may interact with the level of self-concept to influence self-protection (Kernis et al., 1989; Kernis et al., 1992), Time 1 data suggested that this is not the case. Rather, instability seems to operate as a main effect in predicting defensive expectations, reflectivity, and self-handicapping via its higher order construct, uncertain personal control.

8.1.3.3 Ability attributions for failure The quantitative findings in relation to ability attributions for failure are consistent with various studies which show that individuals who attribute failure to ability are more inclined to engage in defensive manoeuvring designed to protect the self-worth 223

(Covington & Omelich, 1979a; Weiner, 1985). Covington and Omelich (1984a, 1984b, 1985) suggested that ability attributions for failure are what render individuals particularly vulnerable to self-worth devaluation. They suggest that of all the attributions, attribution of negative outcomes to ability is the dominant cognition in the sense that shame and humiliation are more closely associated with it than failure attributed to anything else (see also Ames, 1984) and also because ability is directly instrumental in bringing about success (see Covington & Beery, 1976; Covington & Omelich, 1984b). Moreover, Weiner (1994) predicted that individuals attributing failure to ability are more likely to experience shame and embarrassment and the fact that self-handicapping and defensive expectations were underpinned by ability attributions for failure may suggest a link between them and these affective reactions.

The quantitative studys ability attribution findings are also consistent with work by Cantor and Norem (1989) and Showers (1988) who have found that defensive pessimists are prepared to accept responsibility for outcomes. Similarly, Diener and Dweck (1980) found that helpless children (high in ability attributions for failure) tended to underestimate the number of problems successfully solved. As Weiner (1985), conjectures, if conditions (the presence or absence of causes) are expected to remain the same, then the outcome(s) experienced in the past will be expected to recur (p. 556). Given that ability is seen to be relatively stable, the individual attributing failure to this cause can be predicted to hold negative expectations for future performance and the present data reflect this in relation to defensive expectations.

8.1.3.4 Summary of uncertain personal control Taken together, the findings in relation to uncertain personal control contribute to current understanding about the factors underpinning defensive expectations, reflectivity, and self-handicapping. Uncertain personal control reflects individuals uncertainty regarding their ability to perform. When students are uncertain as to their ability to perform, the perceived threat of failure is increased as is a consequent selfworth vulnerability. In response to this vulnerability, the present data show that students set defensive expectations and to a lesser extent engage in reflectivity and self-handicapping. Identifying uncertain personal control as the most salient factor 224

predicting defensive expectations also provides important information for educational practitioners and researchers seeking to address this strategy from an intervention perspective. This issue is pursued in further detail in Chapter 13.

8.1.4 Performance orientation The findings in relation to performance orientation show that the motives to perform based on a need to beat others, to appear competent, or to avoid poor performance and its negative consequences, underpin self-handicapping, defensive expectations, and reflectivity.

8.1.4.1 Ego-orientation It was predicted that ego-orientation would underlie self-handicapping for two reasons. First, because ego-oriented individuals see success as ability-based rather than effort-based (Boyd & Callaghan, 1994; Skaalvik, 1997b), failure is particularly threatening to their self-worth because it implies a lack of ability rather than a lack of effort. Second, because ego-oriented students are concerned with outperforming others, failure in this context is particularly ignominious and damaging to their competence image and consequent self-worth.

The data in relation to self-handicapping support not only these two contentions but also previous empirical work. In the educational domain, Midgley and Urdan (1995) found that a school performance focus and extrinsic goals (hypothesised here to be ego-related dimensions) were associated with self-handicapping and that in a final regression of significant first order predictors, extrinsic goal motivation remained a significant contributor to self-handicapping (see also Midgley et al., 1996). In more sophisticated analyses using multilevel modelling, Urdan et al. (1998) showed classlevel ability goals were significantly associated with self-handicapping (the distinction between class-level and individual-level goals is recognised, but the ego-relatedness was deemed relevant to the present discussion). There are also indirect data to support the present findings. Pyszczynski and Greenberg (1983) found that when performing tasks characterised as being high in ego-relevance and low in success probability, participants were less inclined to invest effort. In the sporting domain, Rhodewalt and colleagues (1994) found that self-handicappers were more inclined to practice less 225

(self-handicap) before important (ego-relevant) sporting meets. The present data, then, add to these studies and confirm the hypothesised link between ego-orientation and self-handicapping.

The quantitative and qualitative data also showed that ego-orientation underlay a tendency to engage in defensive expectations. This too, supports previous research, which, whilst not centrally related to defensive expectations, do deal with the issue of negative expectations. For example, Koestner et al. (1987) and Duda (1992) found that ego-involved individuals tend to choose less difficult tasks than task-involved individuals. Newton and Duda (1993) reported that ego-oriented participants entertain more negative expectations about performance in upcoming sporting activities. Similarly, Diener and Dweck (1978) found that children adopting a performance orientation were more likely to focus on failure and made causal attributions for this failure in much the same way as defensive pessimists by attributing it to low ability. Again, whilst the evidence in these studies is indirect, they do support the present data in demonstrating a relationship between ego-related dimensions and negative expectations/orientations.

The role of ego-orientation may also clarify recent conceptualising regarding approach and avoidance goals. Elliot and Church (1997; see also Elliot & Harackiewicz, 1996; Skaalvik et al., 1994) proposed that relative to task-orientation and avoidance orientation, ego-orientation is a complex construct. Indeed this is evident in the way it has been found to be associated with adaptive (see Elliot & Church, 1997; Harackiewicz et al., 1997; Midgley & Urdan, 1995; Roeser et al., 1996; Skaalvik, 1997b) and not so adaptive (see Ames & Archer, 1988; Duda & Nicholls, 1992) educational outcomes. Elliot and Church (see also Elliot & Harackiewicz) suggest that the nature of the educational challenge determines whether ego-orientation yields adaptive or not so adaptive outcomes. In achievement situations in which there is the threat of failure, it may be that an ego-orientation is more akin to avoidance-oriented performance leading to negative outcomes such as self-handicapping. In achievement situations that are more a challenge than a threat, ego-orientation may be more arousing and adaptive. According to Elliot and Church, performance approach goals [e.g., ego-orientation] can be quite deceptive in that the same phenotypic regulatory 226

form can represent diverse genotypic motivational tendencies (p. 228). In the context of the present study, given that the university is a competitive climate (Harackiewicz et al., 1998) and the maths domain is one that arouses anxiety (Bessant, 1995; Pajares & Urdan, 1996; Schneider & Nevid, 1993; Vance & Watson, 1994), it is proposed that such a climate is somewhat threatening to the student and so ego-orientation reflects more of an avoidance orientation. Hence it loaded strongly on the same higher order factor as avoidance-oriented performance (also see the .56 correlation between performance-approach goals and performance avoidance goals in the Middleton & Midgley, 1997, study; but see the weak association between the two in Elliot & Harackiewicz, 1996) and through this higher order factor was positively associated with self-handicapping and defensive expectations.

8.1.4.2 Avoidance-oriented performance Quantitative and qualitative findings also support the hypothesised association between avoidance-oriented performance and both self-handicapping and defensive expectations. Avoidance-oriented performers are motivated by a fear of poor performance and its negative consequences. Because poor performance represents evidence of low ability and hence low self-worth (from a self-worth motivation perspective), such individuals are more inclined to negotiate performance scenarios in a protective manner.

Atkinson and Feather (1966) were amongst the first to examine the issue of fear of poor performance and its negative consequences as well as the strategies individuals use in response to these fears. According to them, the individual who fears failure opposes any and all sources of positive motivation to undertake the customary competitive activities of life (p. 14). In terms of this study, self-handicappers oppose these motivations by engaging in activities that actually diminish the chances of success. Later work confirmed Atkinson and Feathers analysis. For example, Solomon and Rothbaum (1984) reported that contributing to the tendency to procrastinate was individuals fear of failure represented by anxiety in meeting their own and others expectations and a lack of self-confidence.

227

Atkinson and Feather (1966) also commented on individuals levels of aspiration (expectations) in response to a fear of poor performance and its negative consequences. They cite an example of the individual who exhibits a dramatic decrease in his level of aspiration, a retreat to the safest of ventures (p. 370) in response to a fear of poor performance (see also Baumgardner & Brownlee, 1987). Teevan and Smith (1975) later reported that individuals who are motivated out of a fear of failure tend to set goals that reduce the possibility of failure (see also FriedBuchalter, 1992; Tseng & Carter, 1970). Passer (1983) reported that those high in trait anxiety (arising from, it is argued by Solomon & Rothbaum, 1984, fear of failure and negative consequences), relative to those low in trait anxiety, expected to perform less well, worried more frequently about performing poorly, and expected more negative evaluation following failure. In an Australian context, Feather (1967) found that students with a motive to avoid failure were more likely to regulate (strategically lower or raise) their expectations in response to high performance variability. Taken together, this indirect evidence supports the relationship between an avoidanceoriented performance and defensive expectations derived in the present study.

8.1.4.3 Public self-consciousness The association between public self-consciousness and self-handicapping supports present contentions that self-handicapping arises partly from a motivation to protect the public competence image. The qualitative data also reflected high selfhandicappers concern with how they were viewed by others. Low self-handicappers, on the other hand, were more interested in adhering to and attaining personal standards. There has been some debate as to whether self-handicapping is primarily a private or a public strategy. Two pieces of evidence in the present study suggest that it very much is a publicly-oriented strategy. First, self-presented self-handicapping was just as salient as active self-handicapping in loading on a higher order selfhandicapping factor (this is discussed below). Second, public self-consciousness was found to underpin self-handicapping via the higher order performance orientation factor. Thus, the present data clarify the public dimension of self-handicapping and lend support to a small body of work dealing with the issue. Kolditz and Arkin (1982) found that study participants presented with an insoluble problem tended to selfhandicap when the experimenter was present and when they believed the experimenter 228

would have access to their score. Strube (1986; see also Ferrari, 1992; Schill et al., 1991) found that self-handicapping correlated with public self-consciousness, social anxiety, and other-directedness. Midgley and Urdan (1995) also found public selfconsciousness to be associated with self-handicapping, but that the effect was not particularly strong, as was also the case in the present study. Shepperd and Arkin (1989a) found that subjects high in public self-consciousness were more inclined to choose a self-handicap when performing a task which was characterised as a valid indicant of academic success. These findings, then, and those of the present study, are consistent with Baumeister and Schers (1988) conclusion that public attention to the self is associated with self-destructive behaviours.

Consistent with hypotheses and previous research, a concern with how one is viewed and appears to others also underlies a tendency to hold defensive expectations. Polak and Prokop (1989) found that individuals concerned with their public selves in a protective sense were more inclined to be defensive pessimists. Similarly, research investigating the related issue of worrying has found that worriers are higher in public self-consciousness (Pruzinsky & Borkovec, 1990). The qualitative data were also illustrative of a public dimension underlying defensive expectations in the sense that defensively expectant students were particularly concerned with others expectations of them and that defensive expectations were a means of regulating others expectations. For example, one student felt that her parents expectations of her were onerous and communicating her own negative expectations to them was an attempt to have them lower their expectations of her and indeed, minimised the chances of them being disappointed in her.

8.1.4.4 Esteem-relevant competence valuation When individuals self-worth is contingent on competence, they have as a high priority the need to protect this competence (Beery, 1975; Covington, 1984a, 1992; Covington & Beery, 1976). Consistent with hypotheses, the extent to which individuals feelings about themselves were dependent on their experience or demonstration of competence underlay the tendency to self-handicap and form defensive expectations. According to Beery (1975), when individuals feel they are not able to meet the competence standard required to maintain their self-worth, they tend 229

to have an excuse in the ready, to make sure that their limitations in performance do not reflect on their ability (p. 200). Self-handicapping is a direct means of establishing such an alibi. Forming defensive expectations is another a priori strategy which can mitigate the implications of failure: Setting lower and safer standards minimises the chances of failure and cushions the subsequent blow to the private and public competence image that would occur in the event of such failure.

8.1.4.5 Summary of performance orientation In sum, data in relation to performance orientation contribute to current understanding about the factors underpinning self-handicapping, defensive expectations, and reflectivity. But despite the fact that it positively predicted self-handicapping and defensive expectations, it cannot be concluded that it is necessarily inimical to adaptive functioning because it also positively predicted reflectivity. This is suggestive of the rather complex nature of performance orientation. Moreover, while performance orientation represents a focus on performance, the basis of this focus is not grounded in the most ideal of motives. Specifically, despite the fact that performance-oriented students are driven to perform, this is motivated out of a need to outperform others, avoid poor performance and its negative consequences, or be cast in the best possible light in the public domain.

8.1.5 Level of self-concept Reflectivity was the only strategy significantly predicted by level of self-concept. It has been shown that individuals higher in self-concept are more inclined to manoeuvre in ways that maximise their control and enhance mastery of their environment (Bandura, 1997). It has also been found that when confronted with obstacles, individuals low in self-concept prematurely relinquish efforts to manage the problem whereas individuals high in perceived competence are more likely to master such obstacles (e.g., Bandura & Cervone, 1983; Cervone & Peake, 1986; Peake & Cervone, 1989). Reflectivity, it is argued, is one way of attempting to master such obstacles: By thinking through potential outcomes, individuals gain control (Norem & Illingworth, 1993) and are better placed to successfully meet academic challenges presented to them. The present data, then, identify the strategic means by which individuals high in self-concept seize control. 230

The present study operationalised self-appraisals using domain-specific self-concept rather than global self-esteem. In the review of literature, however, it was suggested that predictions in relation to self-concept could be inferred from the self-esteem literature. To date, the evidence regarding the relationship between self-appraisals and self-handicapping has been mixed. Three competing predictions have been advanced. First, individuals low in self-concept are more likely to entertain the possibility of failure and therefore engage in self-protective strategies. Indeed, it has been found that self-protection is a central concern of individuals low in self-concept (Baumeister et al., 1989) and recent work has shown that perceived competence is negatively associated with self-handicapping (Urdan et al., 1998). Indeed, extrapolation from evidence relating to self-efficacy (Bandura, 1982, 1986) would suggest that individuals low in self-efficacy are more inclined to self-handicap. The second hypothesis is that failure for individuals high in self-concept would be a greater blow to their self-concept than failure for individuals low in self-concept and so it could be predicted that the former group would self-handicap to protect themselves in such a situation.

The present data support the third possibility which holds that self-handicapping is relatively independent of the individuals self-concept. Indeed, this is consistent with Rhodewalt et al. (1991) who found that discounting attributions (protective selfhandicapping) were made independent of individuals self-esteem (see also Midgley et al., 1996, in relation to positive self-esteem; Rhodewalt & Fairfield, 1991; Rhodewalt & Hill, 1995). Moreover, Berglas and Jones (1978) argued, consistent with present findings, that it is not so much level of self-concept that influences the tendency to self-handicap, as the stability of that self-concept.

Whilst it was hypothesised that self-concept would be inversely associated with defensive expectations, there was no significant relationship between the two in the process model. Indeed, given that (a) the correlation between self-concept and defensive expectations was negative, (b) previous research has found that defensive pessimists are lower in self-esteem (Eronen et al., 1998; Norem & Cantor, 1986b), and (c) no interviewees high in defensive expectations in the qualitative study reported a 231

high self-concept, the non-significant effect is somewhat surprising. Perhaps this is a function of the difference between global self-esteem and domain-specific selfconcept. Global self-esteem may be more associated with a pessimistic orientation whilst domain-specific self-concept may be more directly associated with academic outcomes such as future academic plans and persistence (as was demonstrated in the Time 1 process model). It seems, then, that further work is needed to explore the differential effects of global self-esteem and domain-specific self-concept not only in relation to defensive expectations but also self-handicapping.

8.1.6 Beliefs about the nature of intelligence Contrary to hypotheses, beliefs about the nature of intelligence did not significantly predict self-handicapping or defensive expectations. Two reasons are proposed to account for this. First, in theoretical discussions of entity and incremental beliefs, it has been implied that their effects may be mediated by students motivation orientations (Dweck, 1991; Dweck & Leggett, 1988). Indeed, Rhodewalt (1994) found that self-handicapping was positively associated with a performance orientation and contended that this was the case because of performance-oriented participants view that intelligence is a fixed and immutable entity. On the other hand, a learning orientation was negatively associated with self-handicapping and Rhodewalt suggested this was because learning-oriented participants saw intelligence as incremental and that an investment of effort led to enhanced performance. It may be, then, that the incorporation of ego- and task-orientation as predictors in the study explains variance in self-handicapping and defensive expectations that intelligence beliefs otherwise would. This rationale is supported in part by the fact that entity beliefs about the nature of intelligence are positively correlated with self-handicapping and defensive expectations (see Table 6.17) but yield no significant predictive effect in the central process model (see Figure 6.2) once statistically controlled for by the presence of other predictors such as ego-orientation.

The second possible reason for the non-significant role of intelligence beliefs relates to the nature of the item wording. The intelligence-beliefs items were worded such that they reflected beliefs about other students (e.g., Some students wont be smart no matter what) rather than about the respondent him or herself. It may be that in the 232

context of a model centrally self-related, beliefs about other students are neither salient nor relevant. It is recommended that further research be conducted that examines the influence of students beliefs about their own intelligence and the role of these beliefs in a model along the lines of that explored in the present study. At any rate, students beliefs about the nature of intelligence was not a significant factor in predicting any of the three strategies. Accordingly, it was dropped not only from subsequent estimations of the Time 1 process model but is also dropped from the instrument at Time 2.

8.2 To what extent do self-handicapping, defensive expectations, and reflectivity predict academic outcomes? While the data have highlighted the negative consequences that follow from selfhandicapping and defensive expectations, they also show that reflectivity is associated with more adaptive academic outcomes. Moreover, the finding that self-handicapping was more strongly negatively associated with academic behaviours than defensive expectations indicates, consistent with previous contentions, that self-handicapping is the least adaptive of the self-protection strategies. Thus the findings identify the diverse ways in which self-protection can affect academic behaviours and also provide insights into the important distinctions between defensive expectations, selfhandicapping, and reflectivity.

8.2.1 Self-handicapping, self-regulation, and persistence As hypothesised, the quantitative data indicated that self-handicapping negatively predicts self-regulation and persistence. In terms of self-regulation, the qualitative findings showed that high self-handicappers tended not to set goals, did not plan their approach to study, nor monitored their progress. Low self-handicappers, on the other hand, reported that when given an assignment or exam question, they divided the question into manageable units, constantly revised what they had written to ensure its relevance, and monitored their progress as they went about their study. In terms of persistence, the qualitative data were illustrative of how self-handicappers tended to readily abandon tasks in the face of challenge.

233

These findings are consistent with Covington (1984a) who argued that ultimately selfhandicapping tactics are not successful in terms of academic outcomes. It seems that students who tend to place obstacles in the path to success are also inclined to follow this through with low persistence and low self-regulation. If one is looking for an alibi and excuse for failure, a lack of persistence (effort) is a classic candidate. If failure occurs in the face of persistence, it cannot be attributed to low effort and so ones ability remains in question. Persistence implies effort, failure following effort implies low ability (Covington & Omelich, 1979b), and low ability, according to the selfworth theory of motivation, implies low self-worth (Covington, 1984a, 1992). Much the same can be said for self-regulation. Individuals who are inclined to self-handicap do so to deflect responsibility away from their ability and onto less threatening factors such as effort. If individuals engage in self-regulation it follows that they are expending effort. If failure occurs in this context, individuals cannot attribute it to low effort and so there is the risk that failure will be attributed to low ability. It therefore follows that individuals who are inclined to self-handicap are unlikely to be selfregulators.

These findings, however, are contrary to some previous research. A number of investigators have found that self-handicapping can have adaptive consequences in the sense that it minimises anxiety (Leary, 1986), enhances self-esteem (Leary, 1986) and performance (Snyder et al., 1981), and increases task involvement and consequent task enjoyment (Deppe & Harackiewicz, 1996). Similarly, Thompson (1993) found that when given a face-saving opportunity, the performance of self-worth protective study participants improved. The findings in the present study may run counter to this line of evidence because of the way in which self-handicapping was operationalised. The focus in this study is on self-protective self-handicapping and not on selfenhancing self-handicapping. In fact, consistent with Midgley et al. (1996) and Urdan et al. (1998), the present study explicitly invoked the self-protective dimension through item wording (. . . so I have an excuse if I dont do as well as I hoped). It is hypothesised that because of this explicitly self-protective orientation - as opposed to self-enhancing orientation - the chances it would lead to adaptive outcomes are slim.

8.2.2 Defensive expectations, self-regulation, and persistence 234

Consistent with hypotheses, quantitative and qualitative data also confirmed a significant negative association between defensive expectations and self-regulation. Similarly, survey and interview data indicated that students high in defensive expectations were not inclined to persist with tasks in the face of challenge. These findings are contrary to much research on defensive pessimism. For example, Garcia et al. (1995) found that defensive pessimists tend to increase effort and because of this their performance is subsequently unimpaired. Norem and Cantor (1986b) argue that defensive pessimism is a motivating force (p. 1209), while Cantor and Norem (1989) argue that defensive pessimism motivates continued persistence (p. 93). Showers and Ruben (1990) argue that in contrast to real pessimists withdrawal of effort, defensive pessimists show extensive effort (p. 387). Showers (1992; see also Cantor & Norem, 1989) also noted the strategic increase in effort that accompanies defensive pessimism.

Notwithstanding this rather consistent line of evidence, the point proposed here to be crucial in interpreting the present findings is the distinction between reflectivity and defensive expectations. It is further proposed that it is not defensive pessimism as a whole which lets students down, but the expectation component (as opposed to the reflectivity component) which brings about reduced persistence and self-regulation. Thus the present findings may reflect the importance of maintaining a distinction between reflectivity and defensive expectations: Reflectivity was associated with high self-regulation and persistence while defensive expectations inversely predicted these outcomes.

Indeed, this is consistent with Goodhart (1986) who found that thinking negatively resulted in enhanced performance whereas negative expectations resulted in poor performance. Goodhart suggested that individuals who form expectations become committed to the level of performance they predict they will attain and exert only the amount of effort required to attain this level (p. 122). Similarly, Swann (1987; see also Swann et al., 1989) has argued that individuals are inclined to behave in ways that verify their negative self-views and the negative effects of defensive expectations may be consistent with this. Sherman et al. (1981) also found that those who entertained (rather than expected) negative outcomes tended to perform well, whereas 235

those who actually formed negative expectations were inclined to perform more poorly (see also Campbell & Fairey, 1985). Showers (1992) also argued that the effect of thinking on performance may be determined by the type of thinking that occurs. According to her, it may be expectations (defensive expectations) which are not adaptive and not simply entertaining possibilities (reflectivity). Showers (1992) found that thinking negatively (rather than holding negative expectations) for defensive pessimists did not bring about maladaptive outcomes in a social situation. Rather, it resulted in more social engagement and more positive ratings by others.

Furthermore, in the longer term it has been found that defensive pessimists performance tends to decline. Norem and Cantor (1990b) found that in a three-year follow-up study, defensive pessimists performed more poorly than optimists and aschematics, and also experienced more global life stress, more psychological symptoms, and less life satisfaction. Also, Eronen et al. (1998) found that while there was a significant association between defensive pessimism and achievement at Time 1, this was not the case at Time 2. Possible reasons for the negative longer term effects have been proposed (Norem & Cantor, 1990b), however, one explanation which has not been invoked is the fact that over time, the expectations component of defensive pessimism may come to be more salient than the reflectivity component. That is, these expectations over time may take precedence over the thinking-through component and that it is these expectations, rather than the thinking-through which elicits these negative outcomes. The present study supports this thesis in that defensive expectations were associated with negative outcomes whereas reflectivity was associated with academic outcomes in quite the opposite way. At any rate, in the light of the present findings, it appears that negative outcomes arising from defensive expectations are not necessarily restricted to just the longer term.

8.2.3 Reflectivity, self-regulation, and persistence The finding that reflectivity, or the thinking-through process, was found to be positively associated with self-regulation and persistence is consistent with a good deal of research and theorising about defensive pessimism. Norem and Illingworth (1993) found that when asked to list their thoughts (engage in reflectivity), defensive pessimists reported feeling more positive and that when this reflectivity was disrupted, 236

they performed more poorly. In fact, Norem and Illingworths discussion of defensive pessimism seemed to imply that it is the reflectivity component of defensive pessimism which is pivotal to successful performance: Through this process [reflectivity], defensive pessimists feel better, feel less anxious and more in control, and their performance should thus be less likely to be disrupted by anxiety (p. 823). Indeed, as noted above, Showers (1992) outlined the important differences between simply thinking about an outcome and actually expecting it (see also Campbell & Fairey, 1985; Goodhart, 1986; Sherman et al., 1981). It seems, then, that the reflectivity component of defensive pessimism is the component which enables the student to take control through self-regulation and persistence and that it is the expectations component of defensive pessimism which is inversely related to these outcomes.

The fact that reflectivity was most strongly associated with self-regulation (the second strongest path in the model) may suggest how it is that individuals high in reflectivity take control: They are more inclined to plan their approach, set goals, monitor the extent to which they are meeting these goals, and make the appropriate adjustments if necessary. Indeed, Norem and Illingworth (1993) proposed that the thinking-through process may enhance problem-solving processes and that by asking students to engage in reflectivity, we may have facilitated self-regulation (p. 833). Consistent with this, Showers (1988) proposed that when we engage in how thinking (proposed here to be akin to reflectivity), we emphasize specific, concrete actions and events leading up to an outcome (p. 227) - processes very much akin to self-regulation. The students who hold defensive expectations, on the other hand, tend to engage in what appears to be avoidance behaviour such that they are less likely to self-regulate and are less likely to persist in the face of challenge.

8.3 Active and self-presented self-handicapping and defensive expectations The quantitative and qualitative data provided important insights into the active and self-presented dimensions of self-handicapping and defensive expectations,

particularly in terms of the extent to which active forms of self-handicapping and defensive expectations are so highly related to the public/self-presented forms. To date, the active and self-presented dimensions of self-handicapping and defensive 237

expectations have not been directly examined in questionnaire research and the present data show that the distinction between them is not great. Indeed, on this basis, they were incorporated into higher order self-handicapping and defensive expectations factors.

The public and self-presented dimension of self-handicapping is supported by a rather abundant research literature. Kolditz and Arkin (1982) were amongst the first to propose that self-handicapping is very much an impression management strategy (see also Tice & Baumeister, 1990). Later, Strube (1986; see also Ferrari, 1992; Schill et al., 1991) found that public self-consciousness was correlated with self-handicapping. Baumgardner et al. (1985; see also Shepperd & Arkin, 1989a) also suggested that selfhandicapping is very much a strategy designed to protect ones public image. In addition to this, the interview data showed that self-handicappers were inclined to communicate the handicaps to others, particularly those close to them such as their family. This is consistent with Arkin and Oleson (1998) who proposed that the primary audience of self-handicappers is comprised of the significant others in their lives. Taken together, these findings provide a basis for the unity of active and selfpresented self-handicapping and suggest that it is defensible to propose that the two contribute to the one overarching self-handicapping construct.

While Covington (1992) discussed students crafty tendency to communicate unrealistically low expectations to others, to date, there appears to be no research conducted to explore the association between active and self-presented defensive expectations. In this sense, the study contributes to current understanding in the area. There is, however, some indirect evidence to assist interpretations. Illingworth and Norem (1991) found that defensive pessimism significantly correlated with fear of negative evaluation and this implies an other-directedness on the part of the defensive pessimist. Polak and Prokop (1989) found that self-monitoring (particularly protective self-monitoring) was associated with defensive pessimism. Individuals who selfmonitor, they contended, are concerned with the regulation and control of their public selves in the service of creating the desired social impression (p. 285). Indeed, the qualitative data indicated that students high in defensive expectations were concerned about (downwardly) regulating others expectations of them so that the 238

chances of disappointing them were minimised. Moreover, in his discussion of selfverification theory, Swann (1987) outlined how individuals are inclined to behave in ways that not only verify their own self-view but also control how others see them.

On a more general level, it has been argued that a major underpinning of a fear of failure and its negative consequences is a concern with how one is viewed by others. In early work on fear of failure, Birney et al. (1969) identified three forms of fear of failure, one of which, was fear of social devaluation (or fear of a threat to ones public image). Given that the use of self-handicapping and defensive expectations are argued to arise partly from this fear of public devaluation, it is not surprising that selfpresented dimensions of these strategies load on a higher order factor also comprising active dimensions. Indeed, in the qualitative study, respondents high in selfhandicapping and defensive expectations communicated a need to protect themselves from others disapproval and disappointment.

The fact that the self-presented dimension of self-handicapping and defensive expectations was just as salient as the active dimension also underscores the strategic nature of self-handicapping and defensive expectations. The congruence between active and self-presented self-handicapping and defensive expectations highlights the fact that when self-handicapping and defensively expectant students go about their studies they must negotiate not only the private terrain that represents their personal standards but also the public terrain in a bid to alter the meaning of potential failure in the eyes of a perceived audience.

8.4 Domain specificity of constructs The data collected from the first administration of the questionnaire (Nepean, n=204) clarified the way in which the constructs central to the study operated in math and verbal academic domains. This constitutes an important contribution not only to literature in self-handicapping, defensive expectations, and reflectivity, but also in terms of other constructs in the model which occupy salient roles in a good deal of empirical research in educational, personality, and social psychology. From a measurement perspective, the domain-specificity analyses also contribute to current

239

understanding of confirmatory factor analysis and the way in which correlated uniquenesses for parallel questionnaire items can be addressed.

With one notable exception, the math-verbal dimensions of the constructs were highly correlated and indicated domain generality rather than specificity. Consistent with the Internal/External Frame of Reference Model (I/E Model) of math-verbal self-concept (see Marsh, 1990, for a review) which posits that math and verbal self-concept should be orthogonal, the one exception was self-concept. Thus, whilst the development of self-concept may rely on a frame of reference, the other constructs seem to be more general, particularly within the academic context.

In some ways this domain specificity is surprising because a number of these measures are, to varying degrees, associated with ones level of self-concept. Also, given that affect underlies many of the present constructs and that this is also the case for self-concept (Marsh, 1988b), it might have been expected that content specificity as found for self-concept, would be evident in at least some of the other constructs. The present results were also surprising given that previous research identified in the review of literature has shown some academic content specificity for a number of constructs (e.g., attributions: Marsh, 1984; Marsh et al., 1984; mastery orientation: Galloway et al., 1996; Gottfried, 1982; implicit theories: Stipek & Gralinski, 1996) found in the present study to evince little specificity.

In other respects, however, the domain specificity is not so surprising. Firstly, whilst some research has explored the constructs on an intra-academic basis (e.g., math or verbal, and in some cases, math vs verbal), for the most part, these constructs have typically been operationalised as domain-general measures. For example, motivation orientation has been proposed to operate and develop in schoolwork in general, rather than in maths or verbal domains in particular. Where domain specificity has been examined in relation to motivation orientation, it is usually across more diverse domains such as academic and sporting settings (Duda & Nicholls, 1992). A second point to consider is that the I/E Model is able to make predictions about math and verbal self-concept because self-concept is so heavily based on competence, whereas

240

the other constructs presented here, whilst to varying degrees dependent on competence, are also very much underpinned by other factors.

There is of course also the possibility of method bias in the study. The format of the first questionnaire was such that to each item students responded to a math column and a verbal column. Arising from this may have been some level of automaticity in responses giving rise to identical responses to both math and verbal domains. Yet, if there was something systematic in students responses, it certainly was not manifest in the self-concept measure. The math-verbal correlation for this construct was precisely as would be predicted by the I/E Model. Moreover, the fact that items for each of the constructs were randomly distributed throughout the questionnaire obviates the potential for method bias in the form of item-block responding.

Not only have the data contributed to current understanding of the domain specificity/generality of a variety of constructs employed in educational, personality, and social psychological research, they have also raised important methodological issues. The upward bias in correlations due to parallel item uniqueness was demonstrated in this component of the study such that in a number of cases correlations were above unit value. Improper correlations were attenuated by estimating the correlations between parallel item uniquenesses. Given this and given that parallel item formats in questionnaires are not uncommon in psychometric research (e.g., self-other agreement studies and longitudinal research - Marsh & Hau, 1996), these results underscore the importance of employing statistical techniques that are able to estimate (a) uniquenesses associated with questionnaire items and (b) the correlations between these uniquenesses (Marsh, 1993a, 1993c; Marsh et al., 1997). Essentially, this involves the application of CFA procedures and these procedures were used throughout the remainder of the study. Models without correlated uniquenesses did not fit the data well, and, perhaps more importantly from a substantive perspective, produced systematic biases.

A number of implications for future research and educational practice arise from the findings in the domain-specificity component of the study. First, it may be that interventions designed to promote more adaptive motivation styles and strategies may 241

operate more generally rather than for specific academic subjects. For example, promoting task-oriented, rather than competitive approaches, may be best conducted on a general academic basis. Second, and distinct from the first point, interventions designed to promote a more positive academic self-concept, must recognise its multidimensionality. Third, results demonstrate how upward bias in correlations can occur across two subscales utilising parallel wording. Following from this, the present results underscore the importance of CFA, particularly as it enables correlation of uniquenesses, which controls for the upward bias in inter-domain correlations.

8.5 Time 1 qualitative analyses Qualitative data were collected that illuminated the ways in which the Time 1 quantitative findings were manifested in students academic lives from students personal perspectives. While a good deal of these findings have been addressed in the discussion thus far, it is considered important to focus more closely on selfhandicapping and defensive expectations because a number of new and interesting insights emerged from the interviews about these strategies.

8.5.1 Self-handicapping The qualitative data add to existing knowledge of the idiosyncratic ways in which students self-handicap as well as the various lengths to which students will go to protect their self-worth. What these data also underscore is that there is no shortage of potential handicaps at these students disposal. Moreover, it is clear that whilst selfhandicapping can constitute quite obvious forms of avoidance behaviour wherein students overtly establish an alibi before the fact, it is reflected also in circumstances over which the student appears to have little control. The data in relation to low selfhandicappers were also particularly illuminating in that these students seemed to be well aware of the distractions that posed a threat to their study. Importantly, however, whilst the low self-handicappers were well aware of the potential distractions surrounding them, they were quite unaware of the strategic ways in which these distractions could be used for self-protection purposes. In this sense, the qualitative data lend support to quantitative research findings which show that individuals who are not inclined to self-handicap are unaware of its strategic value and as a

242

consequence are more likely to accept at face value the self-handicapping strategies of others (Smith & Strube, 1991).

In addition to addressing the central research questions, other issues relevant to selfhandicapping emerged from the interviews that add to the key findings. Firstly, it was interesting to note that when invited at the end of the interview to add or emphasise anything about the way they went about their studies, a number of high selfhandicappers voluntarily returned to their self-handicapping behaviour. This seemed to imply that the strategy was salient in their academic lives. Secondly, some noted the costs of their self-handicapping, beyond simply not performing well. For example, one student commented on the panic and stress she experiences when she wastes time before exams and assignments. Thirdly, others seemed to feel little or no control over their ability to desist from self-handicapping. For example, the most striking aspect of one interviewees self-handicapping was that he felt so powerless in the face of it. According to him, it was deep set and something which he might change for a week or two but which returns: It just happens. Consistent with this, Covington (1985) notes that the temporary suspension of stressful emotions (e.g., humiliation) resulting from the use of these failure-avoiding tactics sets up an almost irresistible, selfreinforcing incentive for their repetition whenever the threat of failure arises (p. 371). Fourthly, another interviewee indicated that behaviour that might appear to be studious can also be self-handicapping. For example, she identified the lengths to which she goes in planning for upcoming exams: Id sit there and draw this absolute Picasso of a study plan and because I spent so much time planning, I wont do any study. For me that would be a brilliant time-waster. Finally, some students would be quite calculating in their use of self-handicapping. For example, one students choice of an obstacle would be based on its social acceptability. She would choose activities that were socially laudable. As she put it, saintly activities such as visiting a sick grandmother would be the strategy of choice.

8.5.2 Defensive expectations Not only do the data contribute to current understanding about the strategic nature of self-handicapping, they also underscore and confirm quantitative data concerning the strategic nature of defensive expectations (Cantor & Norem, 1989; Norem & Cantor, 243

1990b). For example, consistent with previous research indicating that defensive expectations can motivate students (Garcia & Pintrich, 1994; Norem & Cantor, 1986b), some students high in defensive expectations used their expectations partly to motivate them into action. Defensive expectations were also strategic for others in that they obviated disappointment in the event of poor performance - confirming quantitative data that it may cushion the individual in the event of failure (see Norem & Cantor, 1986b) - and also established lower and safer standards against which to be judged (see Baumgardner & Brownlee, 1987). Indeed, the fact that some students high in defensive expectations reported that they knew they were not going to fail and yet felt they were going to also suggested some cognitive posturing in support of the strategic nature of defensive expectations.

The results also indicated that the role of significant others can contribute to students defensive expectations. Not only did family members inadvertently influence the tendency to engage in defensive expectations, they quite directly contributed to these expectations. For example, one students parents explicitly taught her that by holding lower expectations for upcoming events she can minimise her disappointment if she fails. Indeed, early quantitative data demonstrated that expected negative outcomes are not so hard-hitting as unexpected ones (Feather, 1969). In fact, not only do results illuminate the family factors that can underpin defensive expectations, they also demonstrate how individuals react to others high expectations in a defensive way, confirming previous quantitative findings (Baumgardner & Brownlee, 1987).

In addition to data derived directly from the central research questions, two other interesting issues emerged in relation to defensive expectations. First, the communicated negative expectations seemed to be partly aimed at receiving positive feedback from others and that this was more to disconfirm a poor self-concept than to enhance or confirm a positive one. Second, one student noted that there were social costs of communicating negative expectations to others. According to her, everyone knows youre not going to do badly and they tend to get annoyed with you if you say you are. Similarly, another student reported that her friends were beginning to become savvy to the technique and that when she tells them she is not going to perform well they dismiss the claim saying that she is merely trying to attract attention to herself. 244

There are, then, social costs associated with the communication of unrealistically negative expectations. Indeed, Covington (1985) reported that the willingness of some students to endure such social stigma, especially in light of the counterveiling need to present a positive social image, indicates something of the extent to which individuals can be driven to avoid the implications of failure (p. 371).

8.6 Issues to be addressed at Time 2 The aims of Time 1 analyses were to explore a proposed model of the processes related to self-protection. This was achieved through quantitative analyses using confirmatory factor analysis and structural equation modelling as well as through qualitative analyses involving in-depth interviews with students high or low in selfhandicapping and defensive expectations. Time 1 data also yielded information about the domain specificity of a variety of educationally-relevant constructs and the factor structure underpinning a variety of affective and motivational factors proposed to be indicants of a self-worth motivation.

Analyses at Time 2 follow the same students through to their second year at university. The quantitative process model is tested once more and the congruence between the models at Times 1 and 2 is assessed. In addition to testing the Time 2 process model, the data at Times 1 and 2 are incorporated into a longitudinal model which tests the processes as they operate over time. In addition to these quantitative analyses, a qualitative study is also incorporated at Time 2 which utilises the longitudinal nature of the data by selecting students evincing large shifts (upwards and downwards) in self-handicapping and defensive expectations across Times 1 and 2. This component of the study seeks to explore not only concomitant shifts in the affective and motivational predictors proposed in the study, but also invites students to identify other factors which they see as pivotal in bringing about such shifts.

Whilst the Time 2 questionnaire is essentially the same as that administered at Time 1, following analysis of Time 1 data, some changes were made. First, because they played no significant role in the Time 1 process model, entity and incremental beliefs about the nature of intelligence are dropped from the Time 2 questionnaire. Second, a measure of anxiety (Pintrich & DeGroot, 1990) is incorporated in the Time 2 245

questionnaire and this is proposed to positively predict self-handicapping and defensive expectations. Third, given recent emphasis on the motive to avoid failure and its negative consequences (Skaalvik, 1997a, 1997b; Elliot & Church, 1997; Elliot & Harackiewicz, 1996; Middleton & Midgley, 1997; Midgley et al., 1998), the Time 2 questionnaire incorporates a second measure of avoidance orientation (proposed by Elliot & Church, 1997) and seeks to explore the distinctiveness of this in the context of the other affective and motivational predictors. Fourth, the study at Time 2 seeks to incorporate more objective data on academic outcomes and so the grades students earned at the end of their first year (1997) were collected and subsequently incorporated into the longitudinal analyses.

246

9. TIME 2 QUANTITATIVE ANALYSES 9.1 Method 9.1.1 Participants Students who participated at Time 1 of the investigation were again surveyed approximately one year later (1998). Respondents at Time 2 were 489 second year Teacher Education undergraduates. Most students (n=186, 38%) were enrolled at UWS Nepean, while 143 (29.3%) and 160 (32.7%) were enrolled at UWS Macarthur and University of Sydney respectively. Females (n=419, 85.9%) represented the majority of the sample (males=70, 14.1%). The mean age of participants was 21.42 years (SD=5.2). Most (n=311, 63.6%) were enrolled in a Primary education program, 111 (22.7%) were enrolled in a Secondary education program, and the others (n=67, 13.7%) were enrolled in an Early Childhood program.

9.1.2 Materials and Procedure Some adjustments were made to the instrument following the Time 1 administration. Items pertaining to entity and incremental beliefs about intelligence were dropped as they did not play a statistically significant role in the process model at Time 1. All other items included at Time 1 were retained for the Time 2 administration. Three measures were added to the instrument. First, anxiety, a construct potentially relevant to self-handicapping (see Harris et al., 1986) and defensive expectations (see Norem & Cantor, 1986b; Norem & Illingworth, 1993), was included in the instrument at Time 2. The anxiety subscale incorporated is that devised by Pintrich and DeGroot (1990 - see Appendix J). Second, students were asked to record the grades they received at the end of 1997 (the previous academic year). Because many students study broadly-based subjects that incorporate both maths and verbal domains, collecting maths-specific grades was difficult. Consequently, the performance measure used throughout the majority of analyses is a representation of academic grades generally rather than maths grades in particular. Third, to clarify the status of avoidance orientation in the context of the other Time 2 predictors, the Elliot and Church (1997) avoidance subscale was included (items appear in Appendix J). Importantly, however, given recent concerns raised regarding this subscale (Midgley et al., 1998), its construct validity was assessed before formally incorporating it into

247

central analyses. Results indicated that the Elliot and Church items loaded on the other affective and motivational predictor constructs better than on its own factor (see Appendix K for further details) and was subsequently dropped from further analyses. The method of questionnaire administration was the same as that at Time 1.

9.2 Results Results at Time 2 comprised three broad components. The first component is parallel to central Time 1 analyses involving first order and higher order CFAs of the items in the Time 2 questionnaire and structural equation modelling exploring the proposed process model. The second component of analyses is parallel to Time 1 supplementary analyses that involve tests of the distinctiveness of defensive expectations and reflectivity and the interaction of self-concept and defensive expectations. The third component of analyses involves an assessment of the Time 1 and Time 2 constructs in a proposed longitudinal model.

9.2.1 First order CFA of Time 2 subscales The pool of items incorporated in Time 2 analyses was the same as that at Time 1 (without the intelligence beliefs items) in addition to the anxiety items. Consistent with Time 1 analyses, item parcels were used and these were computed such that they were strictly parallel to the parcels used at Time 1. The factor structure of these Time 2 subscales was tested using CFA and the model fit the data acceptably (chi square=3926.35, df=1870, RNI=.91, TLI=.90, RMSEA=.048). Factor loadings are presented in Table 9.1. These factor loadings are broadly similar to those obtained at Time 1. Correlations between the five first order strategies and both predictors and outcomes are presented in Table 9.2, while correlations amongst the predictors and outcomes are presented in Table 9.3. Correlations presented here are also broadly similar to those obtained at Time 1.

248

Table 9.1 Factor loadings of Time 2 subscales

Item Parcel
Self-handicapping 1 Self-handicapping 2 Self-handicapping 3 Self-handicapping 4 Self-pres self-handicapping 1 Self-pres self-handicapping 2 Self-pres self-handicapping 3 Self-pres self-handicapping 4 Defensive expectation 1 Defensive expectation 2 Defensive expectation 3 Defensive expectation 4 Self-pres defensive expect 1 Self-pres defensive expect 2 Self-pres defensive expect 3 Self-pres defensive expect 4 Reflectivity 1 Reflectivity 2 Reflectivity 3 Private esteem-rel comp val 1 Private esteem-rel comp val 2 Private esteem-rel comp val 3 Public esteem-rel comp val 1 Public esteem-rel comp val 2 Public esteem-rel comp val 3 Ability - failure 1 Ability - failure 2 Ability - failure 3 External - failure 1 External - failure 2 External - failure 3 External - success 1 External - success 2

Factor loading
.80 .83 .86 .83 .73 .84 .90 .85 .82 .85 .88 .85 .74 .82 .93 .92 .50 .70 .68 .88 .86 .77 .84 .92 .86 .71 .67 .81 .51 .61 .80 .54 .81

Item Parcel
External - success 3 Ego-orientation 1 Ego-orientation 2 Ego-orientation 3 Public self-conscious 1 Public self-conscious 2 Public self-conscious 3 Task-orientation 1 Task-orientation 2 Task-orientation 3 Avoidance orientation 1 Avoidance orientation 2 Avoidance orientation 3 Anxiety 1 Anxiety 2 Control future fail 1 Control future fail 2 Control future fail 3 Self-concept 1 Self-concept 2 Self-concept 3 Self-concept stability 1 Self-concept stability 2 Self-concept stability 3 Self-regulation 1 Self-regulation 2 Self-regulation 3 Persistence 1 Persistence 2 Persistence 3 Future plans 1 Future plans 2 Future plans 3

Factor loading
.76 .82 .88 .87 .81 .91 .86 .66 .77 .79 .73 .76 .81 .73 .80 .85 .87 .81 .83 .85 .85 .79 .88 .85 .73 .76 .82 .73 .73 .84 .79 .89 .89

249

Table 9.2 Correlations among Time 2 strategies and both predictors and outcomes

Active SH

Self-presented SH

DE

Selfpresented DE

Reflectivity

Active SH Self-presented SH DE Self-presented DE Reflectivity

.91 .48 .63 .13

-.03 Private enhancing s-wth .23 Public enhancing s-wth .36 Ability - failure .48 External - success .56 External - failure .20 Public self-consciousness .30 Ego-orientation -.48 Task-orientation .37 Avoid-oriented perform .27 Anxiety .29 Control future failure -.11 Self-concept .33 Instability self-concept -.41 Self-regulation -.51 Persistence -.20 Future plans SH=Self-handicapping DE=Defensive expectation Coefficients > .12 are significant at p<0.05

.54 .76 .75 .20 .49 Predictors and Outcomes .06 .19 .32 .40 .39 .60 .47 .58 .56 .51 .31 .43 .35 .33 -.42 -.17 .45 .51 .35 .67 .33 .50 -.11 -.21 .36 .54 -.36 -.19 -.43 -.36 -.20 -.18

.42 .18 .43 .48 .52 .56 .44 .44 -.24 .53 .48 .41 -.10 .47 -.23 -.40 -.13

.56 .52 .42 .41 .32 .56 .41 .37 .51 .57 .34 -.01 .47 .34 -.04 -.02

250

Table 9.3 Correlations among Time 2 predictors and outcomes

1 2 3 1.Private s-wth 72 2.Public s-wth 25 34 3.Ability - failure 13 29 56 4.External - success 10 32 47 5.External - failure 71 93 37 6.Public self-consc 50 71 28 7.Ego-orientation 54 15 03 8.Task-orientation 43 61 41 9.Avoid-orient perf 29 33 61 10.Anxiety 17 29 64 11.Control failure 07 03 -38 12.Self-concept 25 29 48 13.Instab self-conc 21 03 -05 14.Self-regulation 05 -15 -33 15.Persistence 01 -08 -26 16.Future plans Decimal omitted Coefficients > .12 are significant at p<0.05

10

11

12

13

14

15

74 33 32 -20 50 41 49 -19 50 -29 -55 -31

29 36 -27 41 42 48 -07 39 -25 -36 -20

62 22 67 41 28 01 36 04 -12 -10

02 50 29 22 13 20 -07 -25 -09

01 12 -06 17 04 61 52 28

50 42 -08 41 -11 -35 -24

50 -28 48 02 -20 -14

-26 41 -18 -31 -19

-19 14 33 61

-10 -30 -15

58 22

49

251

9.2.2 Time 2 higher order CFAs Consistent with Time 1 analyses, the same first order factors were hypothesised to load on higher order factors. In terms of the models predictors: Ego-orientation, avoidance-oriented performance, private and public esteem-relevant competence valuation, and public self-consciousness loaded on the performance orientation higher order factor; external attributions for success and failure loaded on the external attributional orientation higher order factor; unstable self-concept, ability attributions for failure, and low control over future failure loaded on the uncertain personal control higher order factor. In terms of the strategies: Active and self-presented selfhandicapping and defensive expectations loaded on higher order self-handicapping and defensive expectations factors respectively. CFA was used to test the fit of this hypothesised model to the data. The model fit the data acceptably (chi square=4495.67, df=2003, RNI=.90, TLI=.88, RMSEA=.050). Higher order factor loadings are presented in Table 9.4 and correlations among these factors are shown in Table 9.5. Factor loadings are broadly consistent with those obtained at Time 1 and their strength, in conjunction with the model fit, constitute justification for pursuing this higher order structure in subsequent models. Correlations are, for the most part, broadly consistent with correlations obtained at Time 1.

9.2.3 Time 2 process model The structural relations tested in the Time 1 process model were also tested using the Time 2 data. The Time 2 model to be estimated was the same as the Time 1 model except that the anxiety subscale was incorporated into the Time 2 model as an additional predictor. Consistent with procedures at Time 1, additional parameters were freed if (a) modification indices indicated a large estimated change and (b) these parameters were conceptually defensible. Following a first analysis, one additional path was freed: That between self-concept and future plans (consistent with Time 1). Also consistent with Time 1, to enhance parsimony, structural parameters not significant at the p<0.1 level were dropped from the model. The final model fit, whilst not as good as the Time 1 model, was acceptable (chi square=4648.14, df=2028, RNI=.88, TLI=.88, RMSEA=.051). The structural relations in the model are presented in Figure 9.1.

252

Table 9.4 Time 2 higher order factor loadings

First order factors Ego-orientation Avoidance orientation Private comp valuation Public comp valuation Public self-conscious Ability failure attrib Control future failure Unstable self-concept External failure attrib External success attrib Self-concept Task-orientation Anxiety Active SH Self-pres SH Active DE Self-pres DE Reflect Self-regulation Persistence Future plans

Perform orient .70 .68 .74 .96 .96

Uncertain control

External

Self-concept

Task-orient

Anxiety

SH

DE

Reflect

Self-reg

Persist

Future plans

.82 .69 .65 .83 .90 1.00 1.00 1.00 .92 .98 .84 .90 1.00 1.00 1.00 1.00

253

Table 9.5 Time 2 correlations among higher order and first order factors

1 2 3 1.Performance orient .48 2.Uncertain control .39 .78 3.External attributions .03 -.40 -.17 4.Self-concept .21 .01 -.26 5.Task-orientation .41 .73 .48 6.Anxiety .31 .49 .60 7.SH .52 .77 .71 8.DE .60 .57 .44 9.Reflect .04 -.14 -.31 10.Self-regulation -.16 -.43 -.55 11.Persistence -.10 -.10 -.31 12.Future plans SH= Self-Handicapping DE=Defensive Expectations Coefficients > .12 are significant at p<0.05

10

11

.17 -.28 -.12 -.16 -.01 .14 .33 .61

.11 -.45 -.24 .36 .61 .52 .28

.34 .63 .59 .02 -.20 -.14

.76 .19 -.39 -.46 -.21

.52 -.24 -.44 -.17

.33 -.05 -.02

.58 .22

.49

254

Public selfconsciousness

.96 .70 .68 Performance orientation .19 .27 .27 -.41 .29 -.28 .40 .86 .85 External orientation Reflectivity .67 -.28 -.16 Selfhandicapping Self-regulation Active self-handicapping .93 .97 Self-pres self-handicapping

Egoorientation Avoidance orientation Private comp valuation Public comp valuation External External causes causes of of success success External causes of failure Ability causes of failure Control future failure Instability self-concept

.74

.96

Task-orientation

.29
-.68 .32 -.28 .25 .11* .28 .88 Self-pres defensive expectations .09* .54 Defensive expectations 84 Active defensive expectations Future plans Persistence

.83 .69 .65

Uncertain personal control

.20

Anxiety

Selfconcept

Figure 9.1 Significant structural relations (and higher order factor loadings) in Time 2 process model Note. * p<0.10. All other paths significant at p<0.05 255

For the most part, the structural relations are consistent with those obtained at Time 1. Performance orientation positively predicts self-handicapping, defensive expectations, and reflectivity. Mastery orientation negatively predicts self-handicapping and defensive expectations and positively predicts reflectivity. External attributional orientation positively predicts self-handicapping and to a lesser extent defensive expectations. Uncertain personal control and anxiety positively predict defensive expectations and reflectivity. To a lesser extent, anxiety positively predicts selfhandicapping. Self-concept strongly predicts future academic plans and to a limited extent predicts reflectivity.

Unfortunately, structural relations between strategies and outcomes were problematic in parts, and suggestive of some statistical suppression. For example, in marked contrast to correlations presented earlier (see Table 9.5), self-handicapping does not significantly predict self-regulation, while defensive expectations and reflectivity are extremely strong predictors of self-regulation. For these reasons, it was considered important to assess the three strategies individually in three separate submodels. Process Submodel 1 examined the Time 2 process model excluding self-handicapping and reflectivity. Process Submodel 2 examined the model excluding defensive expectations and reflectivity. Process Submodel 3 examined the model excluding selfhandicapping and defensive expectations. Each of the three submodels fit the data acceptably. Process Submodel 1 (chi square=3329.81, df=1394, RNI=.89, TLI=.89, RMSEA=.053) is presented in Figure 9.2. Process Submodel 2 (chi square=2577.79, df=962, RNI=.89, TLI=.88, RMSEA=.059) is presented in Figure 9.3 (note that uncertain personal control and its first order indicants are dropped from this model because it yielded no significant effects). Process Submodel 3 (chi square=2108.99, df=754, RNI=.89, TLI=.88, RMSEA=.061) is presented in Figure 9.4 (note that external attributional orientation and persistence are dropped from this model because they yielded no significant effects).

With particular focus on the relationships between strategies and outcomes, results in Process Submodel 1 (Figure 9.2) indicate that defensive expectations strongly

256

Public selfconsciousness Egoorientation Avoidance orientation Private comp valuation Public comp valuation External External causes causes of of success success External causes of failure Ability causes of failure Control future failure Instability self-concept

.96 .70 .68 .74 .24 .96 Task-orientation Self-regulation -.29 Performance orientation

.86 .86

External orientation

.27 -.28 -.52 Defensive expectations 86 .28 .85 Self-pres defensive expectations .53 Active defensive expectations Future plans -.13

Persistence

.84 .70 .64

Uncertain personal control

.25

Anxiety

Selfconcept

Figure 9.2 Significant structural relations (and higher order factor loadings) in Time 2 Process Submodel 1 - Defensive Expectations 257

Public selfconsciousness Egoorientation Avoidance orientation Private comp valuation Public comp valuation External External causes causes of of success success External causes of failure

.96 .70 .68 .74 -.43 Selfhandicapping -.53 -.17 .41 .86 .86 Persistence External orientation -.43 Self-regulation Performance orientation .18 Active self-handicapping .95 .94 Self-pres self-handicapping

.96

Task-orientation

.10* Future plans

Anxiety

.53 Selfconcept

Figure 9.3 Significant structural relations (and higher order factor loadings) in Time 2 Process Submodel 2 - Self-Handicapping Note. * p<0.10. All other paths significant at p<0.05 258

Public selfconsciousness Egoorientation Avoidance orientation Private comp valuation Public comp valuation

.96 .67 .68 .74 Performance orientation

.26

.96

Task-orientation .31 Self-regulation

.36 Reflectivity

.32 Ability causes of failure Control future failure Instability self-concept .83 .70 .63 Anxiety Future plans Uncertain personal control .18*

.12*

.54

Selfconcept

Figure 9.4 Significant structural relations (and higher order factor loadings) in Time 2 Process Submodel 3 - Reflectivity Note. * p<0.10. All other paths significant at p<0.05 259

negatively predict persistence and to a lesser extent negatively predict self-regulation and future plans. Results in Process Submodel 2 (Figure 9.3) indicate that selfhandicapping strongly negatively predicts both persistence and self-regulation and to a lesser extent negatively predicts future academic plans. Finally, reflectivity, in Process Submodel 3 (Figure 9.4), positively predicts self-regulation.

9.2.4 Testing interactions in the Time 2 process model Consistent with Time 1, the latent interactions between defensive expectations and reflectivity and between level and stability of self-esteem were examined in the Time 2 process model (see Appendix F1 for technical details of these procedures). The defensive expectations x reflectivity interaction was located alongside the three strategies (self-handicapping, defensive expectations, and reflectivity) and the level x stability of self-concept interaction was positioned as one of the affective and motivational predictors. The level x stability of self-concept interaction term did not significantly predict any of the strategies in the model, nor was the defensive expectation x reflectivity interaction significantly predicted by any of the affective and motivational predictors. The defensive expectations x reflectivity interaction did, however, significantly predict self-regulation.

This interaction was followed up using the approach recommended by Aiken and West (1991, see Appendix F2 for technical details). These analyses indicated that the interaction effect lies largely in the distinctiveness of the Low defensive expectations/High reflectivity group from the High defensive expectations/Low reflectivity group and these two from the other two groups (see Figure 1 in Appendix F2). Specifically, students low in defensive expectations and high in reflectivity are the highest self-regulators while students high in defensive expectations and low in reflectivity are the lowest self-regulators. These data are broadly consistent with present contentions that defensive expectations are not particularly adaptive in terms of self-regulation and that it is reflectivity which is the most adaptive of the two.

260

9.2.5 Assessing a model in which defensive expectations predict reflectivity As foreshadowed in the review of literature and tested at Time 1, an alternative conceptualisation of defensive pessimism is one in which defensive expectations relax students sufficiently to enable them to engage in reflectivity which in turn positively predicts academic outcomes (Norem, personal communication, March 26, 1997, March 7, 1998). Essentially this can be tested by freeing a beta path between defensive expectations and reflectivity rather than simply correlating their residuals as in the central model presented above. To test this issue, the trimmed model (see Figure 9.1) was estimated, however, rather than correlating defensive expectations and reflectivity, the defensive expectations construct was freed to predict reflectivity (as outlined at Time 1, however, the two models are statistically equivalent, but the second approach is carried out for the purpose of estimating a beta path explicitly). Results generated by this model (chi square=4653.22, df=2029) indicate, consistent with suggestions by Norem that defensive expectations significantly predict reflectivity (=.35). An important qualification of this finding, however, is that the indirect positive path between defensive expectations and self-regulation via reflectivity (.35 x .67: =.24) is a good deal weaker than the direct negative path between defensive expectations and self-regulation (=-.68). Thus while defensive expectations via reflectivity yield positive effects, this is overshadowed by the strong direct negative path between defensive expectations and self-regulation. However, given that some statistical suppression was evident in the central model (Figure 9.1), caution is recommended when interpreting the Time 2 predictive path between defensive expectations and reflectivity. In view of this, emphasis is given to the Time 1 results in which no statistical suppression was evident and in which defensive expectations did not significantly predict reflectivity.

9.2.6 The longitudinal data (Time 1 - Time 2)

9.2.6.1 Participants Data for both Time 1 and Time 2 were available for a total of 328 respondents. Most respondents (n=291, 88.7%) were female, 37 (11.3%) were male. The mean age was 21.31 (SD=4.84). Just under forty percent (n=129, 39.3%) were from UWS Nepean, 106 (32.3%) were from UWS Macarthur, and the remainder (n=93, 28.4%) from the 261

University of Sydney. Primary students (n=240, 73.2%) represented the majority of this sample, 57 (17.4%) were enrolled in the Secondary program, and 31 (9.5%) in an Early Childhood program.

9.2.6.2 Enhancing parsimony for Time 1 - Time 2 analyses To estimate the longitudinal model at item level requires many more parameters to be freed than is ideal given the sample size. For this reason, constructs which in previous process models were incorporated into higher order factors, were subsequently estimated in the longitudinal model as first-order factors represented by composite scores. Thus, for example, Performance Orientation - previously a higher order factor in Time 1 and Time 2 process models comprised of five first-order factors and fifteen item parcels - was now incorporated into the longitudinal model as a first-order factor comprised of only five indicators. To do this, composite scores for each of egoorientation, public self-consciousness, avoidance-oriented performance, and (private and public) esteem-relevant competence valuation were computed in SPSS and these were used as indicators of Performance Orientation - now a first-order factor comprised of the five composite scores. In cases where constructs were not represented as higher order factors (e.g., self-concept and task-orientation) in previous models, item parcels were retained as indicators of these constructs and these constructs were retained as first order factors. Importantly, however, when composite scores were computed for these analyses, items for each score were not given equal weighting. Rather, for each construct, one factor congeneric CFAs were computed and the regression weights for each of the indicators of that construct were generated. These weights were then used in the computation of the composite scores. This method of generating composite scores for subsequent structural equation modelling is explicated in greater detail elsewhere (Joreskog, 1971; Rowe, Hill, & HolmesSmith, 1995).

9.2.6.3 CFA of this parsimonious set of constructs Before examining this parsimonious set of constructs in the hypothesised longitudinal model, correlations between these constructs were first generated in a CFA (which also correlated parallel Time 1/Time 2 uniquenesses). This model yielded an

262

acceptable fit to the data (chi square=3038.84, df=1736, RNI=.91, TLI=.89, RMSEA=.048) and correlations are presented in Table 9.6.

9.2.6.4 The longitudinal model The central relationships in the hypothesised longitudinal model are presented in Figure 9.5. In addition to parameters shown in Figure 9.5, three other sets of parameters were freed to be estimated. These were (a) test-retest paths between parallel Time 1 and Time 2 constructs (e.g., between Time 1 task-orientation and Time 2 task-orientation), (b) correlations between parallel Time 1 and Time 2 uniquenesses (e.g., between the uniqueness of each Time 1 task-orientation item parcel and its Time 2 counterpart), and (c) correlations among constructs at the same point in the model (e.g., among Time 1 predictors or among residuals of Time 1 strategies). Essentially, this model is one that adheres to a process-orientation in which the effects of Time 1 constructs on Time 2 constructs operate via the processes outlined in the models presented earlier. For the most part, then, the longitudinal model is one of mediation in which, for example, the relationships between the outcomes at Time 1 and strategies at Time 2 are mediated by predictors at Time 2.

Time 1

Time 2

Predictors

Strategies

Outcomes

Predictors

Strategies

Outcomes

Grades

Figure 9.5 Basic structure of the hypothesised longitudinal model

263

Table 9.6 Correlations generated in the CFA of the parsimonious Time 1 - Time 2 data set
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1. T1PO 42 2. T1UC 37 68 3. T1EA 07 -38 -06 4. T1SC 44 15 02 23 5. T1TO 21 39 57 -14 -32 6. T1SH 48 73 70 -15 02 57 7. T1DE 61 56 41 05 52 07 40 8. T1REF 11 -08 -33 18 46 -49 -18 42 9. T1SR -06 -27 -47 34 44 -58 -34 10 67 10. T1P 01 -15 -13 63 20 -14 -10 06 25 33 11. T1FP 30 30 07 14 28 38 38 -10 -18 -10 64 12. T2PO 28 48 -30 -06 40 47 35 -20 -36 -26 50 68 13. T2UC 17 38 -08 -13 46 42 19 -24 -43 -18 40 73 66 14. T2EA 10 -33 -15 16 -11 -10 02 13 31 48 09 -38 -10 74 15. T2SC 26 16 -08 15 -24 03 27 32 36 19 19 -02 -24 20 59 16. T2TO 11 25 37 -05 -29 37 05 -34 -28 -11 29 44 56 -07 -44 67 17. T2SH 31 50 51 -11 -15 49 29 -22 -32 -12 57 71 70 -15 -23 68 67 18. T2DE 53 47 54 -01 16 28 50 03 -21 -10 72 58 51 02 26 26 69 63 19. T2REF 17 03 -24 14 34 -37 -05 32 47 22 05 -14 -27 18 55 -37 -19 15 72 20. T2SR 01 -09 -32 22 33 -40 -16 11 52 32 -15 -45 -52 37 50 -45 -41 -17 55 75 21. T2P 11 -08 -14 39 20 -10 04 04 15 30 -05 -27 -28 63 33 -18 -15 -06 22 50 57 22. T2FP 08 -16 -29 -01 15 -23 -06 04 16 12 -14 16 -21 -19 03 16 -20 -07 10 18 13 -07 23. 97GR Note. Decimal omitted; Test-retest correlations in bold T1=Time 1; T2=Time 2; PO=Performance Orientation; UC=Uncertain Control; EA=External Attributions; SC=Self-Concept; TO=Task-Orientation; SH=Self-Handicapping; DE=Defensive Expectations; REF=Reflectivity; SR=Self-Regulation; P=Persistence; FP=Future Plans; 97GR=Grades at end of 1997 (first year) Coefficients > .13 are significant at p<0.05

264

The hypothesised longitudinal model was tested and modification indices inspected for large estimated changes. Three parameters were freed following this procedure: In both Time 1 and Time 2, private esteem-relevant competence valuation was free to also load on task-orientation, while self-concept at Time 1 was freed to predict persistence at Time 1. Having freed these parameters, the model was re-estimated and yielded an acceptable fit to the data (chi square=3294.50, df=1905, RNI=.90, TLI=.89, RMSEA=.047). The structural parameters are presented in Figure 9.6.

Relationships between Time 1 predictors and Time 1 strategies are largely consistent with relationships obtained in the Time 1 process model (see Figure 6.2). Relationships between Time 2 predictors and Time 2 strategies are consistent with those obtained in the Time 2 process model (see Figure 9.1) but are generally weaker. The reason for these weaker parameters is largely because the bulk of the variance of Time 2 strategies has been explained by their Time 1 counterparts. Interestingly, these test-retest path coefficients are not as large as those for other constructs indicating that the affective and motivational predictors are effective in explaining variance in the strategies over and above that explained by the Time 1 strategies. The relatively lower test-retest paths for the strategies are even more striking when considered in the context of the test-retest correlations presented in Table 9.6. One of the more pivotal points of the longitudinal model is the link-points between the two process models. In relation to this, academic grades positively predict Time 2 performance orientation and negatively predict uncertain personal control. Also, persistence negatively predicts Time 2 uncertain personal control and external attributional orientation and positively predicts Time 2 task-orientation. Particular attention is directed to the significant negative association between both self-handicapping and future plans at Time 1 and later academic grades.

265

[.65] Perform Orn Perform Orn .11 .28 .26 .22 -.26 -.45 Self-Reg -.52 -.23 .53 External .37 Def Expect -.15* Persist -.17 .11* .44 -.17 [.55] Taskorientation .32 Reflect Fut Plans Taskorientation -.29 -.23 [.32] .16 Reflect .42 .12* .26 .62 [.74] Self-concept Self-concept [.36] Fut Plans [.54] External .26 .16* [.30] Def Expect .20 [.65] Persist Grades -.20 .15 [.59] Uncert Control .43 Self-Hand .41 -.11* Uncert Control .29 .32 -.22 .43 .28 [.44] Self-Hand -.21 Self-Reg [.62]

-.44 -.15

Figure 9.6 Significant structural relations in longitudinal model Note. Time 1 - Time 2 test-retest paths are indicated in brackets [ ] * p<0.10. All other paths significant at p<0.05

266

As with the Time 2 process model, relationships between the strategies and outcomes are problematic. The primary issue that requires further consideration is the predominantly non-significant role of defensive expectations - particularly in the context of the prominent role it played in the Time 1 and Time 2 process models (and Process Submodel 1) and the correlations with Time 1 and Time 2 outcomes. Its nonsignificant role in the longitudinal model seems largely an artefact of collinearity with self-handicapping and reflectivity (note in Table 9.6 that the relationships among the three strategies are generally much higher at Time 2 than at Time 1 - particularly that between reflectivity and defensive expectations) and this gives rise to the suppression of defensive expectations effects. It was therefore considered important to examine three submodels (consistent with submodels explored in relation to the Time 2 process) designed to assess the relative independent effects of self-handicapping, defensive expectations, and reflectivity. In Longitudinal Submodel 1, selfhandicapping and reflectivity at Time 1 and Time 2 are excluded from analyses. In Longitudinal Submodel 2, defensive expectations and reflectivity at Time 1 and Time 2 are dropped. In Longitudinal Submodel 3, defensive expectations and selfhandicapping are dropped. These submodels are presented in Figures 9.7, 9.8, and 9.9 respectively. Longitudinal Submodel 1 fit the data well (chi square=2470.49, df=1349, RNI=.91, TLI=.90, RMSEA=.050), as did Longitudinal Submodel 2 (chi square=2364.42, df=1348, RNI=.91, TLI=.91, RMSEA=.048), and Longitudinal Submodel 3 (chi square=2595.58, df=1457, RNI=.90, TLI=.89, RMSEA=.049). When defensive expectations, self-handicapping, and reflectivity are dealth with in this way, their respective roles - independent of their collinearity - are clarified.

267

[.62] Perform Orn Perform Orn

.27

-.19 .12 [.61] -.09* .28 [.72] .29 Self-Reg

Uncert Control

Self-Reg .38

Grades

Uncert Control

-.20 -.21 External .40 Def Expect -.31 Persist .14 -.18 -.13 Taskorientation Fut Plans [.56] Taskorientation .42 .23 .61 [.74] Self-concept Self-concept -.23 [.35] Fut Plans -.20 [.55] External .23 [.27] Def Expect -.19 Persist [.67]

Figure 9.7 Significant structural relations in Longitudinal Submodel 1 - Defensive Expectations Note. Time 1 - Time 2 test-retest paths are indicated in brackets [ ] * p<0.10. All other paths significant at p<0.05 268

[.63] Perform Orn Perform Orn .12 .21 -.27 -.19 .14 [.62] Uncert Control Self-Hand -.50 Self-Reg -.52 -.21 .51 External Persist -.44 -.17 Taskorientation Fut Plans .12 -.29 [.55] Taskorientation .41 .24 .62 [.74] Self-concept Self-concept [.34] Fut Plans -.16 [.57] External -.12 .27 [.65] Persist Grades -.09* Uncert Control [.43] Self-Hand -.14 Self-Reg -.24 [.66]

Figure 9.8 Significant structural relations in Longitudinal Submodel 2 - Self-Handicapping Note. Time 1 - Time 2 test-retest paths are indicated in brackets [ ] * p<0.10. All other paths significant at p<0.05 269

[.65] Perform Orn .25 -.21 .14 [.64] Uncert Control Self-Reg .39 -.24 [.58] Persist .12* .39 -.13 Taskorientation .36 Reflect Fut Plans [.55] Taskorientation .21 [.37] Reflect .42 .12* .27 .61 [.74] Self-concept Self-concept [.35] Fut Plans .16 -.24 External .20 [.75] Persist Grades -.09* Uncert Control .32 [.69] Self-Reg .19 Perform Orn

Figure 9.9 Significant structural relations in Longitudinal Submodel 3 - Reflectivity Note. Time 1 - Time 2 test-retest paths are indicated in brackets [ ] * p<0.10. All other paths significant at p<0.05

270

In Longitudinal Submodel 1, Time 1 defensive expectations negatively predict selfregulation and persistence at Time 1 and to a lesser extent Time 2 defensive expectations negatively predict persistence at Time 2. In Longitudinal Submodel 2, self-handicapping quite strongly negatively predicts self-regulation and persistence at Time 1 and negatively predicts later academic grades. To a lesser extent, selfhandicapping negatively predicts self-regulation and persistence and weakly negatively predicts future academic plans at Time 2. In Longitudinal Submodel 3, Time 1 reflectivity positively predicts self-regulation and to a lesser extent positively predicts persistence. At Time 2, reflectivity positively predicts self-regulation.

9.3 Summary of chapter findings Time 2 process model findings were broadly consistent with those obtained at Time 1: Performance orientation positively predicted self-handicapping, defensive

expectations, and reflectivity; task-orientation negatively predicted self-handicapping and defensive expectations and positively predicted reflectivity; external attributional orientation positively predicted self-handicapping, and to a lesser extent, defensive expectations; uncertain personal control positively predicted defensive expectations and reflectivity; self-concept positively predicted future academic plans, and to a lesser extent, reflectivity. The Time 2 process model also incorporated anxiety which was found to positively predict self-handicapping, defensive expectations, and reflectivity. To clarify the relationships between strategies and outcomes, three submodels were explored. Findings showed that self-handicapping and defensive expectations negatively predicted self-regulation, persistence, and future academic plans, while reflectivity positively predicted self-regulation.

The longitudinal model (n=328) provided a test of the proposed way in which the two processes operate consecutively (and cumulatively) with particular focus on the pivotal connections between the two processes over time in addition to the effects of Time 1 strategies and outcomes on academic grades obtained at the end of 1997 (first year). The findings in the Time 1 and 2 process components of the longitudinal model were largely congruent with results in the Time 1 and 2 process models examined using the entire Time 1 (n=584) and Time 2 (n=489) samples. Pivotal relationships in 271

the longitudinal model which are not specified in the separate Time 1 and Time 2 process models include: A negative path between both self-handicapping and future plans at Time 1 and later academic grades; a positive path between grades and Time 2 performance orientation; a negative path between grades and Time 2 uncertain personal control; a negative path between Time 1 self-regulation and Time 2 performance orientation; a negative path between Time 1 persistence and both uncertain personal control and external attributional orientation at Time 2; and, a positive path between Time 1 persistence and Time 2 task-orientation. Importantly, these models yielded a good fit to the data and support the proposition of the longitudinal model as one comprised of processes in which the effects of one set of constructs are mediated by the constructs immediately following them in the process. This longitudinal model, then, may be seen as a mediational process model which makes explicit predictions about how the effects of predictors, strategies, and outcomes operate over time.

272

10. EXPLORING SHIFTS ACROSS TIME 1 AND TIME 2 FROM A QUALITATIVE PERSPECTIVE

The purpose of the Time 2 qualitative study is to explore shifts in self-handicapping and defensive expectations from first to second year, concomitant shifts in key process-model predictors, and the salience of these constructs in students academic lives in second year. The second key aim is to identify reasons, from respondents perspectives, as to why upward and downward shifts on these constructs occurred across first and second year. The third aim is to explore respondents comparative perceptions of first and second year and how these perceptions relate to the identified shifts in self-handicapping and defensive expectations. Finally, supplementary interview questions sought to extend the Time 1 qualitative analysis of the role of cultural factors in influencing students approaches to study (see Table 10.1 for a summary of key research issues).

It was discussed in Chapter 7 how qualitative data complement quantitative data both in illuminating and providing insights into the personal perspectives of a target group (Patton, 1987; Strauss & Corbin, 1990). Qualitative data can also introduce and provide insight into issues that have not been addressed by the quantitative data and, perhaps, move the substantive domain in potentially fruitful new directions. Accordingly, the qualitative data presented at Time 1 were used not so much to test the quantitative findings as to flesh out the quantitative data from students personal perspectives. The Time 1 qualitative analyses focused on students high and low in self-handicapping and defensive expectations and explored each component of the Time 1 process model from the personal perspectives of these students.

One issue unexplored as yet concerns the personal accounts of students who yielded relatively large shifts (upwards and downwards) in self-handicapping and defensive expectations across Time 1 and Time 2. Whilst the longitudinal data addressed these shifts through estimation of test-retest parameters, students perceptions as to what accounted for such shifts could yield important data that contribute to an understanding of the process of defensive manoeuvring as gleaned from the

273

Table 10.1 Qualitative research issues addressed at Time 2 Broad themes

Salience of students academic strategies in second year and perceived shifts in these strategies across Times 1 and 2 Shifts in defensive expectations

Specific aspects of the qualitative study

Shifts in self-handicapping

Research issues to be addressed

Students views on self-handicapping in their academic life in second year; Students views on their self-handicapping in their second year relative to first year; Students perceptions as to why shifts have occurred. Students views on defensive expectations in their academic life in second year; Students views on their defensive expectations in second year relative to first year; Students perceptions as to why shifts have occurred.

Salience of higher order affective and motivational factors in second year and shifts in these across Times 1 and 2

Performance orientation

Students views of the importance and satisfaction derived from performance and their ability to demonstrate competence; Students views of their performance orientation in second year relative to first year.

Task-orientation

Students views of the importance and satisfaction derived from learning and gaining mastery; The salience of mastery in second year relative to first year.

External attributional orientation

The salience of respondents external attributions; Students external attributional orientation relative to first year.

Uncertain personal control

Students views regarding the certainty of their personal control over achieving success and avoiding failure; The extent of this (un)certainty relative to first year.

Anxiety

The salience of anxiety in students academic lives; The level of this anxiety relative to first year.

Other salient factors

Significant events identified by students Culture

Students perceptions of the major differences between first and second year; Significant events accounting for identified shifts between first and second year. The role of cultural factors in influencing the way students go about their studies.

274

quantitative models. It will also be recalled that the Time 1 qualitative component comprised interviews with students on the basis of first order factors (e.g., egoorientation, self-concept stability, public self-consciousness) rather than the higher order factors (e.g., uncertain personal control, performance orientation). In the Time 2 qualitative phase, data focused on the more overarching higher order factors and explored students perceptions of these factors, as they relate to the identified shifts across Time 1 and Time 2.

10.1 Qualifications for the Time 2 qualitative study Much the same qualifications that were made at Time 1 also hold here and the reader is urged to revisit the discussion of these qualifications (which involve the representativeness of the interviewees, the purpose of illuminating rather than testing the quantitative data, the issue of cause and effect, and the congruence of reports by self-handicappers and defensively expectant students - see Section 7.4). In addition to these qualifications, Time 2 presented issues particular to the longitudinal processes addressed. Asking respondents to compare their academic approach in second year to the one they used in first year proved difficult at times. At various points in the interviews, some respondents experienced difficulty remembering exactly how they went about their studies in first year. Indeed, Nisbett and Wilson (1977) have questioned the extent to which individuals are aware of internal processes and states and their position would cast doubt on the validity of data based on comparing accounts of current internal processes with those recalled from a year earlier. Notwithstanding this, where students were capable of doing so and where their comments were relevant to the quantitative findings, the qualitative study reports such data.

275

10.2 Method 10.2.1 Sample Time 1 and Time 2 self-handicapping and defensive expectation scores were computed for each student in the quantitative sample by generating the mean of the set of items comprising the subscale. A self-handicapping and a defensive expectations difference score for each student was then computed by subtracting the Time 1 score from the Time 2 score. To select students evincing large shifts in self-handicapping, students were ranked on the basis of their self-handicapping difference score. Those at the top of the list (who had also recorded their name and telephone number on the Time 1 questionnaire) were selected for the participant pool of those representing the largest increases in self-handicapping. Those at the bottom of the list (who had also recorded their name and telephone number on the Time 1 questionnaire) were selected for the pool of respondents evincing the greatest decrease in self-handicapping. The same procedure was carried out to generate an eligibility list of students exhibiting the greatest shifts in defensive expectations. From these lists, 16 respondents were selected for and agreed to an interview. Details of this sample appear in Table 10.2. Important to note is that the potential pool of respondents was restricted to those students who completed both Time 1 and Time 2 questionnaires and who had recorded their telephone number on the Time 1 questionnaire. Because of this, it was not possible to access students who evinced the greatest shifts in self-handicapping and defensive expectations, only those towards the top and bottom of the lists.

10.2.2 Procedure The semi-structured interview schedule is presented in Appendix L. Otherwise, the Procedure is the same as Time 1, detailed in Section 7.3.2. Importantly, consistent with Time 1, at no stage were the terms self-handicapping and defensive expectations used in interviews. Rather, descriptions of such behaviour were used.

276

Table 10.2 Sample details for Time 2 qualitative study

Name*
Ellie (DNDE) Stephanie (DNDE) Joan (DNDE) Martina (DNDE) Samantha (UPDE) Isabelle (UPDE) Theresa (UPDE) Craig (UPDE) Helen (DNSH) Glen (DNSH) Patrick (DNSH) Cathy (DNSH) Jodie (UPSH) Katrina (UPSH) Dominique (UPSH) Danny (UPSH)

Designation**
DNDE DNDE DNDE DNDE UPDE UPDE UPDE UPDE DNSH DNSH DNSH DNSH UPSH UPSH UPSH UPSH

University*
1 1 1 2 1 1 1 2 2 3 3 3 2 1 1 3

Difference***
Down 3.38 Down 3.00 Down 4.25 Down 2.38 Up 1.75 Up 1.25 Up 2.25 Up 1.50 Down 1.13 Down 2.25 Down 1.38 Down 1.00 Up 0.88 Up 2.12 Up 2.00 Up 2.63

* Pseudonyms used and University not specified to ensure confidentiality ** DNDE=Decrease in defensive expectations; UPDE=Increase in defensive expectations DNSH=Decrease in self-handicapping; UPSH=Increase in self-handicapping *** Self-handicapping difference range=-2.88 to 2.88; SD=0.84; Defensive expectations difference range=-4.75 to 2.88; SD=1.19

277

10.2.3 Data analysis The general procedures used in category formation (see Appendix M) and analytic software are the same as at Time 1, detailed in Section 7.3.3. 10.3 Results 10.3.1 Shifts in self-handicapping and defensive expectations11

10.3.1.1 The salience of and shifts in self-handicapping Students evincing upward shifts in self-handicapping tended to report that selfhandicapping behaviour was not only salient in their second year of study but also had increased from first year. As it turned out, most of the self-handicapping behaviour students related involved procrastination. Dominique (UPSH) reported that when tackling assignments and study for exams, I do leave things not to the very last minute, but close to it. According to her, I think, OK Ive got time, Ill do it tomorrow and something else comes along and I think, Ill do it tomorrow. When asked how much control she felt she had over her procrastination, Dominique (UPSH) reported that it just happens. Indeed, while she recognised that, if I start earlier and be more prepared, then I would have better marks, she reported that, Ive always been used to doing it my way. Similarly, despite acknowledging that she is capable of doing the assignments, Jodie (UPSH) reported that once I sit down to do the assignment, I think, Ill go and watch a bit of TV. Jodie (UPSH) reported that compared to first year she was increasingly leaving study to the last minute. This, she reports, is because her workload is heavier. Importantly, however, now in second semester, Jodie (UPSH) has recognised that her self-handicapping is not actually amenable to academic success and is trying a different approach which involves pacing things out rather than leaving them to the last moment. Other students have also recognised that their procrastination is a problem for them and have taken steps to address it. According to Katrina (UPSH), Ive actually written up the subjects and the

11

In this section, just the nature of shifts in self-handicapping and defensive expectations are described. Possible reasons for and factors underpinning these shifts that are relevant to the hypothesised process model (e.g., task-orientation, anxiety etc.) are explored in Section 10.3.2. In Section 10.3.3, students own opinions about the factors underpinning shifts between first and second year are examined.

278

weeks. Like in one part it tells me to do one part of this assignment in this week and then I tick it off. But unfortunately, Im behind in that [strategy] Ive noticed.

Students demonstrating downward shifts in self-handicapping, on the other hand, tended to report a more adaptive work ethic in second year. Whereas reports by UPSHs reflected avoidance, reports by DNSH students implied a success-orientation. For example, Helen (DNSH) reported that, I dont mess up for the week before the assignment and suddenly have to cram it in. Cathy (DNSH) who has heavy work commitments reported that while she can sometimes put things off, she generally plans well in advance and feels she leaves adequate time to do the assignments. Consistent with their questionnaire data, students selected as exhibiting downward shifts in self-handicapping reported being less inclined in second than first year to engage in behaviour such as procrastination and strategic work avoidance. Again, their reports indicated that they were adopting a more positive work focus. Three students Glen, Helen, and Patrick - were illustrative of this clearer sense of purpose and commitment. Glen (DNSH) reported that he left his work to the last minute in first year and that because of this his results suffered. In second year, however, he decided that he would be well prepared and organised for his assessment tasks. Helen (DNSH) reported that it took some time in first year to get into the swing of study and that this led to poor study practices. She also cited negative peer influence as a reason for why she self-handicapped: Just the talk amongst ourselves like, Oh I wish we could give up, I dont have time for this - some fairly negative thinking can drag you down. Patrick (DNSH) withdrew effort - particularly study for exams - in first year, because I couldnt be bothered and I didnt know if I was going to stay. I was sort of sick of it [study] and I didnt like it too much. Now in second year, Patrick (DNSH) wants to show to myself and people I know here that I can do it and I will be good at it.

10.3.1.2 The salience of and shifts in defensive expectations Students selected as demonstrating upward shifts in defensive expectations not only recognised the centrality of defensive expectations in their second year, but also generally acknowledged that their defensive expectations were more salient in second than first year. Despite passing all his subjects in first year, Craig (UPDE) held 279

negative expectations about how he would perform in second year. According to Craig (UPDE), at the beginning of second year,

I could tell as soon as I started the course that they [his subjects] were going to be way harder and the school I was teaching at was way worse than the school I had last year. As a result, that pulled me down. Then I had family troubles and a lot of work came in.

Other students reports were illustrative of the tendency to hold negative expectations that seemed to run contrary to their previous performance. Samantha (UPDE) performed at Distinction level in first year and when she received much the same in second year she was surprised and flabbergasted. According to Samantha (UPDE), the advantage of thinking in this way is that when I get good results, I think, Oh, hello! . . . Ive never been one to set my goals too high. I keep them sub-low. So that I will be surprised. In response to probing about her defensive expectations, she also recognised that her negative expectations motivated her in a twisted kind of way. Similar to Samantha (UPDE), Theresa (UPDE) was explicit in stating that I usually dont set myself up too high - just to be surprised and just not to be disappointed. I find that if expect to do well, I end up failing . . . Its a self-confidence thing, and also so people dont say, Oh look at you, you said you were going to do well and look what happened. So I tend to be a little bit cautious.

The students selected as demonstrating downward shifts in defensive expectations between first and second year were clearly more optimistic in second year. Ellie (DNDE) reported that she was hopeful that she will do well and was generally pretty confident about how she would perform. In second year, Ellie (DNDE) has been applying herself and is associating and working with a group of students who work hard. As a result she is more optimistic about how she will perform in upcoming assessment. Other students provided accounts also reflecting optimism. Joan (DNDE) was aiming for Distinctions in second year, which is quite an improvement on her first year grades. Joans experience of first year is that if I think negative, then everything will turn out negative for me. Whereas . . . [in second year] Ive thought about things positively, I havent got stressed about things, and things have worked out for the 280

best. Optimism also seemed to mark Martinas (DNDE) second year more than her first year. She reported that because she is attending to all the readings, lectures, and tutorials, she is really hoping she will do better than first year and is feeling positive about how she will perform. Compared to first year, Martina (DNDE) is more optimistic because she is investing more effort in her studies. This optimism was also reflected in Stephanies (DNDE) account who, similar to Martina (DNDE), believes this to be the case because she is better prepared in second year.

10.3.2 Concomitant shifts in affective and motivational constructs identified in process models

The interviews also focused on the salience of and shifts in the affective and motivational constructs identified in the process models. These data were very much parallel to the interview data presented at Time 1 and so they are only briefly addressed here with focus on selected students as illustrative cases. The interview data in relation to anxiety, however, is new to the study and so this is dealt with in greater detail than data pertaining to the other affective and motivational constructs.

10.3.2.1 The salience of and shifts in external attributions Students evincing upward shifts in self-handicapping were inclined to see academic outcomes as due to external factors. Moreover, these students were inclined to see these external factors as playing a more salient role in second year than first year. These external factors, however, did vary from respondent to respondent. Some UPSH students saw their results as due to variations in marking standards whereas others saw teaching quality as pivotal. For example, according to Joan (UPSH), I think it does depend on how they mark. Some markers you find are harder and some are easier. Like, you can get marks for pretty presentation, whereas Katrina (UPSH) reported that its just her teaching style that made me do so well. In contrast to these selfhandicappers, students demonstrating downward shifts in self-handicapping were less inclined to see their academic outcomes as due to factors external to them. Cathy (DNSH) reported that the lecturers marking was fair and that whilst there were some tough markers, relative to what everyone else is getting I think its a fair mark. Patrick (DNSH) reported that it comes down to how much study you do . . . I dont 281

really see luck as having too much importance. Whether you like uni or if you dont, its how much effort you put in.

Much like the UPSH students, external attributions for academic outcomes were also salient in second year for students demonstrating upward shifts in defensive expectations. Craig (UPDE), who had failed the majority of his subjects in second year, reported that these grades were entirely attributable to outside forces. In relation to her grades in second year, Isabelle (UPDE) implied that her results were due to unfair marking: Maybe the lecturers know what they want . . . its just the way the lecturer perceives that assignment. Samantha (UPDE) too attributed her grades to luck and factors such as teaching: When asked what contributed to her success, she replied, Id say a lot of it is luck and with one subject the teaching was outstanding. When Theresa (UPDE) was asked why she failed one of her subjects, she reported that it was due to bad luck. When asked about her good grades, she responded, Maybe its luck. Maybe its my Mums prayers. In contrast to UPDE students who clearly saw external factors as salient in influencing academic outcomes, students demonstrating downward shifts in defensive expectations were less likely to see external factors as salient in their academic lives. Reports by these students reflected the fact that they were exerting control over their studies and assuming much greater responsibility for their study. According to Ellie (DNDE), you can never justify saying that this lecturer is an easier marker. I always find that if I put the effort in, I get the marks. An internal attributional profile was reflected in Joans (DNDE) comments. According to her, its sort of an inside personal thing that the more I work, the more time I put in, and the more I concentrate on a specific thing, the better Im going to do on the actual subject.

10.3.2.2 The salience of and shifts in uncertain personal control Respondents were asked about the certainty of their personal control in achieving success and avoiding failure. For the most part, students evincing upward shifts in self-handicapping reported being uncertain in their personal control in second year and also reported that their uncertainty was greater in second year than it was in first year. The reasons for this uncertainty seemed to vary. Some students cited the perceived independence of effort and outcome, while others cited low confidence and poor self282

esteem. For example, Dominique (UPSH) was no longer certain of grades she would receive: Even if I do think Ill do very well, it turns out the opposite. You know, I think Ill get a really good mark and it ends up a fail. Thats why Ive given up thinking, Ill definitely pass. Students demonstrating downward shifts in selfhandicapping reported greater certainty over their studies in second year than first year. This increased certainty was largely a result of their increasing familiarity with the learning environment. As a consequence, they gained confidence in their capacity to meet upcoming challenges and went about their studies in a much more organised and efficient fashion. For example, in second year Cathy (DNSH) has learnt to become an independent learner and is exerting more control over her studies. Glen (DNSH) and Helen (DNSH) reported that their increasing familiarity with the learning environment contributed to their control. According to Helen, I feel more in control because Im more familiar with the system, more used to it. I know the game. I know the library better. I know where the rooms are. The lecturers know me - I know the system.

Much like the high self-handicappers, students who were higher in defensive expectations in second year were relatively uncertain in their personal control. Craig (UPDE), whose personal control was shaky and uncertain, was a good example. In first semester second year, Craig (UPDE) reported, I didnt get a grasp of the gist of what they were talking about. Like I was in a daze about all the things happening. Because of this, his control in second year is way more uncertain than in first year. Students evincing downward shifts in defensive expectations were quite certain in their personal control in second year and generally reported that this was the case to a greater extent than first year. Much like the DNSH students, these students seemed to have difficulty adjusting in first year and now in second year they had found their feet and a concomitant elevation in perceived control. For example, Joan (DNDE) reported that Im pretty certain about it [her personal control] and that this is because

I can see that what I work to achieve, Ive achieved. Its sort of an inside personal thing . . . Last year I was still stepping, still feeling everything out,

283

working out whats going on and whats expected. This year I feel I have more control from what I had last year for sure.

10.3.2.3 The salience of and shifts in task-orientation Students were asked about the extent to which learning, gaining new knowledge and skills, and solving problems were important to them and satisfied them. Danny (UPSH) reported that at the time of the Time 2 survey, I didnt really care. All I wanted to do was just do an assignment, not so much to get anything out of it, but just to hand it in and get a Pass. The data in relation to other UPSHs, on the other hand, reflected a preference for mastery rather than performance. For example, contrary to indications in her questionnaire data, Dominique (UPSH) reported that mastery was quite important and that outperforming others was not so important. In contrast to the mixed findings in relation to the UPSH students, students shifting downwards in selfhandicapping were quite clearly task-oriented. Generally, these students recognised that mastery was not only important to their academic development but also gained a good deal of satisfaction in mastering given material and extending their knowledge and skills. For example, Cathy (DNSH) reported that mastery was crucial to her because, Ill have to transfer that knowledge to other [school] students. Also, she is more task-oriented in second year because subjects she is studying are more relevant to her future career as an educator and as a result she has more interest in them.

Much like the UPSH students, the extent to which students shifting upwards in defensive expectations were interested in mastery was mixed. Some reported that mastery was pivotal to their capacity to perform successfully in the workplace while others reported that mastery was not as salient in their academic lives in second year compared to first year. Craig (UPDE) was not so interested in mastery and reported being more interested in getting through. Whilst Samantha (UPDE) was interested in mastery, this only referred to subjects in which she was interested. Moreover, mastery was important to her only to the extent that it contributed to her performing better and enhanced her curriculum vitae. On the other hand, Isabelle (UPDE) commented that mastery was important because if you dont master the information, then youre not going to know it when you go out in the workforce. Those evincing downward shifts in defensive expectations, much like those shifting downwards in 284

self-handicapping, were quite clearly task-oriented. As with previous reports, some students recognised the relevance of mastery to their effectiveness as educators and with this in mind, gained satisfaction in developing skills and gaining new knowledge. According to Joan (DNDE), mastery is more salient in second year: Last year it wasnt a major issue to me . . . this year Im thinking that the more knowledge I get, the more Im going to pass onto kids when I get out in the classroom. Similarly, Martina (DNDE) reported that with the demands of employment looming, by the time youre in second year, youre getting close to the end and I think, Stuff it, if I dont ask now, Im never going to know. So you tend to lose a lot of your inhibitions. I really want to learn - after all, thats why Im here.

10.3.2.4 The salience of and shifts in performance orientation Students were asked about the extent to which performance-related factors (such as outperforming others, demonstrating competence, being seen to be competent, and performing to avoid poor performance and its negative implications) were salient in their academic lives in second year. On the whole, students demonstrating upward shifts in self-handicapping reported that such factors were important to them and that they had also experienced upward shifts in their performance orientation since first year. For example, performance, according to Dominique (UPSH), was very, very important, while Jodie (UPSH) reported that how she performs influences the way she views herself. Students selected as shifting downwards in their self-handicapping were not so concerned with how they were seen by others and their ability to outperform others. Generally, these students seemed more inclined to adhere to private standards and rely on personal criteria for feelings of satisfaction. According to Cathy (DNSH), when she is presented with a difficult problem, Im not worried about approaching the teacher. Im not worried about them thinking, This girls stupid. In terms of competitiveness, Helen (DNSH) reported, Im not interested in outperforming others. The grades are not comparative as far as Im concerned.

Performance-related concerns were salient in second year for students evincing upward shifts in defensive expectations. UPDE students also indicated that this was the case to a greater degree than when they were in first year. In general, these students seemed to be concerned about how they were viewed by others and were 285

particularly keen not to be seen as incompetent. Craig (UPDE) was a case in point: Im not one to seek help from a lecturer. I feel if you seek help, the lecturer will maybe think youre a dumb-arse and I dont want that to happen. When asked if avoiding being thought of as dumb was a priority, Craig (UPDE) reported, maybe I am [dumb], but I just dont want other people to know. Also, while Theresa (UPDE) recognised the importance of mastery, she concluded, I think performing well. Just because of the self-esteem again. I have to have people say to me, Oh youre doing well, for me to believe it. Moreover, when reflecting on her performance, she was fearful of people judging me if its not good and that because Ive only got Passes, I get scared that theyre going to think the worse of me.

Downward shifts in defensive expectations were not associated with such concerns with performance. DNDE students were generally not concerned with how they performed relative to others and similar to the DNSH students, seemed more interested in adhering to personal standards. According to Ellie, if I feel like Ive done well, it doesnt bother me what anybody else is doing. If I got sixty percent and a friend got eighty percent, depending on the work Id put into it, Id be happy for my friend. You get whatever you deserve. Other DNDE students accounts did not reflect the competitive concerns that were more typical of UPDE students. Joan (DNDE) reported that she was less performance-oriented in second year:

Last year I was still a bit competitive because Id just come out of high school and I was still used to that and Id be running up to others and saying, Oh what did you get for that assignment or that subject? . . . This year I havent worried about it.

10.3.2.5 The salience of and shifts in anxiety At Time 2, anxiety was incorporated into the quantitative process model and was found to predict self-handicapping and defensive expectations. Anxiety was not addressed at Time 1 and so interview data presented here are dealt with in greater detail than other affective and motivational constructs which were dealt with in Time 1 interviews. When questioned about their anxiety, students exhibiting upward shifts in self-handicapping seemed to identify with the profile of the anxious and stressed 286

student. In response to questions as to whether she experienced anxiety in second year, Dominique (UPSH) replied,

of course, who doesnt? Lots of stress. From the beginning I see the assignments and I go, Oh my God, what am I going to do? Every second its on my mind, What am I going to do? What am I going to do?

Other UPSH students were also well aware of the anxiety in their academic lives. For example, Jodie (UPSH) has found that her anxiety has become so salient in second year that she is now taking steps to alleviate it by organising herself better. Her account was reflective of classic test anxiety: I study and know everything, but when I get in there, I think, Oh, I didnt look at this and I didnt look at this either. I freeze up. Katrina (UPSH) also has experienced great anxiety in previous years, so she has made stress management a priority in second year and feels she is now managing this anxiety better than she has in the past. In contrast to Jodie and Katrina who have tried to address their anxiety, Danny (UPSH) surrendered to his anxiety and his account bordered on failure acceptance. While he was quite anxious about his studies at the beginning of second year, towards the end of first semester in second year, his anxiety had dissipated because he had virtually given up: Towards the end, I just didnt really care.

Whilst students exhibiting downward shifts in self-handicapping were not without anxiety, they were distinct from students shifting upwards in that they seemed to be less anxious in second than first year. Consistent with their comments regarding uncertain personal control, now that DNSH students had become more familiar and comfortable with university life, their anxiety had dissipated. Cathy (DNSH) reported that she was definitely more anxious in first year because of her uncertainty regarding exam formats and lecturers expectations of her. While Cathy (DNSH) reported that anxiety is still present in second year, this is less the case than first year. Glen (DNSH) also felt overwhelmed in first year, reporting that he was more anxious then because he was just thrown into it [university] with a whole bunch of students . . . I never knew where I was going. By contrast,

287

Im less anxious this year cause Ive gone through one year of uni and I know how uni exams are. Last year I had no idea how exams were. And you didnt know how they were going to mark you or what sort of exam you were going to face. So I was stressed over that.

For similar reasons Patrick (DNSH) experienced less anxiety in second year than first year: Im definitely less anxious because now Im aware of the system and how they mark.

Students higher in defensive expectations in second year saw anxiety as a more salient factor in their academic lives in second year. Much of this anxiety revolved around exams but was also present when assignments were due. For some, this anxiety seemed to interfere with their ability to address the task at hand. Craig and Samantha were two cases in point. Craig (UPDE) reported that in second year he was very stressed out . . . not understanding things that had been presented. Not being able to actually think about whats been presented. Not really confident. When asked how this years anxiety compared to last years, Craig (UPDE) reported that he was way more stressed. Samantha (UPDE) identified assignments as a perpetual stressing thing for me. When theyre due, I get really nervous and I always think that for some reason Im off track, when Im really on-track . . . I put more stress on myself. In terms of test anxiety, while Samantha (UPDE) reported that I just quiver in exams - I just fall to pieces, she performs well in them. In fact, Samantha (UPDE) has abandoned strategies that would typically mitigate anxiety associated with exams and assignments. For example, she has found that preparing too early for exams and assignments can induce anxiety: I tried planning on a white board, but it would just make me nervous, so I rubbed it all off and got rid of it - I didnt need to see that in my office at home. Much like Samantha, Isabelle (UPDE) has found that approaching assignment deadlines are reasons for panic stations. As the due date approaches, she thinks, Ah its due! Oh Ive got to get that done. Have I got everything ready for it? She then proceeds to invest effort into these assignments which in turn alleviates the anxiety. Theresa (UPDE) is also much like Samantha (UPDE) in terms of exams: I cant eat the night before an exam. All night I have nightmares that Im going to fail and that Dads going to come after me . . . I get 288

intimidated really quickly and think the worst and so I get very very anxious. Theresa (UPDE) also reported that she experiences more anxiety in second year than she did in first year.

Students evincing downward shifts in their defensive expectations were not as anxious in second year as those evincing upward shifts and their experience of anxiety in second year was not as high as that in first year. Again, the recurring theme of experience and familiarity with academic demands and a confidence in their ability to meet them was pivotal to the downward shifts in anxiety from first to second year. Reports by Joan and Martina (DNDEs) reflected this. Joan (DNDE) reported that when exam and assignment time comes, Im pretty relaxed about the whole thing . . . ever since second semester last year Ive sort of relaxed a lot because Ive realised what is involved in exams and assignments at uni and what is required. Martina (DNDE) was much the same: Now Im more or less used to the way uni works and you just go with the flow. Moreover, Martina (DNDE) indicated that her stress was more energising than debilitating: I love getting stressed. I need to feel the pressure on me when Im doing things. Other students seemed to find an enthusiasm that was not present in their first year. Ellie (DNDE) was illustrative: This year I dont stress anymore about exams - I know what Im going to do and Im enthusiastic about things this year, whereas last year I dont think I was enthusiastic. Similarly, Stephanies (DNDE) account reflected an enthusiasm and interest in her subject matter: Actually, there are a lot of assignments that I enjoy reading for.

10.3.3 Distinguishing features of first and second year university While much of the Time 2 interviews focused on the constructs that were salient in the process and longitudinal models, they were also concerned with exploring, from respondents own perspectives, the reasons for the identified shifts between first and second year. Thus far, the data show that features of students experiences of first and second year were influential in their use of self-handicapping and defensive expectations. In the light of this, respondents were asked more directly about their first and second years at university.

10.3.3.1 Upward shifts in self-handicapping 289

Students exhibiting upward shifts in self-handicapping generally did not find first year as difficult as those exhibiting downward shifts. Danny (UPSH) found first year was good and enjoyable. Jodie (UPSH) also enjoyed it. For Dominique (UPSH), first year was really good. According to her, in first year I was really excited and I did quite well in it too. Rather different accounts, however, were provided by UPSH students about second year. Some found that the academic demands were greater in second than first year. For example, Danny (UPSH) reported that he felt more pressure on him in second year than first year because everythings due at the same time this year. He also felt that the support from lecturers was not the same as first year: I think the lecturers arent as flexible. A similar account was provided by Dominique (UPSH) who said of second year:

[It is] harder . . . A lot of stress and lots of requirements . . . the assignments arent as straight-forward as the first year. They dont tell you much. They give you the paper and say that its up to you to find what to do . . . youre confused. You dont know what to do . . . Its tougher than first year.

Difficult personal issues also marked second from first year for UPSH students. These difficulties seemed to leave their academic motivation flagging. For example, the transition between first and second year was a tumultuous one for Danny (UPSH):

I got kicked out of home twice. Im back there now - so twice in the space of six weeks. Then at one stage I lived with my father and that didnt work out. So all this stuff was going on and it was new to me cause all through my life Id had nothing go wrong.

Other UPSH students identified failure in first year as a blow to their confidence and subsequent capacity to function effectively in second year. For example, failing practicum in first year was a great blow to Katrina (UPSH). In fact, her practicum was so traumatic that I was basically burnt out in the third week of prac and I actually had to take a day off cause I got so burnt out I got sick. I couldnt eat. I couldnt sleep. It just got on top of me. This coupled with the fact that I didnt have a qualified person supervising me, meant that the prac was just a disaster. While difficult personal 290

issues do not feature in most students accounts, the data show that when they arise, they impact on students academic lives in quite substantial ways.

10.3.3.2 Downward shifts in self-handicapping In contrast to the UPSH students, students shifting downwards in self-handicapping from first to second year tended to find their first year at university to be quite difficult. Generally, these students found the transition from school to university as academically demanding and socially isolating. Some even questioned whether they had made the right decision to study at university. Cathy (DNSH) found it to be tough and because the subjects in first year were not related to her intended career in Physical Education, I didnt enjoy it much. In fact, she became so disillusioned that there were stages when I thought, Yeah, Im going to drop out. I cant do this anymore . . . It was very different - such a transition from school to uni blew me away. It was such a big change. She also found it socially isolating: First year was lonely. Scary. You dont know who to approach . . . So yeah, first year was pretty tough. Hard to get to know people. Glen (DNSH) found university much more demanding than high school and had difficulty with the hectic workload and demands. Patrick (DNSH) found much the same: The first few months I hated it. I think cause university is so big . . . In the first few months I thought Id made a bad mistake. In a different way, Helen (DNSH) found her first year was difficult in that it lacked stimulation and challenge and was boring, very boring.

For DNSH students, second year was a much better one in terms of being more confident about their decision to study at university and their capacity to successfully negotiate the academic demands assigned to them. According to Cathy (DNSH),

second year is more enjoyable. You know your lecturers one-on-one cause there are smaller classes, so you feel like they care . . . Its more enjoyable and youre more focused. You know where youre going, whereas last year I was really vague. Like, whats happening, what am I doing here?

This issue of experience was a common theme amongst the DNSH students. A major factor that assisted their passage between first and second year seemed to be the 291

experience they accrued in first year. For example, according to Glen (DNSH), I knew what it was like and what to expect. Similarly, Helen (DNSH) found that accruing one year of university experience was a help to her and that she felt she was on the downhill run. The amount left to do isnt as overwhelming as first year.

The social isolation some experienced in first year was not so marked now that they were in second year. For example, Cathy (DNSH) did not experience the loneliness in second year that she did in first year. She also found that with a years academic experience behind her, she has found that her study skills have improved: Im learning a bit better. My learning strategies are more efficient. Patrick (DNSH) also identified making friends in second year as a reason for positive shifts between first and second year: I think friends at uni are important in you having a positive perception of uni. You realise that other people are in the same boat doing the same assignments. He is also more familiar with his learning environment. In first year, youre a bit disoriented, you dont know how the system works, but as time goes on you settle into it, you know where you stand.

292

10.3.3.3 Upward shifts in defensive expectations Much like the UPSH students, first year was generally an easier and more enjoyable one than second year for students evincing upward shifts in defensive expectations. These students generally found the passage between school and university a relatively smooth one. For example, according to Craig (UPDE), I settled well into uni. It was a pretty good transition from Year 12 to uni. The lecturers were pretty helpful. The subjects were relatively easy, not that hard. I found it a pretty good experience According to Samantha (UPDE), last year was a breeze cause I had given up work and she could concentrate on her studies.

But this was not the case now that UPDE students were in second year. Similar to UPSH students, UPDE students tended to find second year to be more demanding. According to Craig (UPDE), second year is way harder than first year. Theres more work . . . Youre not helped as much by your lecturers cause they think youve done your first year and you should know whats expected of you. According to Samantha (UPDE), I think its harder because every day I want to give up. I just cant see the end and you just get more work.

Similar to some UPSHs, difficult life events also marked one UPDE students transition between first and second year. In the Christmas break at the end of first year, Craig (UPDE) went overseas which

messed me up . . . I had a lot of freedom in [Country X]. I think my mind was more there in [Country X] than it was here . . . Things happened when I was in [Country X] - personal things happened - and I think that affected me in a bad way and so my whole life outlook changed. So first semester [second year] was shocking.

Difficulties with compulsory aspects of the course impacted on Samanthas (UPDE) general approach to her studies. Much the same as Katrina (UPSH), Samantha (UPDE) has been experiencing great difficulty with her practicum. She has been assigned to a DSP (Disadvantaged School Program) school which is

293

the worst of them all. I guess thats the bottom line to why I feel so negative. Theyre unruly, theyre out of control. Im scared for my own life. Im terrified of being in that environment and I dont know how to handle it cause nobody here will help me. Its taking the pleasure out of my prac Im dreading going there.

Also, the fact that many of her assignments are based on the practicum means that other areas of her academic life are suffering.

10.3.3.4 Downward shifts in defensive expectations In contrast to the accounts of UPDE students, students selected as exhibiting downward shifts in their defensive expectations from first to second year found first year more difficult than second year. These accounts were broadly parallel to those provided by the DNSH students: Some questioned whether they had made the right decision even being at university, let alone doing the right course, while others found the transition between school and university to be quite difficult. Ellie, Stephanie, and Martina (DNDEs) were illustrative. Ellie (DNDE) reported that in first year I knew from Day 1 I was depressed and I didnt want to be here. It was like a waste of time . . . If I didnt have to do them [some subjects] to get my degree, Id be doing something else. She also was unsure of her place at university: Last year I was trying to figure out what the hell I was doing here. Stephanie (DNDE) was much the same. First year for her was, What am I doing here? . . . It was just very hard. Also, she did not find the lecturers as supportive as she expected. Martina (DNDE) also had difficulty adjusting to the change in support given to her between school and university; at school she felt the teachers are really standing behind you all the way, whereas at university it was very much up to her.

These students accounts of second year, however, were quite different. Generally, DNDE students were now well familiar and comfortable not only with the academic demands before them but also the university environment in general. Moreover, the extent to which the course was relevant to their careers and capacity to function in the workplace also marked DNDE students second year from first year. For example, Joan (DNDE) has found that because she is more involved in subjects that are directly 294

relevant to her career as a teacher, she is more enthusiastic about second year. Similarly, Martina (DNDE) is increasingly seeing the implications her coursework holds for her future career in teaching. In second year, she realises she is only a year away from graduating and that she must focus on her studies more now than she did in first year:

When it really comes down to it you think, Well, this time next year Im going to be out there, Im going to be teaching kids and if I dont know what the hell Im doing, Im just going to fall down. So I have to really settle down.

Stephanie (DNDE) identified her more advanced use of resources and ability to learn independently as demarcating second from first year. According to her, the difference between first and second year was

knowing what to skim over and what to take in. Last year I would sit there and read the whole thing and spend hours on one book and not get anything out of it. Whereas this year, I can flick through a book and think, Yeah that looks good and photocopy that page. So its a lot less time wasted and I know how to use the time a lot better.

10.3.4 The role of cultural factors Data derived from interviews at Time 1 indicated that cultural factors may also influence the ways students go about their studies. At Time 1, however, this issue was not directly addressed and one purpose of the Time 2 qualitative study was to explore it more fully. Three students in the Time 2 qualitative sample were from nonAustralian backgrounds (Dominique [UPSH] - Middle East; Theresa [UPDE] Eastern Europe; Ellie [DNDE] - Southern Europe). These students were asked about the extent to which their cultural background influenced the way they go about their studies. The purpose of presenting these data is not so much to imply cause and effect - the students involved did not make any direct links between their cultural (and family) background and self-handicapping or defensive expectations. Rather, this

295

component of analyses is more supplementary and the derived data are suggestive of a need for further in-depth analyses into the role of culture in students academic lives.

Dominique (UPSH), who had first completed one year of a Science degree - which she did not enjoy at all - reported that she enrolled in Science because her parents believed it to be a higher status career than teaching and she felt that this concern with status was culturally derived. Importantly, she felt that there was great pressure on her to perform well and when asked how her culture would view poor performance she replied:

Its terrible if somebody fails subjects at uni. Failing means youre a loser. It was a lot of pressure for me to get to uni - mostly cultural. And now that youre going to uni youre now very high according to other [Middle Eastern] people . . . To get to university is more pressure than anything, cause I thought it was a matter of life and death. Once I got into it, I have to keep on being successful.

Thus, while not directly linking cultural expectations to her self-handicapping, Dominique did feel extreme pressure on her to perform well and to avoid at all costs that which her culture sees as particularly ignominious: failure.

Theresa (UPDE) feels that her cultural background has influenced her tendency to hold unrealistically negative expectations. She has found that holding and communicating such expectations is a way of regulating her familys (particularly her fathers) very high expectations of her. The reason for their high expectations, according to her, is based on family and cultural pride and a fear of what others in their ethnic group will think if she does not perform highly:

He [her father] wants me to do the best that I can, but my best isnt good enough . . . Dads a bit, I expect more from my daughter, we expect you not to shame us or the community. Its a shame I couldnt be a doctor. That was his goal [for me]. To be a doctor. I have all this on my shoulders.

296

In fact, when she does do well, her father does not commend her:

At uni, people are telling me, Youre doing well, then I go home and my Dad says, Im not going to praise you cause its going to go to your head. So Im kind of in two worlds. I dont know who to believe.

To minimise the impact of her familys high expectations of her, she does not tell them when she has an exam,

cause everyone will put pressure on me . . . In second semester I had two exams and I passed them because I didnt tell them [her parents]. Last semester, the same: I passed because I didnt have that pressure on me. Dont get me wrong, I hate lying . . . I have to go to confession every week and say, I lied, I lied.

Ellie (DNDE) has found that her family and cultural background has instilled in her a value of the education she receives in Australia:

My Dad, hed say, In the [Country Y] system we did this and we did that, and I think, Yeah, alright Dad. But in a way it does, because when I went overseas last year [to her parents homeland] I found that they value their education, whereas we take it for granted . . . I think in Australia its good to have opportunity, thats great. But sometimes too much opportunity can go to someones head and they forget the struggle.

Now in second year, Ellie (DNDE) has found that her parents values have influenced her ability to settle down from a previous year in which she did not apply herself to her studies. According to Ellie, their value of education did come through this year.

10.4 Summary of Time 2 qualitative findings The Time 2 qualitative study utilised the longitudinal data in a way that large shifts in self-handicapping and defensive expectations could be further explored from students personal perspectives. These data (see Table 10.3 for a summary) were particularly 297

illuminating in the ways upward and downward shifts in self-handicapping and defensive expectations were accompanied by concomitant shifts in an external attributional orientation, uncertain personal control, task and performance orientation, and anxiety. In addition to exploring students academic lives within the parameters of the process model predictors, the Time 2 interviews were sufficiently open-ended to enable respondents to offer their own insights as to reasons why the identified shifts occurred as they did. These data illustrated the difficulties students experience in making the transition from school to university, their social isolation, the rigours of academic demands, the powerful impact of academic failure (and fear of it), and in some cases, aspects of their personal lives.

The Time 2 interview data also illustrated how students defensive manoeuvring is very responsive to changes in their personal and academic fortunes: For example, students who reported personal difficulties or had experienced particularly ignominious failure were inclined to respond through self-handicapping or defensive expectations. On the other hand, students who in second year were now familiar with their learning environment and confident in their capacity to meet academic challenges assigned to them, were less inclined to manoeuvre in a self-protective fashion. Supplementary analyses at Time 2 involved a closer assessment of the influence of culture. Data in relation to this indicated that not only can students cultural (and family) backgrounds impose additional pressures on them but these influences can also impact positively on their academic lives by instilling in them an adaptive work ethic.

Essentially, then, the strength of the Time 1 and Time 2 qualitative studies lies in their exploration of self-handicappers and defensively expectant students motivations and behaviours. In listening to students stories about the diverse ways they go about their studies in a self-protective manner, we can understand better the factors other than ability which lead to success or failure at university. Moreover, through students stories, we are reminded that the essence of any large scale exercise in data collection (such as the Times 1 and 2 quantitative studies) has at its crux individual lives comprising unique personal and academic experiences. The qualitative data have been crucial in reflecting these individual and unique accounts.

298

Table 10.3 Summary of Time 2 qualitative findings

The issue Salience of and shifts in strategies UPSHs and UPDEs12 UPSHs reported that strategies such as procrastination were salient in second year and that this was more the case than first year. UPDEs reported that despite generally successful performance in first year, they were pessimistic about their performance in second year. External attributional orientation Uncertain personal control UPSHs and UPDEs saw outcomes as due to a variety of external factors and this was the case to a greater degree in second than first year. For a variety of reasons, UPSHs and UPDEs were uncertain about their personal control in second year and that compared to first year, this was the case to a greater degree. Task and performance orientation Whilst UPSHs and UPDEs for the most part were less task-oriented and more performance oriented, in some respects these findings were qualified. Anxiety UPSHs and UPDEs were generally anxious students who for a variety of reasons were more anxious and reported greater stress in second year. First and second year and key events and factors in these years Generally, UPSHs and UPDEs found second year to be more difficult than first year. UPSHs and UPDEs identified a variety of key events and factors (e.g., personal problems, failure in some subjects) that made second year particularly difficult for them.
12

DNSHs and DNDEs DNSHs reported that the extent to which they impeded their chances of success had decreased from first year. DNDEs tended to be more optimistic in second year than they were in first year.

DNSHs and DNDEs were inclined to see academic outcomes as due to factors within their control. DNSHs and DNDEs tended to be certain about their ability to achieve success and avoid failure.

DNSHs and DNDEs were quite clearly task-oriented in second year and this was more the case than in first year. The inverse was generally the case for performance orientation. Whilst DNSHs and DNDEs reported some anxiety, this was not as marked as anxiety reported by UPSHs and UPDEs. DNSHs and DNDEs found the transition from school to first year quite difficult and were more settled in second year. They identified key events and factors (e.g., familiarity with university demands) that marked first year as a difficult year and second year as more manageable.

Trends for UPSHs and UPDEs were deemed sufficiently comparable to deal with them conjointly. The same applies to DNSHs and DNDEs

299

11. DISCUSSION OF TIME 2 FINDINGS

The central aims of Time 2 quantitative analyses were to examine the Time 2 process model and to test the proposed longitudinal model. The hypothesised structure of the Time 2 process model was very much consistent with that of the Time 1 model: Affective and motivational factors were represented as first and second order factors which predicted self-handicapping and defensive expectations and reflectivity. These three strategies in turn predicted self-regulation, persistence, and future academic plans. Anxiety was also introduced as a predictor in the Time 2 process model. The longitudinal model was one which represented these processes in the one analysis across Times 1 and 2 and also incorporated grades as an additional outcome measure.

11.1 The Time 2 process model Given that the Time 2 process model sought to replicate the Time 1 model, it is not surprising that many of the parameters are much the same as those obtained at Time 1. In view of this, the discussion presented earlier in relation to Time 1 findings (see Chapter 8) is also relevant here. Notwithstanding this, the present discussion briefly addresses the significance of the generality of the previous findings. Also discussed is the role of anxiety (not incorporated in Time 1 analyses) in predicting selfhandicapping, defensive expectations, and reflectivity.

11.1.1 Generality of process models across Time 1 and Time 2 The most striking feature of the Time 2 process model is the fact that, with few exceptions, it is very consistent with the Time 1 model. That the two process models operate in much the same way demonstrates the generalisability of the processes over time. This is significant for two primary reasons. First, given the present evidence, it seems that these processes are more generic across students academic lives and not just particular to their first year of study. Without such longitudinal data it could be speculated that the self-protective process might operate differently across year levels. For example, first year students are dealing with many issues that are not characteristic of second year (as Time 2 interview data clearly showed): They are immersed in a very different learning climate, are unfamiliar with the academic challenges before them, and are uncertain as to their capacity to meet these challenges. In a sense, then, 300

the stage is set for self-protective behaviour. Indeed, Thompson (1994) has argued that self-worth protection is a primary motive of students in a transition phase (such as that between high school and first year university). By second year, students are wellfamiliar with the learning climate, have a greater understanding of the challenges that are presented to them, and have some belief in their capacity to successfully meet these challenges. The generality of the processes over time, however, suggests that despite the marked differences between first and second year, the defence-related processes appear largely applicable to both.

The generality of the models over time is also significant because it shows that students are dealing with the issue of self-protection and its related processes throughout (at least half of) their undergraduate life. Essentially, then, experience of first year and a successful transition from first to second year does not inoculate them from threats to their self-worth and the various strategies they use to deal with these. That essentially the same process was played out in second year as it was in first year is testament to the fact that self-worth motivation is an ongoing process. Indeed, the fact that students must have passed first year to have been surveyed in second year indicates that the process is to some degree independent of academic success.

11.1.2 The role of anxiety in predicting academic strategies Incorporating anxiety (Pintrich & DeGroot, 1990) at Time 2 was an important addition to the process model. As it turned out, anxiety significantly predicted all three strategies, self-handicapping, defensive expectations, and reflectivity. Of these, anxiety most strongly (positively) predicted defensive expectations and this supports previous contentions (Garcia et al., 1995; Norem & Cantor, 1986b; Norem & Illingworth, 1993) that individuals may respond to their anxiety by strategically setting unrealistically low expectations. According to Norem and Cantor (1986b), defensive pessimism may be thought of as a method by which individuals are able to cope with their anxiety (p. 1209). The present data, then, confirm the hypothesised centrality of anxiety to individuals proclivity to hold defensive expectations.

Whilst the present data are consistent with Norem and Cantors (see also Eronen et al., 1998) contention that individuals respond to their anxiety with defensive expectations, 301

they are not consistent with their contention that it generally leads to adaptive outcomes. Rather, these defensive expectations negatively predicted self-regulation and persistence. Indeed, the fact that students are less inclined to self-regulate and persist in response to their defensive expectations to some extent challenges Norem and Cantors contention that they harness their anxiety as a motivating force (1986b, p. 1209). Motivated students are inclined to self-regulate and persist and in terms of the present data this was not the case for students high in defensive expectations.

This is not to suggest that responses to anxiety must necessarily be associated with lower self-regulation and persistence. Individuals can respond to their anxiety with reflectivity. In response to this thinking-through process, students were in turn more likely to engage in self-regulation. Indeed, previous contentions that defensive pessimists harness their anxiety to adaptive ends may be more applicable to the reflectivity component of the defensive pessimism construct than the defensive expectations component. This finding is also significant from an applied perspective. Reflectivity may represent a fruitful focus for interventions designed to assist students in dealing with their anxiety: They can be taught how to respond to this anxiety in an adaptive way by thinking through a variety of outcomes. This not only prepares them (cognitively and affectively) for these outcomes but also leads to self-regulatory behaviour that ultimately better positions them for academic success.

Anxiety, to a lesser, but significant extent, also positively predicted self-handicapping. Whilst this supports a good deal of previous research into self-handicapping (e.g., see Berglas, 1985, 1987; Jones & Berglas, 1978), it is perhaps surprising that it predicted self-handicapping to such a limited extent. Pivotal to interpreting this finding are the negative associations between self-handicapping and persistence, self-regulation, and grades (discussed below). One perspective on these associations is that the selfhandicappers have to a large extent abandoned efforts to engage in their studies and that this is indicative of failure acceptance rather than failure avoidance (Atkinson, 1964; Covington, 1992, 1997). Importantly, Covington (1997) has discussed how failure accepters tend to be lower in anxiety, hold doubts about their ability, have poor study skills, and spend little time studying. This profile closely resembles that of the self-handicapper identified in the present study. It may be that self-handicappers 302

experience less anxiety than students high in defensive expectations and reflectivity because they have accepted failure and no longer fear it. The hypothesised connection between self-handicapping and failure acceptance is further tested in the following chapter.

11.2 The longitudinal data Essentially the longitudinal model is a mediational one in which the constructs operate via the two process models across time. The longitudinal model, then, underscores the centrality of the hypothesised process models. One of the many strengths of structural equation modelling procedures is that in addition to simultaneously testing many parameters, they can also provide rigorous tests of overall model fit. The fact that the longitudinal model fit the data well indicates that the hypothesised processes across time (in addition to test-retest parameters) are a viable means of representing the longitudinal process of self-protection.

This longitudinal model also provides a new perspective on self-protection in that it explicitly assesses self-protection in an extended process-like fashion. Thus the longitudinal data provide support for Buss and Cantors (1989) model of the processes giving rise to behaviour which to date has not been operationalised to test substantive issues such as those in the present study. Moreover, the Buss and Cantor model has apparently not been assessed over time, and so in this sense, the present study extends their conceptualisation of the processes giving rise to behaviour.

Given that the longitudinal model essentially represents the two process models over time, earlier discussion of the majority of findings in relation to these two models also applies here. The focus of the following discussion, then, is upon the new relationships in the longitudinal model. These relationships are significant because they represent new effects of a number of variables beyond the substantial effects of these variables at Time 1. These relationships primarily involve the pivotal links between the two process models: The relationships between the Time 1 outcomes and Time 2 affective and motivational predictors and between both Time 1 strategies and outcomes and grades obtained at the end of students first year of study.

303

11.2.1 Predictors of academic grades The inclusion of academic grades in the study was an important extension of the outcome constructs incorporated in the model. The outcome measures to this point have been self-reports of self-regulation, persistence, and future academic plans and so no objective outcome data had been collected. At Time 2, however, data were collected on students grades for the end of their first year at university. These grades were incorporated into the longitudinal model such that Time 1 strategies and outcomes were freed to predict them. Results showed that the only strategy significantly predicting grades was self-handicapping (negatively) whilst the only Time 1 outcome predicting grades was future plans (negatively).

The link between self-handicapping and academic grades is consistent with the Time 1 and 2 process model data. Specifically, in line with the strong negative associations between self-handicapping and both self-regulation and persistence, self-handicapping in the longitudinal model was found to negatively predict grades. Indeed, this confirms self-handicappings status as the least adaptive strategy of the three strategies incorporated into the present study. This finding is particularly significant because the negative effects of self-handicapping to this point in the study have been couched in terms of self-report questionnaire constructs. The negative association with grades substantiates these findings and also shows that the negative impact of selfhandicapping is quite comprehensive, influencing both behavioural and cognitive engagement as well as actual performance. Whilst these findings are consistent with hypotheses presented earlier in the study, it must be emphasised that the present selfhandicapping measure is explicitly self-protective. Further research is required that assesses the effects of self-enhancing self-handicapping on self-regulation, persistence, and academic grades.

The finding that future academic plans negatively predicts academic grades is unexpected and given the nature of the present data it can only be speculated as to the reason. Most of the present sample are required to teach maths once they begin their duties as school teachers. It may be that those who report a greater willingness to pursue maths in further studies are those who are weakest in the area and therefore in most need of further education. It may not be surprising, then, that it is these students 304

whose grades are relatively lower. Further analyses were conducted to substantiate this speculation. If the negative relationship between future maths plans and grades is a result of a weakness in maths (and therefore domain specific), it therefore follows that future verbal plans should not negatively predict academic grades. To test this, the Time 1 Nepean data (which incorporated verbal measures) were reanlaysed with a view to ascertaining the extent to which future verbal plans predict grades. A process model in which future maths plans was replaced by future verbal plans indicated that the latter positively predicted academic grades (quite opposite to findings related to future maths plans). Moreover, analysis of data using maths-only grades as the outcome measure replicated the finding that both self-handicapping and future academic plans negatively predict grades. These results support speculation that it is the students who are more inclined to pursue maths in the future who are weaker in this area and therefore receive lower grades generally and also in maths specifically.

11.2.2 Factors predicting affective and motivational constructs at Time 2 Whilst achievement is typically lauded as an educational outcome, it may also constitute something of a double-edged sword: Because high achievement evoked a performance orientation and a performance orientation was positively associated with maladaptive self-protective strategies (e.g., self-handicapping), it appears that there is something about achievement which has the potential to lead the student into counterproductive manoeuvring. Three reasons are proposed to account for this. First, higher grades promote a focus on the performance dimensions of tasks rather than the content/mastery-related dimensions. Very much related to this is the fact that high achievement or communication to individuals about their competence also promote a competitive orientation (Harackiewicz et al., 1997), a performance orientation (Harackiewicz et al., 1998), or competence valuation (Harackiewicz & Manderlink, 1984; Harackiewicz, Sansone, & Manderlink, 1985) - three constructs subsumed under the higher order performance orientation factor in the present study. The second reason is that the esteem-relevant stakes are raised when students perform well and as a consequence they have more to lose. Given that the esteem-relevant risk they now face has arisen from successful performance, minimising this risk may be seen by the individual to be best achieved through performance-related processes rather than task/mastery-related processes. Third, according to Deci (1975; see also Berglas, 305

1985; Deci & Ryan, 1980; Ryan, 1982), extrinsic rewards shift the focus away from intrinsic orientations (e.g., task-orientation) and onto more extrinsic orientations (e.g., performance orientation). Whilst, grades are not extrinsic in the way that material rewards are (particularly because they provide feedback that has intrinsic value), they typically provide normative information that equates ones own performance with that of others and this may prompt a more extrinsic rather than intrinsic orientation. Further work is required to establish which of these possibilities is the most feasible in explaining the link between achievement and performance orientation.

Given that self-regulation is primarily concerned with activities such as planning ones approach and monitoring ones progress, it is more consistent with a mastery/task-related orientation and somewhat antithetical to a performance orientation. It is therefore not surprising that it negatively predicted performance orientation at Time 2. Persistence at Time 1 also played a key role in linking the two process models by predicting at Time 2 uncertain personal control (negatively), an external attributional orientation (negatively), and task-orientation (positively). The former two findings are consistent with the conceptual bases of these constructs: Individuals high in persistence actively exert control over the target outcome and are concomitantly less inclined to see these outcomes as beyond their control (hence the negative association with both uncertain personal control and external attributional orientation). Also, consistent with the findings of Miller et al. (1996), persistence in the face of challenge implies a focus on the task and an inclination towards mastery (hence the positive association with task-orientation).

These findings are important because not only do they show how the two process models are connected across time, but also set in place the self-protection process at Time 2. In this sense, then, the constructs argued to represent the academic outcomes at Times 1 and 2 are not solely ends in themselves but hold implications for the affective and motivational predictors in the following phase. When viewed in this light, the outcomes in the present study may be more accurately construed as mediating constructs than outcomes per se. Essentially this reflects the fluid nature of the process model over time such that outcomes at one point in time also represent predictors at another time point. This is not to suggest that the selection of the set of 306

predictors and designation of them as antecedents was arbitrary. On the contrary, the selection of predictor constructs was conducted on conceptual bases deriving from a model in which behaviour is influenced by the strategies one adopts which in turn are a response to a variety of affective and motivational factors (see Buss & Cantor, 1989). Present findings have shown that this is not just a face valid means of conceptualising the self-protective process but is also a defensible thesis on empirical grounds. Recognising the centrality of the processes involved in defensive manoeuvring and the way these processes unfold over time is essential for practitioners and researchers seeking to not only better understand students defensive manoeuvring but to target such behaviours in terms of intervention.

11.2.3 Test-retest parameters In addition to the paths comprising the two process models and the links between the two processes, the test-retest parameters were also freed. This involved correlating parallel Time 1 and 2 uniquenesses associated with each item and freeing the path between each Time 1 construct and its parallel Time 2 construct. Methodologically, this is important because it attenuates the positive bias in paths that would result if parallel constructs were not freed to be estimated (see Joreskog, 1979; Marsh et al., 1997; Marsh & Hau, 1996).

Assessing the test-retest paths in the structural model (see Figures 9.6 to 9.9) and testretest correlations (see Table 9.6) provides an important perspective on the stability of constructs in the model. The correlational data indicated that while all constructs in the model are relatively stable, there remains a reasonable degree of variance unexplained. Particular attention is directed to the test-retest correlations between the strategies for which over half the variance remains unexplained. Indeed, the primary objective of the present study was to account for this variance and it will be noted in Figure 9.6 that the test-retest strategy paths are much lower than those for other constructs indicating that the proposed model has done a good job of explaining the variance in the strategies. In one sense, the unexplained variance is reassuring because the defensive manoeuvring in which students engage is not so entrenched that interventions designed to remediate it cannot be successful. On the other hand, these results may suggest that the strategic way students go about their studies can be quite 307

fluid and that given a change in students circumstances, fortunes, or affective and motivational orientation, strategic behaviour that once was adaptive may no longer be so (see the Time 2 interview data).

11.2.4 Correlations among the strategies One further point to consider in relation to the strength of relationships is the generally higher correlations among the three strategies at Time 2 than at Time 1 (see Table 9.6). Indeed, this collinearity required that these strategies be considered separately in three submodels. The most notable example of this is the much higher correlation between reflectivity and defensive expectations at Time 2 (r=.69) than at Time 1 (r=.40 - see Table 9.6). In the review of literature it was discussed how simply thinking about negative outcomes (reflectivity) may not diminish performance whereas actually expecting negative outcomes is inimical to academic success (Goodhart, 1986; Sherman et al., 1981; Showers, 1992). It was further speculated that the negative effects of defensive pessimism in the longer term (Norem & Cantor, 1990b) could result because over time students reflectivity is increasingly accompanied by defensive expectations. The Time 2 correlations seem to support this contention in that defensive expectations and reflectivity at this time are not so distinct. Thus, whereas thinking-through in first year was more independent of defensive expectations, by the time students are in second year this independence is not so marked. This is not to suggest that reflectivity becomes less conducive to academic success over time. Rather, data imply that reflective students over time come to be increasingly defensively expectant. It must be recognised, then, that if the benefits to be derived through reflectivity are to be maintained over time, students should be made aware of the important distinction between simply thinking through outcomes and actually forming expectations about those outcomes.

11.3 Qualitative findings The preceding discussion relates primarily to the quantitative process and longitudinal models. The Time 2 qualitative data also provided key insights into the selfhandicapping and defensive expectations phenomena. The Time 2 qualitative study differed from the Time 1 study in that students were selected for interviews on the basis of upward or downward shifts in self-handicapping and defensive expectations 308

from first to second year. Following this, a series of research issues were explored that were centrally related to the exhibited shifts in self-handicapping and defensive expectations.

The Time 2 qualitative data demonstrated that shifts in defensive expectations and self-handicapping could be associated with concomitant shifts in the key predictors in the Time 2 quantitative process model: For the most part students evincing upward shifts in defensive expectations and self-handicapping reported that in second year, relative to first year, they were more inclined to see outcomes as due to external factors and perceived uncertain personal control over achieving success and avoiding failure. They were also more anxious and performance oriented, and less taskoriented.

The interview data also demonstrated how, consistent with defensive pessimism theory (see Norem & Cantor, 1986a, 1986b), students can use their defensive expectations as a motivator. As one student commented, her defensive expectations motivated her in a twisted kind of way. This is consistent with Time 1 qualitative data which showed that students high in defensive expectations were inclined to get anxious and harnessed this anxiety to motivate them into action. Also consistent with qualitative data at Time 1 and with present contentions regarding the protective dimensions of defensive expectations, some students maintained defensive expectations in second year so that they were not disappointed in the event of failure and were pleasantly surprised in the event of success (see Feather, 1969).

The data also highlighted the role of multiple goals. In the Time 2 qualitative analysis, findings in relation to task-orientation and performance orientation were not as clearcut as those related to other dimensions: Some students gained satisfaction in both mastery and achievement, reporting that both were equally important to them. Indeed, Harackiewicz et al. (1997) have shown that task and performance orientations can lead to different, yet adaptive, academic outcomes and that a well-rounded approach to ones study may be to recognise the importance of dual orientations. It may be advisable, then, to encourage students to learn how to coordinate the two goals (see Harackiewicz et al., 1998). Interestingly, the Time 2 qualitative data supported Time 1 309

qualitative findings which indicated that students saw mastery as pivotal to successful performance and that this was how the two were interconnected. This recognition that the two are not necessarily unrelated and that one is somewhat dependent on the other (see Harackiewicz et al., 1997) may be important to consider in attempts to understand how the two can be coordinated.

In addition to the shifts in key process-model predictors, respondents identified aspects of their first and second year university experiences that clearly demarcated one year from the other. Students evincing downward shifts in self-handicapping and defensive expectations tended to find first year quite difficult and second year to be much more manageable. First year for these students tended to be one of social isolation, unknown academic demands and expectations, an unfamiliar learning environment, and less support from staff than they were accustomed to at high school. In second year, these students reported that they had made a number of friends, were familiar with their learning environment and the academic demands presented to them, and had developed better learning and study techniques. These all serve to underscore the impact of the transition between school and university and are supportive of Thompsons (1994) claims that it is at times of transition when the motive to protect the self is paramount. Indeed, the identified downward shifts seem to support this in the sense that when students had settled into university life, their tendency to engage in self-protective strategies was much reduced.

The Time 2 qualitative data also underscored the powerful effects that failure can have on the way students go about their studies and the strategies they acquire in response to this failure. In a number of cases, students high in self-handicapping and defensive expectations had at some stage in the recent past experienced academic failure and this seemed to influence their tendency to self-protect. Covington and Omelich (1981) have highlighted the impact of failure on students. They found that failure negatively predicted perceptions of ability which in turn led to negative affect and lower expectations for success. Moreover, the more failure cannot be explained by factors other than ability, the more students are forced to acknowledge their low ability. It is not surprising, then, that students reporting experiences of failure in the Time 2

310

qualitative study tended to also be high in self-handicapping or defensive expectations:

Much like Time 1, Time 2 interview data superimposed a very human face on the study. This was particularly evident in cases where students evincing upward shifts in self-handicapping and defensive expectations did so as a consequence of quite difficult personal problems. These data show that the strategies students use when going about their studies do not operate in a vacuum, but rather, are responsive to events and factors operating in other areas of their lives. Indeed, it is important for educators seeking to address students self-protective manoeuvring to recognise the role of factors external to the learning environment which might be influencing students approaches to their study.

The qualitative data also provided some insights into the role of cultural factors in students academic lives. Three students for whom questions regarding the role of culture were relevant, reported that their cultures view of failure and concerns with status and how they were seen within their community, all placed a good deal of pressure on them and the way they went about their studies. The data also illustrated how family expectations (culturally derived) impacted on students own expectations and their tendency to regulate others expectations of them (see Baumgardner & Brownlee, 1987). Whilst some work has been conducted exploring the issue of selfhandicapping in African- and European-American cultures (Midgley et al., 1996), the qualitative data presented here have identified additional culture-related issues that offer some direction for further research into self-protection generally and both defensive expectations and self-handicapping in particular.

Much along the lines of the Time 1 interview data, the profiles of self-handicappers and students high in defensive expectations were broadly congruent. Consistent with the exhibited shifts between first and second year in self-handicapping and defensive expectations, were concomitant shifts on the key predictors (external attributional orientation, task-orientation etc.). Upward shifts in self-handicapping and defensive expectations were also associated with key events and factors (e.g., personal difficulties, failure) that marked students second year at university as more difficult 311

for them. Consistent with Time 1 interview data and the Time 1 and 2 quantitative data, however, defensive expectations were not so inimical to academic success as self-handicapping. For example, along similar lines to interviewees at Time 1, defensive expectations at Time 2 were seen in some cases as a strategy that enhances motivation. Thus, whilst both self-handicapping and defensive expectations are two strategies that students use to self-protect, defensive expectations, relative to selfhandicapping, seems to be a more complex phenomenon that can in some instances enhance motivation (see Time 1 and 2 interview data) and does not significantly impair academic performance (see Time 2 quantitative data).

11.4 Summary of chapter Data at Time 2 were important in demonstrating the generality of the Time 1 processes. Whilst there were some discrepancies between Time 1 and Time 2 process models, the similarities between them were such that the processes seem more generic than year-specific. Time 2 process data were also significant because they incorporated a measure of anxiety which was found to be positively associated with the three strategies at Time 2. The longitudinal data provided important insights into the self-protection process over time both in terms of the generality of constructs in students first and second years at university and in terms of the pivotal links between the two process models. Academic grades were also incorporated into the longitudinal model and these were negatively predicted by both self-handicapping and future academic plans.

The qualitative data were useful in illuminating students perspectives on shifts in their self-handicapping and defensive expectations and the variety of factors that accompanied these shifts. Utilising the Time 2 process model as a starting point, upward and downward shifts in self-handicapping and defensive expectations were accompanied by generally parallel shifts in students external attributional orientation, task and performance orientation, uncertain personal control, and anxiety. In addition to these factors, students identified aspects of first and second year which they felt related to the shifts in self-handicapping and defensive expectations and also provided insights into the effects of events which, whilst extraneous to university life,

312

nonetheless impacted on their tendency to defensively manoeuvre in the academic domain.

If the Time 1 and Time 2 quantitative and qualitative studies can be integrated on common conceptual bases, two dimensions which probably most accurately represent major findings are the motive to avoid failure and motive to approach success. The motive to avoid failure is the dominant orientation pervading the bulk of the present findings, but the data have demonstrated how a dual motive to succeed is also pertinent. Need achievement theory directly addresses these two motivations and is proposed to be a useful framework under which the three strategies can be meaningfully integrated. This is now addressed.

313

12. A QUADRIPOLAR NEED ACHIEVEMENT REPRESENTATION OF SELF-HANDICAPPING, DEFENSIVE EXPECTATIONS, AND REFLECTIVITY

12.1 Brief orientation to need achievement theory and a quadripolar representation Evolving from the Time 1 and 2 quantitative and qualitative analyses is the notion that students can be broadly characterised in terms of their motive to avoid failure and approach success. Indeed, such a conceptualisation is particularly relevant to the three strategies, self-handicapping, defensive expectations, and reflectivity, because the data thus far have indicated that while they vary primarily along failure avoidance dimensions, to varying degrees they can be underpinned by a success orientation. These motivations to approach success and avoid failure are best reflected in the need achievement model first proposed by Atkinson (1957, 1958, 1964; see also McClelland, 1965). The Atkinson model was later extended by Covington and Omelich (1991; see also Covington, 1992, 1997) and their work is particularly relevant to the present study because it integrates self-worth motivation theory with need achievement theory. The work by Covington and Omelich represents need achievement in a quadripolar model in which the two axes represent the motive to approach success and the motive to avoid failure. The basic structure of this model (outlined by Covington, 1992, 1997) is presented in Figure 12.1.
Failure avoidant
High

C. Failure avoider

B. Overstriver

Success oriented
Low D. Failure accepter High A. Optimist

Low

Figure 12.1 Quadripolar model of need achievement (Covington, 1992, 1997)

314

In this model, individuals are proposed to vary along two dimensions: Success orientation and failure avoidance. Success orientation refers to the extent to which an individual strives for and is motivated to attain success. Failure avoidance refers to the extent to which an individual is motivated to avoid failure and (in the context of selfworth motivation theory) its implications for the self-worth. In the first quadrant (Quadrant A) are individuals who are high in success orientation and low in failure avoidance. According to Covington and Omelich (1991), these individuals are generally optimistic. In Quadrant B are individuals high in both success orientation and failure avoidance. These individuals are referred to as overstrivers (Covington & Omelich, 1991) who for the most part achieve as highly as optimists but do so in part to avoid failure. Unlike optimists, overstrivers contemplate failure as well as positive outcomes and are particularly vulnerable in the face of failure as it tends to confirm doubts they have about their ability (Covington, 1992). In Quadrant C are individuals low in success orientation and high in failure avoidance. These individuals are failure avoiders whose motivation is to manoeuvre so as to avoid failure or alter the implications and meaning of that failure. In Quadrant D are individuals who are low in both success orientation and failure avoidance and are referred to as failure accepters. These individuals are generally disengaged from their studies and display a helpless pattern of motivation (see Abramson et al., 1978).

As noted, this model may provide a useful framework by which to integrate the three strategies. First, it is suggested that the overstriver shares characteristics with the reflective student. Because these individuals are both success oriented and failure avoidant (see also Arkin & Oleson, 1998), they tend to contemplate both success and failure - a hallmark of reflectivity. Second, the failure avoiders hold doubts about their ability and are not confident in their capacity to meet upcoming challenges. Thus they are more inclined to contemplate failure. In such circumstances, these individuals are inclined to engage in strategies designed to protect the self (Covington, 1992). Two such strategies involve setting defensive expectations and engaging in selfhandicapping.

315

Working anti-clockwise from Quadrant A, individuals achievement orientations become increasingly failure avoidant and then failure accepting. Whilst motivated towards success, individuals in Quadrant B are also failure avoidant and tend to be debilitated in the case of even isolated failure (Covington, 1992). Individuals in Quadrant C, on the other hand, are primarily failure avoidant and tend to engage in self-defeating strategies such as defensive expectations and self-handicapping aimed at altering the meaning or implications of failure. Moreover, it is proposed that defensive expectations are closer to the overstriving quadrant than the failure accepting quadrant because students high in defensive expectations do not actively set obstacles that obviate success and so are not necessarily opposed to it. On the other hand, self-handicappers are closer to the failure-accepting quadrant than the overstriving quadrant because they actually set obstacles that obviate success and so are also opposed to success orientation. While these predictions are speculative, it is proposed that the data in the present study enable empirical tests of these ideas.

From this perspective, then, the reflective, defensively expectant, and selfhandicapping students all have a failure-avoidance orientation underpinning them. In fact, the two extremes, reflectivity and self-handicapping, have been unified from a failure avoidance perspective not only quite recently by Arkin and Oleson (1998), but also in the original formulation by Jones and Berglas (1978). According to Jones and Berglas, self-handicappers and overachievers are alike in that each is fearful that failure will implicate competence. Each has an abnormal investment in the question of self-worth. One succeeds in avoiding failure through persistent effort, the other embraces failure as an alternative to self-implicating feedback (p. 205). Similarly, according to Arkin and Oleson (1998), while the two are phenotypically quite different - overachievers expend great effort and attempt to avoid failure at all costs, while self-handicappers withdraw effort and increase the likelihood of failure genotypically, the two types of behavior may be inspired by the same motivational force: self-doubt (p. 339).

It is further proposed that each quadrant represents an individuals achievement style which is largely independent of another quadrant (Covington & Omelich, 1991; Covington & Roberts, 1994). Indeed, Norem and Cantor (1986b) outline how 316

expectations, for example, can act as a cognitive set. In a similar fashion, ones achievement style might represent such a cognitive set. Moreover, students choose an achievement style with which they are comfortable and which for a variety of adaptive or not so adaptive reasons works in their service. Thus when a students typical achievement style competes with another style, that student finds it disrupting and in some cases aversive, and it may be that performance which was previously successful may no longer be so.

Thus the quadripolar model of need achievement is proposed here to be a useful means of mapping the three strategies. To test this, the item parcels for active selfhandicapping, active defensive expectations, and reflectivity are incorporated into subsequent analyses. Importantly, while the original quadripolar model (Covington & Omelich, 1991) also incorporated optimism, the present study comprised no such measure. For the purposes of exploring this need achievement framework, positive self-concept is operationalised as a dispositional equivalent. Thus the predicted two dimensional pattern of the four constructs (self-handicapping, defensive expectations, reflectivity, and self-concept) which is tested in the present study appears in Figure 12.2.

Success oriented
C. DEFENSIVE EXPECTATIONS (failure avoider) C. SELFHANDICAPPING (failure avoider) High D (failure accepter) A. HIGH SELF-CONCEPT (optimist) Low High B. REFLECTIVITY (overstriver) Low

Failure avoidant

Figure 12.2 Hypothesised quadripolar model of need achievement incorporating the present studys strategies
Note. Present studys constructs in caps; Covingtons (1992, 1997) constructs in parentheses

317

12.2 Results Multidimensional scaling was used to explore the three strategies from a two dimensional need achievement perspective. Multidimensional scaling assists the researcher in determining the perceived relative position of a set of objects or items13 (Hair, Anderson, Tatham, & Black, 1995). If two objects are rated by respondents as similar, they will be located in multidimensional space in a way that the distance between them is smaller than the distance between other pairs of objects. The resulting perceptual map indicates the relative positioning of all items. The researcher then interprets the underlying dimensions in a way that best explains the positioning of items in the map, particularly as it relates to an underlying theoretical rationale which in this case, is the quadripolar need achievement model. The present analysis used the multidimensional scaling procedure in SPSS which created the similarity matrix from the raw data (item pairs, consistent with previous analyses) using the Euclidean distance measure (Young & Harris, 1994). Given that the analysis seeks to test a two-dimensional model specified a priori, only a two dimensional model was examined.

12.2.1 Time 1 The Time 1 data fit the model well (RSQ=.98, S Stress=.068). Stimulus coordinates appear in Table 12.1 and are mapped in Figure 12.3. Results in Table 12.1 and Figure 12.3 to a large degree conform to the proposed quadripolar need achievement model. As predicted, self-concept was positioned in the optimism quadrant (high success orientation and low failure avoidance) and reflectivity was positioned in the overstriver quadrant (high success orientation and high failure avoidance). The defensive expectations item parcels were positioned in a way that primarily reflected high failure avoidance (but were partly positioned in the high success orientation dimension and partly in the low success orientation dimension). Not only was selfhandicapping diametrically opposed to a success orientation, its relative positioning actually bordered failure acceptance.

13

While multidimensional scaling is typically used to determine similarities amongst a set of objects (rather than self-report questionnaire items), it is considered appropriate for use in the present study not only in terms of its heuristic value, but also in terms of its focus on mapping constructs in multidimensional space as is relevant here.

301

Table 12.1 Time 1 stimulus coordinates for self-handicapping, defensive expectations, reflectivity, and self-concept

Item Pair Dimension 1 Dimension 2 .34 1.01 DE 1 -.22 .93 DE 3 .10 .97 DE 3 .26 1.01 DE 4 1.29 -.02 REF 1 1.49 .16 REF 2 1.17 .42 REF 3 -1.75 -.28 SH 1 -1.68 -.17 SH 2 -1.55 -.14 SH 3 -1.91 -.34 SH 4 .69 -1.08 SC 1 1.00 -1.25 SC 2 .77 -1.22 SC 3 Note. Consistent with CFA and SEM analyses, item parcels are used DE=Defensive Expectations; SH=Self-Handicapping; REF=Reflectivity; SC=Self-Concept
1.5 1.0 dp8.9 dp1.2 dp3.4 dp6.7

Failure Avoidance

.5 0.0

ref6to8 ref4.5 ref2.3

sh7.9 sh3.4 sh1.2 h10.11

-.5 -1.0 -1.5 -2.0 sc1.2 sc5.6sc3.4 -1.5 -1.0 -.5 0.0 .5 1.0 1.5 2.0

Success Orientation
Figure 12.3 Time 1 multidimensional mapping of the three strategies and self-concept
Note. Ref=Reflectivity; DP=Defensive expectations; SH=Self-handicapping; SC=Self-concept

302

Table 12.2 Time 2 stimulus coordinates for self-handicapping, defensive expectations, reflectivity, and self-concept

Item Pair Dimension 1 Dimension 2 .12 .89 DE 1 -.62 .59 DE 2 -.26 .91 DE 3 -.34 .81 DE 4 1.50 .22 REF 1 1.50 .37 REF 2 .89 .62 REF 3 -1.64 -.52 SH 1 -1.59 -.38 SH 2 -1.51 -.32 SH 3 -1.83 -.51 SH 4 1.02 -.87 SC 1 1.60 -.88 SC 2 1.17 -.93 SC 3 DE=Defensive Expectations; SH=Self-Handicapping; REF=Reflectivity; SC=Self-Concept

1.0 tdp8.9 tdp3.4 tref6to8 tref4.5 tref2.3 0.0 tsh7.9 tsh3.4 sh10.11 tsh1.2

Failure Avoidance

.5

-.5

tsc1.2 tsc5.6 -1.0 -2.0 -1.5 -1.0 -.5 0.0 .5 1.0

tsc3.4 1.5 2.0

Success Orientation

Figure 12.4 Time 2 multidimensional mapping of the three strategies and self-concept
Note. tref=Reflectivity; tdp=Defensive expectations; tsh=Self-handicapping; tsc=Self-concept

303

12.2.2 Time 2 The same multidimensional scaling procedures were carried out using the Time 2 data. Again, the data fit the model well (RSQ=.99, S Stress=.05). Stimulus coordinates appear in Table 12.2 and are mapped in Figure 12.4. Time 2 data also conform to the proposed quadripolar need achievement model and are very much consistent with Time 1 findings. Self-concept was positioned in the optimism quadrant and reflectivity was positioned in the overstriver quadrant. Defensive expectations item parcels were located such that defensive expectations primarily reflected high failure avoidance. Self-handicapping was diametrically opposed to a success orientation and in much the same way as Time 1 data indicated, its relative positioning reflected failure acceptance.

This issue of failure acceptance is a new perspective on self-handicapping and for this reason is explored further. It is contended that a strong indicant of failure acceptance is withdrawal from ones studies (Covington & Omelich, 1991). A number of respondents present at Time 1 were not present at Time 2 and it is argued that one primary reason for students non-attendance at Time 2 is that they had withdrawn from their course14. Indeed, most withdrawals occur between the first and second years of candidature - precisely that period under focus in the present study. It is further proposed that non-attendance can be incorporated as an additional outcome construct in the Time 1 process model. It seemed important, then, to test the effects of students strategic manoeuvring on their non-attendance at Time 2. In particular, it is hypothesised that consistent with the above data, self-handicapping, reflecting failure acceptance, positively predicts non-attendance at Time 2.

Accordingly, in addition to estimating the Time 1 process model presented in Chapter 6 (see Figure 6.2), paths between strategies (self-handicapping, defensive
14

It is recognised that non-attendance at Time 2 could be for a number of reasons: Students may have withdrawn from their course, may have been ill at Time 2 questionnaire administration, or were cutting class. It therefore can only be speculated as to the reason why these students were not in attendance at Time 2 and thus non-attendance as an index of withdrawal is to some extent limited. Watkins & Hattie (1985) have discussed how there are three groups of students in follow-up studies: withdrawers, persisting non-respondents, and persisting respondents. The present data were such that they could not

304

expectations, and reflectivity) and Time 2 attendance (1=Present; 2=Not Present) were freed as were paths between Time 1 outcomes (self-regulation, persistence, and future plans) and attendance at Time 2. This model yielded an acceptable fit to the data (chi square=4338.46, df=1967, RNI=.91, TLI=.90, RMSEA=.045). Results indicated that non-attendance was significantly predicted by self-handicapping (=.10) and future plans (=-.09). Specifically, consistent with hypotheses, Time 1 self-handicappers were less likely to be in attendance at Time 2. Not surprisingly, students reporting less inclination at Time 1 to engage in their studies in the future were also less likely to be in attendance at Time 2.

12.3 Discussion One strength of this need achievement framework is that interpretations of the present studys process models are further clarified. First, that reflectivity positively predicts self-regulation and persistence is proposed to be because the reflective student is akin to the overstriver who exerts effort in a bid to strive for success and to avoid failure. Secondly, that defensive expectations negatively predicted self-regulation and persistence is consistent with the finding that defensively expectant students are high on the failure avoidance dimension to a greater extent than they are success oriented. Importantly, however, the fact that defensive expectations were not associated with impaired performance is not inconsistent with this framework: Because students holding such expectations are neither low nor high in success orientation, there is no systematic relationship between these expectations and success in the form of academic grades. Thirdly, that self-handicapping was the strongest negative predictor of self-regulation, persistence, academic grades, and non-attendance is consistent with the quadripolar model in that not only are self-handicappers diametrically opposed to success orientation but may even be failure accepting.

The multifaceted framework is able to further account for findings obtained in previous defensive pessimism research. First, Norem and Cantor (1986b) have found that encouraging defensive pessimists leads to poorer performance relative to defensive pessimists who are not encouraged. As contended earlier, each quadrant
differentiate between withdrawers and persisting non-respondents and so results must be considered with this limitation in mind.

305

marks students achievement styles - styles with which they are comfortable for a variety of adaptive or not so adaptive reasons. Further, when this style is disrupted in a way that places students in a quadrant that conflicts with their preferred style, their typical achievement pattern is disrupted (Covington, 1992). Thus, encouragement for the defensive pessimists in Norem and Cantors study was essentially an optimistic orientation, disrupting their typical achievement pattern. Secondly, Norem and Illingworth (1993) found that when defensive pessimists reflective processes were interrupted, their performance suffered. On the assumption that Norem and Illingworths defensive pessimists were somewhat success-oriented (given that they were in the top third of a self-concept-like screener), this to some extent locates them in Quadrant B (overstrivers). Given that Quadrant B represents an achievement orientation that also involves reflectivity, disrupting this reflectivity may disrupt their typical achievement style.

Another issue in relation to the two dimensional need achievement framework is that self-handicappers seem to become increasingly failure accepting over time. There is, however, one qualification made in relation to self-handicapping and failure acceptance: Whilst these students were abandoning engagement in their studies and investing little effort in the face of challenge, they were nonetheless motivated to strategically manoeuvre so as to protect their self-worth in achievement situations (see also Covington, 1992) and to cast their poor performance in the best attributional light. Unfortunately, however, the temporary relief afforded by these failure-avoiding tactics is illusory, since their repeated use will finally destroy the will to learn. This self-defeating process involves a shift from being failure avoiding to becoming failure accepting (Covington, 1984b, p. 12). In relation to this, the greatest cost of selfhandicapping seems to be withdrawal from university altogether (discussed below).

A number of the findings in the present study in relation to reflectivity, defensive expectations, and self-handicapping parallel Covington and Omelichs (1991; see also Covington, 1997) analysis of need achievement. They discussed how overstrivers were confident, experienced anxiety, displayed good study skills and spent much time studying. This parallels the reflective students who were higher in self-concept, anxiety, self-regulation, and persistence. Failure-avoiding students in Covington and 306

Omelichs analysis were found to be self-doubting, have poor study skills, and experience high anxiety. Similarly, in the present study, students high in defensive expectations were high in uncertain personal control and anxiety and were poor selfregulators. Finally, failure-accepting students in Covington and Omelichs analysis were low in anxiety, had poor study skills, and were self-doubting. Self-handicappers in the present study were lower in anxiety (relative to students high in reflectivity and defensive expectations), poor self-regulators and persisters, and uncertain in their personal control.

As hypothesised, self-handicapping positively predicted non-attendance at Time 2 and this is very much consistent with the failure-accepting perspective on selfhandicapping outlined above. Arguably, of the academic outcomes assessed in the present study, non-attendance is important in that it determines whether a student gains a university qualification. Whilst such outcomes as grades and persistence are also important, students scoring relatively lower on these dimensions at Time 2 can conceivably still pursue a higher education. Students who have withdrawn from their course, on the other hand, constitute the real casualties of higher education. Indeed, this might explain why self-handicapping did not negatively predict academic grades to a greater extent than it actually did: Grade measures were collected at Time 2 and self-handicappers were significantly less likely to be in attendance at Time 2.

The present study, then, is significant in the sense that self-handicapping behaviour at Time 1 was the only strategy which predicted non-attendance. The fact that neither defensive expectations nor reflectivity significantly predicted non-attendance at Time 2, confirms the status of self-handicapping as the most maladaptive of the strategies examined in the present study. Importantly, however, the study has identified a set of constructs that predict self-handicapping and which may represent ideal dimensions on which educational interventions designed to tackle students self-handicapping behaviour can be based (see the following chapter for further discussion). Indeed, this may constitute one significant means of enhancing student retention rates.

12.4 Summary of chapter

307

On both conceptual and empirical grounds, the quadripolar need achievement model has proved to be an effective framework under which the three strategies can be interrelated. For the most part, the data supported the predictions regarding the relative positioning of the three strategies: Reflective students resembled the overstriver high in success orientation and failure avoidance; defensively expectant students resembled the failure avoider high in the motive to avoid failure and not particularly inclined towards success. Consistent with predictions, self-handicapping was located closer to failure acceptance than defensive expectations, but so much so that it actually reflected failure acceptance outright. Indeed, the fact that self-handicapping was the only strategy that significantly predicted non-attendance at Time 2 supports this interpretation of the self-handicapper as failure accepting.

308

13. GENERAL DISCUSSION

The present study has been centrally concerned with exploring two process models and a longitudinal model across students first and second years at university. These models enabled an examination of relationships between a set of affective and motivational predictors, three strategies aimed at protecting self-worth, and the academic outcomes that follow from these strategies. A number of supplementary issues were addressed in the process of assessing these central models and through separate multidimensional scaling procedures. Additionally, two qualitative studies were conducted to illuminate key findings in the Time 1 process model and to explore factors associated with considerable (upward and downward) shifts in selfhandicapping and defensive expectations across Times 1 and 2.

13.1 Synthesis of self-handicapping, defensive expectations, and reflectivity from a self-protection perspective Taken as a whole, Time 1 and Time 2 data offer important insights into the factors that underpin self-protection, the precise ways in which this self-protection manifests itself, as well as the diverse ways in which self-protective strategies predict academic outcomes. Whilst in many respects the self-handicappers may have increased the likelihood of failure, they have managed to change the perceived cause of that failure: It has not occurred because of low ability but rather, it is because of low effort or due to factors seemingly beyond their control. Consistent with a self-worth motivation rationale, because low effort or external causes do not reflect poorly on ones ability, they do not pose a threat to ones self-worth. Because failure following the expenditure of effort can be seen to reflect low ability (Covington & Omelich, 1979b), it is preferable to risk failure due to low effort than to maximise ones chances of success by trying hard.

The qualitative data identified other ways self-handicappers engaged in selfprotection. For example, one student reported that by withdrawing effort before an evaluative threat, she was able to minimise its importance and in a sense, lower the ego-relevant stakes. Indeed, individuals have been found to be more inclined to selfhandicap in tasks characterised as important (Rhodewalt et al., 1984). Further, if 309

individuals do not perform as well as expected, the fact that they have devalued the event a priori implies a lack of personal investment and so their esteem is not on the line to be judged against that failure. In a sense, then, it reduces the involvement of ones esteem in the task thereby reducing the potential threat to this esteem.

The data also identified the self-protective dimensions of defensive expectations. First, setting defensive expectations is a way of avoiding personal disappointment. In interviews, students high in defensive expectations seemed to be fearful of failure and to minimise the disappointment such failure would cause, preferred to set lower expectations. Their self-protection was also evident in their aversion to optimism. Optimism, according to some respondents, simply meant that they had further to fall when they failed. These findings are consistent with Baumgardner & Brownlee (1987) who propose that one way to deal with the threat of failure is to strategically set lower standards that are easier to attain and therefore somewhat safer. Indeed, this is selfprotective in the sense that expected negative outcomes are not so hard-hitting as unexpected negative outcomes (Feather, 1969).

Self-handicapping and defensive expectations also seemed to serve protective purposes for students public image. In the present study it was proposed that individuals who engage in self-handicapping and defensive expectations not only do so to protect their private self-esteem, but also do so to protect their competence image. Indeed, a number of studies have identified the public dimensions of selfhandicapping (Shepperd & Arkin, 1989a; Strube, 1986) and defensive pessimism (Polak & Prokop, 1989; Showers & Ruben, 1990). The present studys operationalisation was unique in that it assessed separately the active and public/selfpresented dimensions of self-handicapping and defensive expectations and contributed to the area in the sense that these two dimensions were sufficiently indistinct that they could be subsumed under a higher order factor (see Section 8.3 for full discussion).

The Time 1 and Time 2 data also provided a new perspective on reflectivity. It is argued that reflectivity is a means of gaining and enhancing control. By thinking through a variety of possible outcomes, individuals prepare themselves for these outcomes and in this sense reflectivity can be seen as somewhat self-protective. But 310

this self-protection is not maladaptive in the way that self-handicapping and defensive expectations seemed to be. Rather, by contemplating a variety of possible outcomes, individuals respond with proactive academic behaviours that primarily took the form of self-regulation. In fact, reflectivity may represent a more adaptive form of selfmonitoring (as opposed to a maladaptive ruminative private self-focus - see Ingram & Smith, 1984; Musson & Alloy, 1988; Pyszczynski & Greenberg, 1987) which better places individuals to monitor their progress and take positive steps to enhance subsequent performance. This self-regulatory dimension of reflectivity may be a means by which individuals successfully respond to the possibility of failure whilst taking steps to maximise the possibility of success.

13.2 Assessing the strategies in the context of existing theory The quantitative and qualitative findings also hold a number of implications for theoretical contentions related to defensive pessimism and self-handicapping. In many cases the present findings are consistent with these theoretical contentions. In notable respects, however, the data are not so consistent with prior theorising.

13.2.1 Defensive expectations and reflectivity In the academic context, the theory of defensive pessimism holds that in anticipation of upcoming achievement situations, students feel anxious and out of control. In response to this, they set unrealistically low expectations which is a strategy aimed at cushioning the blow of potential failure (Norem & Cantor, 1986b). The theory also holds that in conjunction with defensive expectations, defensive pessimists think through a variety of possible outcomes (Norem & Illingworth, 1993). As a consequence of these defensive expectations and reflectivity, students performance is often unimpaired. In a sense, then, defensive pessimists harness their anxiety to adaptive ends such that their expectations do not become self-fulfilling prophecies (Norem & Cantor, 1986a, 1986b).

The fact that anxiety underpinned defensive expectations and reflectivity in the Time 2 quantitative study is consistent with prior theory. Moreover, consistent with the theory which predicts that students high in defensive expectations harness (Norem & Cantor, 1986a, 1986b) their anxiety to adaptive ends, the Time 1 and 2 qualitative 311

data illuminated the ways in which students high in defensive expectations used their anxiety to motivate them into action. Also consistent with theoretical contentions are quantitative and qualitative data which show that defensively expectant students feel outcomes are beyond their control: Uncertain personal control predicted both defensive expectations and reflectivity while an external attributional orientation predicted defensive expectations.

In terms of the consequences, the data are in some respects supportive of the theory in that defensive expectations did not lead to impaired performance. The qualitative data also showed, consistent with theory, that defensive expectations were a means of cushioning the blow of potential failure: Interviewees high in defensive expectations were careful not to be optimistic and strategically regulated their expectations (publicly and privately) to avoid the (public and private) disappointment that would result if they performed poorly. The hypothesised strategic nature of defensive expectations (see Cantor & Norem, 1989; Norem & Cantor, 1986a, 1986b) was also confirmed in interviewees reports of their strategic attempts to regulate their own and others expectations for upcoming performance so as to protect their self-esteem in the event of failure.

A number of other findings, however, were not so consistent with the theory of defensive pessimism. First, defensive expectations were not confined to high performing students. The present studys operationalisation of defensive expectations is apparently unique in its extension of defensive expectations encompassing students of all levels of academic performance. By rewording defensive expectation items, this study extends the potential pool of students high in defensive expectations to those at any level of self-concept. To justify treating defensive expectations in this way, the original operationalisation of defensive pessimism was tested by examining the interaction between defensive expectations and a self-concept item consistent with the self-concept-like screener used in previous research. These analyses yielded no significant interaction effects on self-regulation, persistence, and grades, and consistent with present contentions, indicated that setting defensive expectations is a strategy that operates independently of students previous performance and level of self-concept. Moreover, when the process model was examined separately for students 312

scoring in the top third of the self-concept item, defensive expectations were still negatively associated with self-regulation and persistence. Further research is now needed to explore the reasons why defensive expectations are held by students at various levels of self-concept. For example, it may be that students low in self-concept are more inclined to self-protect to defend themselves from a consistently low selfconcept while students high in self-concept hold defensive expectations because failure is a threat to their high self-concept. This, however, does not quite fit the results derived from the quadripolar model and suggests that the issue may be more complex than simply casting high and low self-concept in terms of a common need to self-protect underpinned by different motivations.

Second, defensive expectations and reflectivity did not interact in ways that are indicated by previous theory (Norem & Illingworth, 1993). Norem and Illingworth have implied that it is individuals who are high in both defensive expectations and reflectivity who are defensively pessimistic. But the finding that no effects on academic behaviours emerged as a function of their interaction suggests that defensive pessimism as previously defined (high defensive expectations/high reflectivity Norem & Illingworth) does not influence academic behaviours in the forms of selfregulation and persistence. Rather, academic behaviours are primarily influenced by the separate main effects of defensive expectations and reflectivity. To date, little research has examined reflectivity as a separate construct and the distinctiveness of findings in relation to it suggests that it be considered on its own merits in future research.

Norem (personal communication, March 26, 1997, March 7, 1998) has also implied that defensive expectations may predict reflectivity such that by setting defensive expectations and acknowledging the possibility of failure, students are sufficiently relaxed to engage in reflectivity which leads to adaptive educational outcomes. This possibility was tested and the data at Time 1 indicated that defensive expectations do not significantly predict reflectivity. At Time 2, while defensive expectations did significantly predict reflectivity, the indirect positive effect of defensive expectations on self-regulation via reflectivity was much weaker than the direct negative effect of defensive expectations on self-regulation. Thus, in terms of (a) the non-significant 313

interactions between defensive expectations and reflectivity, (b) the fact that defensive expectations did not predict reflectivity at Time 1, and (c) the finding that defensive expectations and reflectivity yielded markedly different outcomes as main effects, it seems that the two are substantially independent strategies.

The distinction between defensive expectations and reflectivity is further clarified in the framework that represented the three strategies in two-dimensional space, consistent with the quadripolar need achievement model proposed by Covington and Omelich (1991; see also Covington, 1992, 1997). As noted earlier, this framework showed that whilst both defensive expectations and reflectivity are high in failure avoidance (hence self-protective), reflectivity is also high in success orientation. Because reflective students are high in success orientation they are more inclined to self-regulate and persist in the face of challenge. Importantly, while students holding defensive expectations are not high in success orientation, they are not low on this dimension and so the finding that defensive expectations did not impair performance is quite consistent with this. This multidimensional framework, then, lends clarity to the issue of how defensive expectations and reflectivity are both self-protective (high in failure avoidance) yet lead to quite different outcomes (due to their differing status in success orientation).

It is argued here, however, that despite these strategies being independent constructs, they are nonetheless both a means of preparing the student for upcoming achievement situations. Defensive expectations seem to be aimed at preparing the student for failure and cushioning the blow of this anticipated failure (Norem & Cantor, 1986b). Reflectivity, whilst also a preparatory strategy, is aimed at preparing the self for all possible outcomes rather than just failure. Moreover, both can be separated from selfhandicapping because while self-handicapping is unambiguously inimical to academic success, reflectivity is adaptive in terms of self-regulation and persistence while defensive expectations do not lead to impaired performance. Further consistencies were found at Time 2 where defensive expectations and reflectivity were much more highly correlated than at Time 1.

314

An issue that requires further clarification concerns the primary purpose of defensive expectations. Whilst the present study adopts a self-protection perspective on defensive expectations, thus far there has been proposed a variety of purposes for such self-protection. These include setting lower and safer standards against which to be judged, avoiding disappointment in the event of failure, engaging in some level of self-verification, and steeling the self for failure and cushioning its blow. There is a need for further research that explores the salience of each of these purposes and the conditions under which one purpose may be more pertinent than another.

13.2.2 Self-handicapping The central tenet of self-handicapping theory is that in a bid to protect or enhance the self, students strategically manoeuvre by placing obstacles in the path to success. This is an a priori strategy that changes the attributional profile of the task such that students who fail are not held responsible for that failure and so are protected. Moreover, success in the context of these impediments is seen as more impressive and thus enhances ones competence image. In synthesising the theoretical bases of selfhandicapping, Berglas (1987) proposed a self-handicapping model in which (1) an individual experiences performance anxiety or evaluation apprehension, (2) that individual chooses an obstacle that is known to impede success, (3) the individual then assumes that he or she has nothing to lose, (4) the performance anxiety or evaluation apprehension is alleviated, and (5) the individual then performs successfully.

The quantitative data supported the self-protection perspective of self-handicapping in the sense that a variety of factors hypothesised to underpin a self-worth motivation predicted self-handicapping. The qualitative data provided a more in-depth perspective on students motivation for employing this strategy. Interviewees identified that they were unsure and anxious about their ability to perform in upcoming achievement scenarios and proceeded to outline the variety of ways that they defensively manoeuvred which ranged from visiting ones grandmother at exam time to exhibiting a chronic inability to keep a tidy desk. In all cases, respondents indicated that they were unable to study effectively as a direct consequence of these impediments. 315

Consistent with Berglass (1987) model, in response to anxiety or evaluation apprehension, respondents did choose impediments to successful performance. But the particularly weak strength of the predictive path between anxiety and selfhandicapping was unexpected (although the correlation was a little higher). This may be because protective self-handicapping is more consistent with the failure accepter (who is low in anxiety - Covington, 1997) rather than the failure avoider who is typically more anxious. If, as is presently suggested, protective self-handicapping is more akin to failure acceptance than failure avoidance, theorising about selfhandicapping must duly accommodate this.

The present data were also consistent with early contentions that self-handicapping is strategic (Jones & Berglas, 1978). The qualitative data reflected this in quite a marked fashion. Students reported that in anticipation of upcoming essays and exams they would engage in a variety of activities that could serve as impediments to success but which also reflected strategic a priori attributional manoeuvring that protected them in the event of failure. They also seemed to be savvy to other attributional principles. For example, whilst they were aware that they were the cause of the behaviour, they seemed to be aware of the conditions under which they would not be held responsible for poor performance resulting form their behaviours (see Berglas, 1987). Also, the data indicated that self-handicapping was strategic in the sense that it was very much self-presented and public. The Time 1 interview data were particularly supportive of this in showing that students were inclined to make others aware of the impediments to success and in some cases actually exaggerated these impediments.

The original conceptualisation (Berglas & Jones, 1978; Jones & Berglas, 1987) and Berglass (1987) model also posit that self-handicapping reduces anxiety and that this gives rise to enhanced performance (see also Deppe & Harackiewicz, 1996). In contrast, the Time 2 process model data that incorporated anxiety suggested that selfhandicapping led to reduced self-regulation and persistence. In this sense, then, selfhandicappers responses to anxiety were not consistent with original contentions. Present findings may be more consistent with the notion of the self-handicapper as a failure accepter than as a failure avoider - a distinction which has apparantly not 316

received empirical or theoretical attention to date. This issue of failure acceptance may be particularly relevant to the self-protective form of self-handicapping (. . . so that I have an excuse if I dont do as well as I hoped) explicitly invoked in the present study (see also Midgley et al., 1996). Further research might focus on self-enhancing self-handicapping with a view to ascertaining where in the quadripolar model it is located relative to the other constructs incorporated in this study. Whereas most theory deals with enhancing and protective self-handicapping jointly, need achievementrelated data may suggest that they be considered as two relatively independent constructs.

In sum, then, the quantitative and qualitative data in relation to defensive expectations, reflectivity, and self-handicapping are in many respects consistent with original and current theorising. They do, however, diverge from theory in important ways. In the case of defensive expectations and reflectivity, the data pose some rather complex questions for existing theory. The data in relation to self-handicapping are more consistent with existing theory, but do challenge aspects of theory in terms of the failure-accepting versus failure-avoiding dimensions of self-handicapping and also in terms of its consequences.

13.3 Implications for education The present findings hold a number of implications for educational researchers and practitioners. The predictors of self-handicapping, defensive expectations, and reflectivity point to areas that can be the focus of educational interventions designed to target defensive manoeuvring. In addition to the specific predictors of the strategies, other analyses conducted in the present study not only clarified some of the theoretical issues relevant to constructs in the present study but also provided data relevant to more practical considerations.

13.3.1 The role of the predictors in addressing students defensive manoeuvring 13.3.1.1 External attributional orientation The quantitative and qualitative data extended current understanding as to the extent to which an external attributional orientation underpinned the strategies central to the proposed processes. These data show that an external attributional orientation is not 317

particularly adaptive in that it is a primary predictor of self-handicapping, and to a lesser extent, defensive expectations. It is important, then, for educators to encourage students to develop more of an internal attributional profile and thereby enhance their perceptions of control. Consistent with the clinical tradition, Higgins and Berglas (1990) suggest that self-handicappers need to develop a new attributional logic by learning techniques such as cognitive restructuring (see also Meichenbaum, 1977). In an educational context, a number of attribution retraining studies have shown that encouraging participants to attribute outcomes more to effort (seen as unstable and controllable) than ability (seen as stable and uncontrollable) promotes individuals control over given outcomes. Schunk (1981), for example, has shown that attributing past achievement to effort tends to promote persistence and performance (see also Dweck, 1975). Andrews and Debus (1978) found that effort and ability attributions following failure were positively and negatively associated with persistence respectively. Moreover, Andrews and Debus found that encouraging those who had not previously attributed failure to effort to do so led to enhanced persistence.

What these attribution retraining studies show is that it is imperative to distinguish between controllable and uncontrollable internal dimensions. For example, the present findings indicate that ability attributions for failure (relatively uncontrollable) contributed to defensive expectations and self-handicapping whereas it has been shown that encouraging students to attribute outcomes to effort and strategy (controllable) promotes perceived competence and adaptive approaches to learning (Andrews & Debus, 1978; Craven, Marsh, & Debus, 1991; Dweck, 1975; Schunk, 1981). Thus, when encouraging students to attribute outcomes to internal factors, it is important that educators are cognisant of the key differences between effort- and ability-related causal ascriptions.

13.3.1.2 Uncertain personal control The quantitative and qualitative findings also underscored the centrality of uncertain personal control in predicting defensive expectations and self-handicapping. Moreover, the qualitative data indicated that inconsistent assessment in the educational context contributes to students perception of uncertain control over their ability to avoid future failure. Indeed, this supports research demonstrating the link 318

between non-contingent feedback and self-handicapping (Berglas, 1987; Berglas & Jones, 1978; Higgins & Harris, 1988; Perry et al., 1986; Perry & Dickens, 1984; Rhodewalt & Davison, 1986; Riggs, 1992). The principal outcome arising from noncontingent feedback is low perceived control which has also been associated with pessimism and inversely associated with optimism (Pyszczynski, 1982; Schunk, 1983; Zackay, 1984). It is therefore proposed that promoting individuals perceived control over given outcomes is an important step in addressing self-defeating behaviour. One way of promoting perceived control is to administer reinforcement and feedback that is directly contingent on students performance (Craven et al., 1991; Hattie, 1992; Thompson, 1994). Additionally, as outlined above, addressing individuals causal ascriptions for outcomes is another way of enhancing their perceived control over future outcomes.

Also underpinning this issue of control is the stability of individuals self-concept. It is proposed that individuals whose self-concept is unstable are in doubt as to their ability to meet upcoming challenge. In the face of such uncertainty, students may choose to play it safe and engage in strategies that offer self-protection. The present data do not enable concrete recommendations about promoting a stable self-concept. It is therefore important to determine the factors that give rise to an unstable selfconcept. In the educational context, these may include inconsistent assessment (Kernis & Waschull, 1995) and differential attention and reinforcement to students on the part of the teacher (Beckman, 1970; Felsenthal, 1970). It may also be that a fluctuating self-concept results from a view of self that is partly based on how others respond to us. For example, the symbolic interactionist position holds that our self-concept develops as a consequence of how others respond to us and what we perceive others think of us (Kinch, 1963). When self-concept is based on this and because others opinions are unlikely to be entirely consistent, we are particularly vulnerable to fluctuations in self-concept. Thus an over-reliance on others evaluations of them may render some students vulnerable to an unstable self-esteem (see Rosenberg, 1986). Moreover, Kernis et al. (1993) have found that individuals whose self-worth is based primarily on competence are most susceptible to fluctuations in self-concept. Thus, it might be important to address students concerns regarding others evaluations of

319

them and to de-emphasise the role of competence in students private assessments of their self-worth.

13.3.1.3 Performance orientation A number of dimensions underpinned performance orientation. These included egoorientation, esteem-relevant competence valuation, public self-consciousness, and avoidance-oriented performance. Each of these is discussed with particular focus on how they manifest themselves in the educational context from a self-worth protection perspective in addition to the means by which educators are able to address these factors in altering the self-worth implications of their current educational practice.

Ego-orientation, as one pivotal component underpinning a performance orientation, contributed to the tendency to self-handicap and engage in defensive expectations. Ego-oriented individuals feel most successful when they outperform others and tend to be dominated by competitive concerns. Competition perpetuates the notion that success is derived from ability rather than from effort (Covington & Omelich, 1984b) and so failure in this context reflects directly on ability and hence ones self-worth. Thus competition for those not guaranteed success can represent a climate in which self-protection is a primary objective. Indeed, in the interviews, students high in selfhandicapping and defensive expectations reported that they saw university as a competitive environment and that they felt particularly successful when outperforming others in this climate. Mastery, on the other hand, has as its focus success by way of effort and thus failure in this context is not seen as a threat to ones ability. Rather, it is seen as diagnostic effort-related feedback that one can use to enhance performance in the future (Covington, 1997).

Indeed, Covington (1989, 1997) suggests that classroom rewards be based on the extent to which students become competent rather than distributed on the basis of students ability to be competitive. According to Covington (1992), the classroom reward system should be focused on incentives that draw rather than those that drive. Along these lines, goals that minimise the competitive climate include self-mastery, helping others, expressing creativity, and satisfying curiosity (Covington, 1989, 1997). Other practical strategies include reworking the nature of assessment itself. Covington 320

(1989) has proposed alternative grading systems in which students work at assigned tasks until they have reached a predetermined grade of choice. Covington and Omelich (1984c) also found that mastery-based learning structures such as re-testing were associated with higher performance and motivation. Notwithstanding this, it must be recognised that with re-testing comes the possibility of repeated failure which has been shown to be detrimental to feelings of self-worth (Covington & Omelich, 1981). Another complicating issue is the role of effort in students lives and in the context of what teachers reward. Covington and Omelich (1979b) found that while low effort in the context of failure elicited lower judgements of incompetency and shame from the perspective of students, from teachers perspectives this combination of failure and low effort elicited most punishment. Thus the meaning of effort must be redefined to students such that it does not so directly reflect on ones ability and is seen as a positive value in its own right.

Further to this issue of competitive versus cooperative learning environments, Johnson and Johnson (1989) and Qin, Johnson, and Johnson (1995) have proposed that competitive environments promote comparisons of ones ability with that of others whilst cooperative environments are very task-oriented and focused on mastery through collaboration with ones peers. On the strength of their meta-analyses they concluded that cooperative learning tended to be more adaptive in terms of performance and productivity (see also Moriarty, Douglas, Punch, & Hattie, 1995). It seems then, that it may be more desirable to gear the learning environment towards mastery and cooperation rather than competition (see also Garcia et al., 1995; Thompson, 1994). Indeed, competitive and ego-oriented environments have been found to be maladaptive in a number of other educational respects (see, for example, Elliot & Dweck, 1988; Graham & Golan, 1991; Meece & Holt, 1993). In terms of self-handicapping these contentions are very much supported in recent work by Urdan et al. (1998) who found that an emphasis on relative ability goals (very much akin to a competitive climate) in the classroom as well as teachers use of instructional practices emphasising relative ability were positively associated with selfhandicapping by students (see also Garcia et al., 1995, in relation to competitive university classrooms and both defensive pessimism and self-handicapping). Midgley and Urdan (1995) found much the same: Self-handicapping was positively correlated 321

with school performance goals and negatively correlated with school task goals. Whilst the quantitative data in the present study did not ascertain students perceptions of their learning environment, some interviewees indicated that they saw university as a competitive place and to this extent, implications regarding competition and competitive environments are relevant to the present study.

Notwithstanding these negative associations between performance-related goals and educational outcomes, it must be noted that some level of competition is preferred by students with more adaptive (deep) approaches to learning (Hattie & Watkins, 1988) and performance-related goals can be related to more adaptive educational outcomes under particular circumstances (see Elliot & Church, 1997; Harackiewicz et al., 1997; Meece et al., 1988; Moriarty et al., 1995; Roeser et al., 1996; Skaalvik, 1997a, 1997b). Indeed, Harackiewicz et al. (1998) argue that there is a variety of conditions under which performance goals can be adaptive. For example, Elliot and Church (1997) suggest that the effects of performance-related goals can depend on the nature of the educational challenge before students. In situations in which the threat of failure is salient, performance-related goals may take on a more avoidance profile and be negatively associated with desirable educational outcomes. On the other hand, in achievement situations that present more a challenge than a threat, performancerelated goals may be more adaptive. What is required then, is to recognise the complexity of performance-related goals and the conditions under which these goals are not deleterious to important educational constructs.

Public self-consciousness also contributed to students tendency to self-handicap and engage in defensive expectations. Consistent with previous contentions (Federoff & Harvey, 1976; Shepperd & Arkin, 1989a; Vallacher & Solodky, 1979), it is proposed that individuals who are concerned with how they are viewed and evaluated by others are particularly motivated by the need to self-protect. Publicly self-conscious individuals are very much concerned with their image (Buss, 1980; Cheek & Briggs, 1982) and it is this which renders them vulnerable to self-worth threats. In the context of self-worth vulnerability, it may be preferable that students are encouraged to be less concerned with how they are viewed and evaluated by others and adhere to more private and personal standards that do not render self-worth vulnerable. Higgins and 322

Berglas (1990) imply as much in their discussion of the clinical implications of selfhandicapping. According to them, one of the fundamental goals of therapy with selfhandicappers is to enable them to achieve a sense of self that is independent of social influence and capable of determining acceptable levels of performance on the basis of internally held standards (p. 227). Indeed, the qualitative data indicated that students low in self-handicapping and defensive expectations were more concerned with aspiring to personal goals and standards than with adhering to publicly prescribed standards. Moreover, given that competition is prevalent at university (Harackiewicz et al., 1998), and that competition promotes a concern with how one fares in relation to others, it is speculated that these factors also give rise to a concern with how one is evaluated by others because it is the demonstration of ability and competence which is paramount. A more cooperative, rather than competitive focus, may diminish the extent to which students are publicly self-conscious.

Also contributing to the tendency to engage in self-handicapping and defensive expectations is an avoidance-oriented performance based on a motivation to work and perform to avoid failure and its negative consequences. Fear of failure along these lines has been the topic of study for many years. A motivation to perform based on a fear of failure and its negative consequences is pivotal to a self-worth motivation perspective because it is primarily poor performance that places students ability image in question and which has subsequent implications for self-worth (Beery, 1975; Covington, 1984a, 1992, 1997; Covington & Beery, 1976). One important means by which this process can be short-circuited is to encourage students to see failure more as diagnostic effort-related feedback than performance feedback. In this way the threat to the self-worth that failure poses is minimised.

The extent to which students feelings about themselves were dependent on experiencing and demonstrating competence also contributed significantly to their tendency to engage in self-handicapping and defensive expectations. This is centrally tied to self-worth motivation theory because competence is the entity that influences many students feelings of self-worth. Indeed, it has been shown that children learn early in life that competence and self-worth are closely connected (Covington, 1992; Thompson, 1994). It is important, then, to minimise the link between competence and 323

self-worth. Whilst it is probably idealistic to propose that students be encouraged to see that their self-worth is unconditional, a more realistic proposition is that if educators seek to tie students self-worth to anything, it is best that it be tied to effort and attainment of personal goals. In this way, students gain feelings of self-worth based on exertion of effort and self-improvement. Importantly, however, it is recognised that expenditure of high effort can be seen to imply low ability (Covington & Omelich, 1979b). Therefore, if effort is to be lauded as a goal in the educational setting, it is crucial that educators also make it clear that a given level of effort does not necessarily reflect on ones ability.

13.3.2 A need achievement perspective Findings derived from the multidimensional mapping of the three strategies confirmed an hypothesised conceptual perspective on the strategies that was based on the quadripolar need achievement model (Covington & Omelich, 1991; Covington, 1992, 1997). These data clarified the nature of reflectivity, defensive expectations, and selfhandicapping, particularly in relation to how they can be represented in terms of the individual and joint operation of success orientation and failure avoidance. First of all, whilst reflectivity in the process models led to adaptive outcomes, results also indicated that similar factors underpinned it that also underpinned self-handicapping and defensive expectations. This was suggestive of a self-protective dimension underpinning reflectivity but one which did not manifest itself in deleterious outcomes. The multidimensional framework supported and clarified this contention by showing that reflectivity reflects not only a high success orientation but also some level of failure avoidance. It was further proposed that students high in reflectivity were very much akin to the overstriver proposed by Covington and Omelich (1991; see also Covington, 1992, 1997) in the sense that they were high in both success orientation and failure avoidance. From an educational perspective this is significant because such students are nonetheless vulnerable to failure, no matter how isolated (Covington, 1992). It is suggested here that when the need for failure avoidance becomes more salient than the drive to succeed, the strategic balance is shifted such that these students are then more inclined to hold defensive expectations. Thus, while reflectivity is clearly the most adaptive of the three strategies, students high on this dimension are vulnerable given their concerns with avoiding failure. 324

The multidimensional framework also provided a most important insight into the need achievement profile of the self-handicapper. Data suggested that self-handicapping bordered on failure acceptance in the sense that self-handicappers were low on both success orientation and failure avoidance. Indeed, the interpretation of selfhandicapping as reflecting failure acceptance clarifies its status as the only strategy to predict non-attendance at Time 2. This is parallel to the profile of the helpless student and the earlier discussion dealing with attribution retraining and enhancing perceived control (see Andrews & Debus, 1978; Craven et al., 1991; Dweck, 1975; Schunk, 1981) may be more critical to the self-handicapper than to any other type of student. Importantly, however, whilst reflecting helplessness, the self-handicappers are nonetheless distinct from the helpless students in that they are motivated to protect their self-worth. Such a perspective on self-handicapping is consistent with that presented by Midgley and Urdan (1995) who argue that self-handicappers still care enough to want to appear able to others and that this is a hopeful sign (p. 407) because educators may be able to harness this motivation to adaptive ends.

This need achievement model of reflectivity, defensive expectations, and selfhandicapping demonstrates how students using a variety of strategies can be interlinked on common dimensions using an integrative framework. One strength of such a framework is that educational implications can be advanced using a common conceptualisation that enables specific, yet multifaceted, recommendations regarding pedagogical practice. As Covington and Roberts (1994) speculate, how can the threat to learning be diffused, on the one hand, so that overstrivers and failure avoiders feel less vulnerable; yet, on the other hand, how can learning be infused with powerful incentives that appeal to all, sufficient to transform both disengagement (failureaccepters) and defensive vigilance into personal commitment? (p. 177). To do this, Covington and Roberts argue, the very nature and bases of learning must be changed so that motives become goals and draw rather than drive the student (see also Covington, 1992). Such a program would encourage students to gain knowledge for masterys sake rather than for the sake of performance, encourage students to serve the interests of the group, and to give expression to their creativity and curiosity. Changing the reward system has also been proposed as a way in which the purpose of 325

learning can be altered. Covington and Roberts suggest that reward should be based on students meeting personal standards rather than outperforming others. In a sense, then, the student is encouraged to become success-oriented rather than failure avoiding or failure accepting. When students are success-oriented they are then in a stronger position to learn and as Covington and Roberts note, have even been shown to perform well in competitive scenarios (see also Elliot & Harackiewicz, 1994; Epstein & Harackiewicz, 1992; Harackiewicz & Elliot, 1993; Harackiewicz & Manderlink, 1984).

The precise means by which these pedagogical recommendations are put into practice are somewhat complex. For example, rewards should not be too plentiful. As Covington and Roberts query, if there are no clear losers in the learning game, will it be worth playing? (p. 179). One possible means of circumventing this problem is that students performance standards should increase along with their aspirations and so there always exists a challenge to be surmounted (Covington & Beery, 1976). For example, this could involve encouraging students to view their education in terms of personal bests much along the lines of sportspeople at even the most elite levels. Another example of the complexity of the pedagogical implications relates to the students experiences of cooperative education. In the Time 2 qualitative study some students indicated that they were frustrated with group work, suggesting that it evoked feelings of personal helplessness. One way of overcoming this is to instigate work contracts which ensure that students contribute in an equitable and dedicated fashion. In essence, then, with any set of guidelines aimed at promoting a success-oriented environment, educational practitioners must recognise the complex and multifaceted nature of the variety of pedagogical applications involved.

13.3.3 Summary of educational implications In summary, then, the present findings offer a number of directions for educational practitioners. It is proposed that students be encouraged to attribute outcomes to internal factors but that this be primarily effort- rather than ability-based. It is suggested that feedback, reinforcement, and assessment be as directly contingent on performance as possible thereby enhancing students perceptions of control. In terms of an unstable self-concept, it is proposed that students be discouraged from basing 326

their self-worth on the demonstration of competence and others evaluations of them. Educators should also minimise the competitive focus of the learning environment (or at least recognise the conditions under which it is deleterious to academic outcomes and self-worth) and promote task and mastery foci. In this way, effort, rather than ability, becomes the basis of success, and failure is robbed of its threat to ones ability image. It is also suggested that a cooperative, rather than competitive (or relativeability-based), learning environment may minimise self-protective concerns in the classroom. Encouraging students to adhere to more personal goals and standards is recommended and this is one way to minimise the public dimension of self-worth protection. Instilling in students the notion that failure or poor performance is diagnostic and a springboard for later success rather than evidence of low ability and low self-worth is also identified as an important task for educators. Nevertheless such an approach is probably only workable when students believe that they can surmount such failures and attain success, in which case, some self-efficacy promotion is also recommended.

13.4 Limitations of the study and directions for future research The present data not only provided a number of new perspectives on the many constructs involved in the investigation but also extended current understanding of these constructs. There were, however, limitations which in some respects qualify these findings and in other respects provide a clearer context in which to interpret the data. Moreover, each of the identified limitations has direct bearing on future research and this too is discussed.

13.4.1 Self-report data The first issue to consider is that the data used in the study were derived from selfreports. Whilst many of the constructs involved can only be assessed through selfreport, it is important for future research to examine the present process models using data derived from additional sources. Such data might incorporate ratings about students by others close to them. While this is one avenue for future research, it must be recognised that self-reports of self-handicapping have been validated against actual behaviour (Deppe & Harackiewicz, 1996; Rhodewalt et al., 1984; Rhodewalt & Fairfield, 1991; Strube, 1986). Perhaps more importantly, the present study validated 327

the self-handicapping and defensive expectation constructs through triangulation of data collected using qualitative and quantitative methods. Moreover, research has shown that self-reports about self-related constructs are usually in broad agreement with objective ratings made by others (e.g., see Marsh, 1990, for a review in relation to self-concept). Indeed, in a study of self-handicapping in the educational context, Urdan et al. (1998) raise the issue of self-report accuracy and suggest that because teachers reports of students behaviour were consistent with students reports, students self-reports were deemed to be accurate.

Related to this issue of self-report is the extent to which students are aware of the selfprotective strategies they use. In relation to self-handicapping, for example, consistent with Midgley and colleagues (Midgley & Urdan, 1995; Midgley et al., 1996; Urdan et al., 1998), the present study explicitly asked students about the extent to which they engaged in self-handicapping for self-protective purposes (e.g., I often fool around the night before an exam, so if I dont do as well as I had hoped, I can say that is the reason, italics added). At question is the extent to which students are aware of their use of such strategies, or, if they are aware of their tendency to engage in such behaviour, the extent to which they are aware of the reasons for it. Indeed, Nisbett and Wilson (1977) have argued that there are some aspects of the self, such as thought processes, that cannot be accessed. In the context of the present study, it may be that individuals cannot access their self-protective motivations and so the present data are limited in this respect. However, because self-reports of behaviour such as selfhandicapping have been validated against individuals behaviour, it is proposed that the present data are not as limited as Nisbett and Wilson would hold. Norem and Illingworth (1993) also point to the transparent nature of the items in the defensive expectations subscale as evidence that respondents are aware of what they are being asked. In their study, when participants were asked to guess which strategy group they were in, they were almost always correct (p.829).

Also related to these issues is the possibility that self-protective students may be unlikely to concede to the adoption of self-protection strategies. Asking students about defensive, self-defeating behaviour may evoke defensive self-reports. This would primarily manifest itself on the self-handicapping items which explicitly reflect the 328

most defensive strategy in the study. Indeed, mean scores on this dimension were the lowest amongst all the constructs. Thus the substantive issue is confounded with the measurement component of the study and this must be considered when interpreting present findings. In defence of the present trait measures of self-handicapping, Rhodewalt (1990) has suggested that respondents are willing to confess to selfhandicapping inclinations and yet think of each enactment of self-handicapping as a discrete event in which the audience is not knowledgeable of his or her selfpresentational modus operandi across other situations (p. 92). Additionally, Arkin and Oleson (1998) have proposed that in the pursuit of protecting their ability, individuals are willing to be stigmatised as self-handicappers and by implication would not be unwilling to concede to self-handicapping behaviour in a questionnaire.

329

13.4.2 Measures used in the study When interpreting present findings, it also important to recognise that only the selfprotective dimension of self-handicapping was assessed. Whilst this is consistent with a good deal of research into self-handicapping (Midgley & Urdan, 1995; Midgley et al., 1996; Urdan et al., 1998; Rhodewalt & Davison, 1986) and with the attributional principles that guide self-protection (Baumeister & Scher, 1988; Berglas, 1987; Kelley, 1972), the other dimension which is also of relevance to self-handicapping in an attributional sense is that of self-enhancement (Kelley, 1972). The selfenhancement perspective of self-handicapping holds that individuals select or construct impediments to success so that in the event of successful performance they are seen to be more competent than if they succeeded in the absence of such impediments. It would be informative to re-assess the present process models by juxtaposing protective self-handicapping with enhancing self-handicapping (e.g., I often fool around the night before an exam, so if I do well, I feel/look more successful than if Id studied). It would also be interesting to incorporate this enhancing dimension into the multidimensional framework to ascertain the extent to which this dimension is related to success orientation and failure avoidance. Indeed, of further interest would be an assessment of the private (feel more successful) and public/self-presented (look more successful) dimensions of enhancing selfhandicapping and the relative salience of each in relation to this enhancing orientation.

While the present study did not incorporate original measures into some analyses, conceptually-akin constructs were utilised which were considered to be feasible substitutes. For example, in analyses testing the interaction between defensive expectations and a self-concept item, the present analyses utilised an item drawn from the self-concept subscale rather than including in the questionnaire the typical selfconcept-like screener used in previous research. Also, rather than incorporate a measure of optimism in the multidimensional framework, consistent with that incorporated in Covington and Omelichs (1991) quadripolar need achievement model, the present study incorporated the self-concept item parcels into the multidimensional scaling procedures. Some constructs involved in the present analyses, then, are not strictly congruent with previous formulations and findings in relation to these constructs are to be interpreted with this in mind. Further research is 330

required to establish that present findings are consistent with findings using original measures. Indeed, incorporating optimism into the proposed process model would be a fruitful direction for future research.

The present study placed emphasis on the separation of reflectivity and defensive expectations and in this sense to some extent challenged the conceptual position held by Norem and colleagues. In support of the present position were the findings that reflectivity led to effects quite opposite to those of defensive expectations, there was no significant interaction between the two, and defensive expectations did not significantly predict reflectivity in a model that specifically tested this (or its direct negative effect on self-regulation was stronger than its indirect positive effect via reflectivity). Whilst the weight of present evidence suggests that the two are separate constructs, further research is required to explore exactly how the two work together under a defensive pessimism framework. Indeed, this is not the only reason further research into reflectivity is required. In the context of the history of defensive pessimism research, reflectivity is a much under-studied construct.

Also related to the measurement of defensive expectations is the fact that the present study used an adapted form of the Defensive Pessimism Questionnaire and operationalised it in a way that differed from previous work. Accordingly, while conclusions have been drawn and recommendations made for intervention and future research, these must be placed in the context of the adapted measurement and operationalisation of defensive expectations. Further research is required in which these two (parallel) dimensions of defensive pessimism are incorporated into the one study and analyses along present lines conducted.

While the present study adapted the defensive expectations measure, the original form of the reflectivity scale was used. This scale included three types of items: Items pertaining to thoughts about positive outcomes; items pertaining to thoughts about negative outcomes; and items pertaining to thoughts about both positive and negative items. It may be, then, that reflectivity is best represented as three factors rather than the one factor operationalised in the present study. This three factor model was not pursued in the present study primarily because there were not enough items to 331

adequately represent each factor and also because the study was not intended to test the multidimensionality of the factors contained within it. Further research is required to test the multidimensionality of reflectivity, and if multidimensional, the extent to which component factors are differentially associated with a variety of parallel constructs.

Some qualification of the multidimensional framework is also required. It will be recalled that this framework pivots around the interpretation of the axes as failure avoidance and success orientation. Because no direct measures of failure avoidance and success orientation were incorporated into the study, it can only be inferred that these axes represent the quadripolar need achievement dimensions proposed by Covington and Omelich (1991; see also Covington, 1992, 1997). Indeed, it may be that the primary utility of this component of analyses is its heuristic value. While it is suggested that because the strategies were positioned in the framework largely as predicted in part validates the present interpretation of these axes, further work is required that more directly establishes this.

Related to this issue is the extent to which the variety of affective and motivational predictors are indicants of self-worth motivation. The contention that self-worth motivation underpins these constructs could not be established empirically as it appears that there exists no direct measure of self-worth motivation. Rather, selfworth motivation in the present study is proposed to be more a joint operation of constructs than a single unidimensional measure. Nonetheless, the present study is limited in that no single measure of self-worth motivation was incorporated and so the self-worth motivation bases of the predictors were established more through conceptual rationale than direct empirical operationalisation. In a sense, however, representing self-worth motivation in this way may constitute a contribution to the area in that self-worth motivation is proposed to be multidimensional and subsumed by a variety of affective and motivational constructs.

332

13.4.3 The content domain and nature of the sample When interpreting present findings, it is important to recognise that they relate largely to the maths domain. It is hastily added, however, that the first phase of data analysis quite clearly established that there was very little differentiation between maths and verbal domains. Indeed, given that these two domains represent the diversity of academic demands presented to university students, these tests of content specificity were a rigorous means of assessing the generality of the constructs under focus in the study. For this reason, much of the study was couched in general terms rather than in terms of mathematics. Nevertheless, future research might test these constructs in the context of other academic domains to verify that present contentions regarding domain generality can be sustained. Moreover, as Self (1990) notes, self-handicapping is particularly relevant when an important self-concept is at stake. The extent to which maths is important or students value competence in this domain (see Elliot & Harackiewicz, 1994; Harackiewicz & Sansone, 1991) was not established in the present study and further research is required in relation to this. It may be that defensive manoeuvring in domains that are intrinsically interesting yields positive effects (e.g., see Deppe & Harackiewicz, 1996, in relation to self-handicapping) whereas in areas which evoke anxiety and are perhaps not so interesting (such as the maths domain in the present study), defensive manoeuvring may not be so adaptive.

The present findings are limited to Education students. Whilst these students constituted quite an ideal sample because they all study maths- and numbers-related subjects in addition to reading- and writing-related subjects, students from other faculties do not necessarily study both maths- and verbal-related subjects. What, then, is the pattern of self-protective processes in relation to Science, Arts, or Engineering students where maths and verbal skills are differentially salient for some and not for others? It must be noted, however, that whilst all students in the present sample were studying Education, they comprised three very different groups (enrolled in either Early Childhood, Primary, or Secondary education programs from three universities). Importantly, analyses assessing the relationship between program of study and all constructs in the model (at Time 1) yielded no significant or systematic effects nor did the factor structure vary from university to university. These findings, then, also reflect the generality of the present findings. 333

The fact that future research is needed to assess the proposed processes using students from a diversity of university faculties across all year levels, also applies on a broader level: Further research is needed to explore these processes using students in other educational environments. Presumably, participants in the present study were sufficiently capable and motivated academically to have graduated from high school to university. It would be interesting to explore the proposed process model using students at various levels of school. Whilst some previous work has assessed selfhandicapping in relation to upper elementary school students (Urdan et al., 1998) and secondary school students (Midgley & Urdan, 1995), self-handicapping research has primarily focused on university students as has research into defensive pessimism. Also, Rhodewalt (1990) has suggested that there is a need for research into selfhandicapping in particular and self-protection more generally that adopts a life-span approach (see also Garcia et al., 1995; Midgley et al., 1996 in relation to developmental approaches to self-handicapping). According to Rhodewalt, this concerns not only research into self-handicapping involving young children but also research into the strategic use of self-handicapping amongst the elderly.

The self-protection process is not confined to the educational domain and it is quite conceivable that the present process model could be quite applicable to social, occupational, and sporting domains (see Rhodewalt, 1990). This is not a new proposition. For example, self-handicapping has been studied in the sporting domain (e.g, see Leary, 1986; Rhodewalt et al., 1984) and defensive pessimism has been studied in the social domain (e.g., see Cantor & Norem, 1989; Norem & Cantor, 1990b; Showers, 1992). Norem and Cantor (1990b) suggest future research into defensive pessimism should follow students in and out of qualitatively different life experiences (p. 204). It would be interesting to assess such experiences using a process-oriented approach along the lines of that presented in this investigation.

The role of sex was not emphasised in the present study for three reasons. First, there were relatively few males in the study. This is because Education is typically comprised of females. Second, the invariance of the factor structure across sex was tested and findings showed that there were no marked differences between males and 334

females in terms of the number of factors, the factor loadings, the uniquenesses associated with each item parcel, and the correlations among factors. Third, in the CFA in which the relationship between sex and all other constructs was examined, no systematic effects were obtained and this was considered to provide further justification for not incorporating it in further analyses. The findings, then, apply largely to females and so subsequent samples in which the issues raised in the present study are examined should aim to comprise more males. Indeed, examining the present issues in other disciplines (e.g., Economics, Science) would probably remediate the bias in sample composition.

The present study adopted an individual-differences approach to the factors predicting self-handicapping and defensive pessimism. It is recognised, however, that class-level factors also impact on these strategies. For example, some recent work has assessed class-level ability goals (rather than individuals ego-orientation) and found that these class-level goals not only contribute to the tendency to self-handicap (see Urdan et al., 1998; Midgley & Urdan, 1995), but are associated with maladaptive patterns of behaviour, affect, and cognition more generally (Roeser et al., 1996). Indeed, recent advances in statistical software enable researchers to more accurately assess the relative influence of individual and class-level factors using multi-level modelling (see Bryk & Raudenbush, 1992; Goldstein, 1995) and a recent study into selfhandicapping has done precisely this (Urdan et al., 1998). In fact, researchers have recently incorporated multi-level modelling procedures into their structural equation modelling (Rowe & Hill, 1997). Future research, then, might explore the influence of class-level goal climate on self-handicapping and defensive pessimism using a process-oriented approach that incorporates both structural equation and multi-level modelling. Middleton and Midgley (1997) recommend that such research not only examine the traditional ego- and task-related class level climates but extend existing research by exploring how avoidance orientation operates at a class level.

13.4.4 Statistical issues In some respects, aspects of statistical analyses were complicated by some collinearity amongst constructs which led to suppression effects (see Bollen, 1989; Tabachnick & Fidell, 1996). Whilst for the most part this collinearity was amply dealt with using 335

higher order CFA, in some cases it required the use of submodels to clarify effects primarily related to the strategies at Time 2. Indeed, this is one of the costs of assessing conceptually-akin constructs (e.g., three strategies proposed to be selfprotective). Analyses involving the Time 2 processes, then, were not able to directly assess the joint operation of the three strategies without incurring collinearity effects and in this respect are qualified. Another issue concerns the sample size in relation to the number of parameters estimated in some models. Where possible, the number of parameters to be estimated was kept to a minimum, but in some models this was difficult and so the usual caveats relating to small case-to-parameter ratios apply.

13.4.5 Consequences of the three strategies The focus of the present study was on educational processes related to selfhandicapping and defensive pessimism. Accordingly, the outcome component of the models comprised only academic constructs. There may, however, be other costs involved in self-handicapping and defensive pessimism. For example, Smith and Strube (1991) have shown that others tend to have negative impressions of selfhandicappers, whilst Cantor and Norem (1989) have found that defensive pessimists feel more fatigued after social engagement. Moreover, defensive pessimists have been found to be more variable in emotional reactions across situations than optimists. It appears, then, that there are social and emotional costs involved with each strategy and future research might explore the present process models with a view to clarifying the precise role of these social and emotional issues.

Related to this issue of the consequences of self-handicapping and defensive pessimism is how others perceive these strategies and the extent to which they serve the desired purpose of protection or enhancement of ones competence image. This was not addressed in the present study and has not received much attention in the academic domain. There is a small body of research that has shown self-handicapping to serve its protective purposes from the perspective of an audience (Luginbuhl & Palmer, 1991 - but its success in this respect has not been assessed over time). Importantly, however, the benefits accrued from such strategic manoeuvring are not without their costs. For example, whilst Luginbuhl and Palmer (1991; see also Smith & Strube, 1991) found that self-handicapping reduced others negative attributions to 336

the subjects ability, they were more inclined to think more negatively about other characteristics of the subject (see also Rhodewalt, Sanbonmatsu, Tschanz, Feick, & Waller, 1995; Baumgardner & Levy, 1988, on circumstances that mitigate the positive reception of self-handicapping by an audience). In terms of students high in defensive expectations, the qualitative data showed that whilst others reassured them and affirmed to them that they were competent, one respondent described how her friends were becoming savvy to her strategy and did not indulge her unrealistically low expectations. Research has shown that the extent to which individuals engage in impression management depends on how credible their presentation is in the eyes of others and how much the audience knows about the individual (Frey, 1978; Schlenker, 1975). Clearly, the public or self-presented dimensions of self-handicapping and defensive expectations only go so far and reach a point where they are no longer acceptable to an audience. It would be interesting to explore the defensive process beyond this point to see how individuals behave once their alibi or cushioning strategy no longer suffice.

13.4.6 The qualitative data The qualitative data indicated that some self-handicappers felt little control over their ability to desist from self-handicapping. For example, the most striking aspect of one interview with a self-handicapper was that he felt powerless in the face of his selfhandicapping. According to him, it was deep set and just happens. The data also indicated that some interviewees would quite consciously manoeuvre in a defensive fashion whereas others were less aware of their strategic attempts to protect the self. Perceived control over ones actions and behaviours is pivotal to effective functioning (Conway & Terry, 1992; Gamble, 1994). Greater understanding of students perceptions of control over their defensive manoeuvring is required to better inform educational interventions designed to promote more adaptive functioning.

Cultural factors also underpin some students defensive manoeuvring. For example, some interviewees felt extreme pressure on them to do well, arising, they reported, from cultural mores. They further reported that their unrealistic negative expectations were a way of dealing with this. Whilst some research has examined cultural differences in relation to self-handicapping and negative attitudes (e.g., Midgley et al., 337

1996), this tends to be restricted to ethnic minorities in the American context. Moreover, the Midgley et al. data indicated that the role of ethnicity is a potentially complex one in that it interacted with ego goals in predicting self-handicapping. Further cross-cultural comparisons, then, are needed. Self (1990) notes that some subcultures agree on the excuse-value of various handicaps whilst other handicaps may not be viable. Thus, there should be differential preferences for some handicaps over others as a function of ones subculture and this requires further research.

Related to the issue of culture is the role of the family in influencing students tendency to engage in self-protective strategies. The Times 1 and 2 interviews were suggestive of some family influence in this respect, but further work is required to determine the precise nature and extent of this influence. Midgley and Urdan (1995) make the point that because self-handicapping arises from a history of non-contingent feedback (see also Berglas & Jones, 1978; Jones & Berglas, 1978), it would be interesting to survey parents about the type of feedback they give to their children and the relationship between these reports and childrens self-handicapping. Importantly, the present data indicated that family-related factors are quite relevant to students even at the university level.

An extension of interviewees motivation orientation was their perception of the learning climate. The Time 1 qualitative data indicated that self-handicappers and defensive pessimists tended to see the learning climate as competitive and this evoked a need to outperform others. This supports previous quantitative findings (Midgley & Urdan, 1995; Urdan et al., 1998), but some issues remain unclear. For example, the learning climate and its impact on defensive pessimism has apparently not been addressed to date. Also, the impact of altering the learning climate (e.g., from competitive to cooperative - see Johnson & Johnson, 1989; Qin, Johnson, & Johnson, 1995) on students academic strategies has not been studied and future research might focus on this with a view to identifying interventions that can take place at class and institutional levels.

The qualitative data also illuminated the diversity of self-handicapping behaviour in which students engage. The questions that remain concern the differential influence of 338

various types of self-handicapping on academic outcomes and whether certain factors give rise to particular forms of self-handicapping more than others. Whilst some research has been conducted exploring self-presented forms of self-handicapping (see DeGree & Snyder, 1985; Hirt et al., 1991; Smith et al., 1983) as opposed to more active self-handicapping (see Berglas & Jones, 1978; Higgins & Harris, 1988; Rhodewalt & Davison, 1986; Shepperd & Arkin, 1989a, 1989b; Tice & Baumeister, 1990), the qualitative data suggested that there are a variety of active forms of selfhandicapping and the educational implications of the multifaceted nature of active self-handicapping has not yet been explored (see also Arkin & Oleson, 1998; Self, 1990). Moreover, whilst the present study addressed self-handicapping that students were prepared to acknowledge, there is also the issue of self-deceptive selfhandicapping (Arkin & Oleson, 1998) and this distinction, particularly as it relates to education, has not been assessed to date.

Finally, on a methodological note, the contribution of the qualitative data to quantitative findings underscores the importance of the interface and integration of quantitative and qualitative perspectives. At a micro level, the qualitative component utilised psychometric scales to identify contrasting levels of self-handicapping and defensive expectations, pursued constructs typically incorporated in quantitative research, and sampled students in a way that enhanced representativeness. At a more macro level, the study has shown how qualitative methods can be located within and contribute to a quantitative literature. It is proposed that more work along these lines can enhance current educational practise and research, particularly in those areas that to date have been dominated by quantitative perspectives.

339

14. CONCLUSION

The present study has been centrally concerned with the variety of factors predicting self-handicapping, defensive expectations, and reflectivity, and the effect of these three strategies upon academic outcomes such as self-regulation, persistence, future academic plans, and grades. These issues were examined in process and longitudinal models which drew upon previous theory and research pertaining to the factors giving rise to self-worth protection. These models provided an encompassing perspective on the processes related to defensive manoeuvring and the specific means by which students self-protect. Addressing these issues in such a manner extends the bulk of previous work that has typically examined the predictors and consequences separately.

In the process of presenting this encompassing perspective, the study contributes to current understanding of factors comprising self-worth motivation by systematically unifying a variety of self-worth-related constructs and linking them to selfhandicapping, defensive expectations, and reflectivity. Moreover, the higher order factors which were proposed to be indicants of a self-worth motivation were validated through their hypothesised associations with the three strategies. In addition to the higher order factors underpinning self-worth motivation, active and self-presented dimensions of self-handicapping and defensive expectations were also incorporated into higher order self-handicapping and defensive expectations factors. Representing active and self-presented self-handicapping and defensive expectations in this way extends previous research which to date has not assessed these dimensions using formal measurement techniques such as those incorporated in the present study. Moreover, whilst some research has addressed the issues of active and self-presented self-handicapping, the proposition of self-presented defensive expectations is new and so the present data have contributed to current understanding about defensive expectations in this respect.

In the course of addressing the process and longitudinal models, a number of other issues relevant to self-worth protection were addressed. For example, a quadripolar need achievement model clarified the self-protective dimensions of defensive expectations and reflectivity and also provided a new (failure accepting) perspective 340

on self-handicapping. Moreover, the strength of the (negative) associations between self-handicapping and academic outcomes in addition to the data related to nonattendance on the second occasion of testing further supported self-handicappers status as possibly more failure accepting than failure avoiding. Indeed, this quadripolar framework was useful in accounting for present findings that were discrepant from previous work in the area. This framework and the process models also identified a distinction between defensive expectations and reflectivity and it was concluded that whilst these two strategies are self-protective, their relationships to subsequent academic outcomes are quite distinct.

While the emphasis of the present investigation was on the quantitative components, two qualitative studies provided a number of important perspectives on the nature of self-handicapping and defensive expectations and the factors related to them. The qualitative study at Time 1 illuminated many of the key relationships in the Time 1 process model and provided insight into the specific ways in which such relationships were manifested in students academic lives from their own perspectives. Qualitative data at Time 2 contributed to the longitudinal model by shedding light on key shifts in self-handicapping and defensive expectations and students perceptions as to why such shifts occurred. Taken together, the qualitative data were valuable not only in grounding the quantitative findings in personal accounts of the variety of challenges first and second year students face in their academic lives, but also in providing a very real sense of the rich detail in which these challenges are met by students.

In sum, then, the yields of the present study are multifold: the study has provided confirmatory measurement work on self-handicapping, defensive expectations, and reflectivity and has clarified the factors giving rise to them and the academic outcomes that follow from them; these issues were assessed using a process model that captured both predictors and consequences of self-handicapping, defensive expectations, and reflectivity, thus extending previous research which typically examines predictors and consequences separately; the study extends the self-worth motivation research, not only by directly locating it, theoretically and empirically, in the areas of self-handicapping, defensive 341

expectations, and reflectivity, but also by identifying a variety of (primarily higher order) factors argued to be indicants of a self-worth motivation; the study is among the first to explicitly draw together self-handicapping, defensive expectations, and reflectivity under a common conceptual framework and to assess their respective effects in an hypothesised model of the academic process as is relevant to self-worth protection; in clarifying the constructs underlying self-handicapping, defensive expectations, and reflectivity, a more comprehensive understanding of the process of defensive manoeuvring as it relates to achievement has been advanced; the research adds to the rather small body of work into either self-handicapping or defensive pessimism that assesses these constructs in a longitudinal fashion and through this longitudinal process, the study has contributed to current understanding about and provided new perspectives on the processes related to defensive manoeuvring; the quantitative findings were enriched through qualitative analyses that not only illuminated the relationships derived in the quantitative models, but also provided important insights into shifts on key dimensions that occurred across students first and second years at university; the study extends current understanding about defensive expectations, reflectivity, and self-handicapping by drawing them together in a need achievement model of academic motivation and identifying their respective positions on failure-avoidance and success-orientation axes; and, in addition to these central issues, the quantitative analyses addressed a variety of supplementary issues which have not been given direct empirical attention in previous research and which further clarified the ways in which self-handicapping, defensive expectations, and reflectivity operate in students academic lives.

The quantitative and qualitative studies have provided a great deal of data on the lengths to which students go to protect their self-worth, the factors giving rise to this defensive manoeuvring, and the diversity and complexity of consequences that follow from such behaviour. In addition to these processes, the study has yielded important data relevant to assessments of defensive manoeuvring, attempts to study it, and

342

interventions designed to address it. Taken together, then, the findings of the present investigation hold not only substantive and methodological implications for researchers studying the causes, means, and consequences of self-worth protection, but are also relevant to educational practitioners operating in contexts in which students are motivated to manoeuvre defensively.

343

REFERENCES Abel, M.H. (1996). Self-esteem: Moderator or mediator between perceived stress and expectancy of success? Psychological Reports, 79, 635-641. Abott-Chapman, J. (1993). Is the debate on quantitative versus qualitative research really necessary? Australian Educational Researcher, 20, 49-63. Abramson, L.Y., Seligman, M.E.P., & Teasdale, J. (1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology, 87, 49-74. Aiken, L.S., & West, S.G. (1991). Multiple regression: Testing and interpreting interactions. London: Sage. Albersnagel, F.A., Arntz, A., Gerlsma, C. (1986). Some limitations of the attributional learned helplessness model on understanding effects of (non-) contingency: A controlled study in Dutch adolescents. Advances in Behaviour Research & Therapy, 8, 1-42. Alloy, L.B. (1982). The role of perception and attributions for response-outcome noncontingency in learned helplessness: A commentary and discussion. Journal of Personality, 50, 443-479. Ames, C. (1984). Achievement attributions and self-instructions under competitive and individualistic goal structures. Journal of Educational Psychology, 76, 478-487. Ames, C., & Archer, J. (1988). Achievement goals in the classroom: Students learning strategies and motivation processes. Journal of Educational Psychology, 80, 260-267. Andrews, G.R., & Debus, R.L. (1978). Persistence and the causal perception of failure: Modifying cognitive attributions. Journal of Educational Psychology, 70, 154166. Arkin, R.M., & Baumgardner, A.H. (1985). Self-handicapping. In J.H. Harvey & G. Weary (Eds.). Basic issues in attribution theory and research. New York: Academic Press. Arkin, R.M., & Oleson, K.C. (1998). Self-handicapping. In J.M. Darley & J. Cooper (Eds.). Attribution and social interaction: The legacy of Edward E. Jones. Washington DC: American Psychological Association. Atkinson, J.W. (1957). Motivational determinants of risk-taking. Psychological Review, 64, 359-372. Atkinson, J.W. (1958) (Ed.). Motives in fantasy, action, and society. Princeton, NJ: Van Nostrand.

344

Atkinson, D.W. (1964). An introduction to motivation. Princeton, NJ: Van Nostrand. Atkinson, J.W., & Feather, N.T. (1966) (Eds.). A theory of achievement motivation. New York: Wiley. Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122-147. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. New Jersey: Prentice Hall. Bandura, A. (1997). Self-Efficacy: The exercise of control. New York: Freeman & Co. Bandura, A., & Cervone, D. (1983). Self-evaluative and self-efficacy mechanisms governing the motivational effects of goal systems. Journal of Personality and Social Psychology, 45, 1017-1028. Baumeister, R.F., Hamilton, J.C., & Tice, D.M. (1985). Public versus private expectancy of success: Confidence booster or performance pressure. Journal of Personality and Social Psychology, 48, 1447-1457. Baumeister, R.F., & Scher, S.J. (1988). Self-defeating behavior patterns among normal individuals: Review and analysis of common self-destructive tendencies. Psychological Bulletin, 104, 3-22. Baumeister, R.F., Tice, D.M., & Hutton, D.G. (1989). Self-presentational motivations and personality differences in self-esteem. Journal of Personality, 57, 547-579. Baumgardner, A.H., & Brownlee, E.A. (1987). Strategic failure in social interaction: Evidence for expectancy disconfirmation processes. Journal of Personality and Social Psychology, 52, 525-535. Baumgardner, A.H., Lake, E.A., & Arkin, R.M. (1985). Claiming mood as a selfhandicap: The influence of spoiled and unspoiled public identities. Personality and Social Psychology Bulletin, 11, 349-357. Baumgardner, A.H., & Levy, P.E. (1988). Role of self-esteem in perceptions of ability and effort: Illogic or insight? Personality and Social Psychology Bulletin, 14, 429-438. Beckman, L.J. (1970). Effects of students performance on teachers and observers attributions of causality. Journal of Educational Psychology, 61, 76-82. Beery, R.G. (1975). Fear of failure in the student experience. Personnel and Guidance Journal, 54, 191-203. 345

Bempechat, J., London, P., & Dweck, C.S. (1991). Childrens conceptions of ability in major domains: An interview and experimental study. Child Study Journal, 21, 11-36. Benenson, J., & Dweck, C. (1986). The development trait explanations and selfevaluations in the academic and social domains. Child Development, 57, 1179-1187. Berglas, S. (1985). Self-handicapping and self-handicappers: A cognitive/attributional model of interpersonal self-protective behavior. In R. Hogan & W.H. Jones (Eds.). Perspectives in personality: Theory, measurement and interpersonal dynamics. Greenwich, CT: JAI Press. Berglas, S. (1987). Self-handicapping model. In H.T. Blane & K.E. Leonard (Eds.). Psychological theories of drinking and alcoholism. New York: Guildford Press. Berglas, S. (1990). Self-handicapping: Etiological and diagnostic considerations. In R.L. Higgins, C.R. Snyder., & S. Berglas (Eds.). Self-handicapping: The paradox that isnt. New York: Plenum Press. Berglas, S., & Jones, E.E. (1978). Drug choice as a self-handicapping strategy in response to noncontingent success. Journal of Personality and Social Psychology, 36, 405-417. Bessant, K.C. (1995). Factors associated with types of mathematics anxiety in college students. Journal for Research In Mathematics Education, 26, 327-345. Birney, R.C., Burdick, H., & Teevan, R.C. (1969). Fear of failure. New York: Van Nostrand. Bogdan, R.C., & Biklen, S.K. (1982). Qualitative research for education: An introduction to theory and methods. Boston: Allyn & Bacon Inc. Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley. Boyd, M., & Callaghan, J. (1994). Task and ego goal perspectives in organized youth sport. International Journal of Sport Psychology, 22, 411-424. Brewer, J., & Hunter, A. (1989). Multimethod research: A synthesis of styles. Newbury Park, CA: Sage. Brock, T.C. (1968). Implications of commodity theory for value change. In A.G. Greenwald., T.C. Brock., & T.M. Ostrom (Eds.) Psychological foundations of attitudes. New York: Academic Press. Brookover, W.B., Schweitzer, J.H., Schneider, J.M., Beady, C.H., Flood, P.K., & Wisenbaker, J.M. (1978). Elementary school social climate and school achievement. American Educational Research Journal, 15, 301-318. 346

Brophy, J. (1981). Teacher praise: A functional analysis. Review of Educational Research, 51, 5-32. Brown, J., & Weiner, B. (1984). Affective consequences of ability versus effort ascriptions: Controversies, resolutions, and quandaries. Journal of Educational Psychology, 76, 146-158. Bryk, A.S., & Raudenbush, S.W. (1992). Hierarchical linear models. Newbury Park, CA: Sage. Busemeyer, J.R., & Jones, L.E. (1983). Analysis of multiplicative combination rules when the causal variables are measured with error. Psychological Bulletin, 93, 549-562. Buss, A.H. (1980). Self-consciousness and social anxiety. San Francisco: Freeman. Buss, D.W., & Cantor, N. (1989). Personality psychology: Recent trends and emerging directions. New York: Springer-Verlag. Byrne, B.M. (1994). Structural equation modeling with EQS and EQS/Windows. Thousand Oaks: Sage. Campbell, D.T. (1975). Degrees of freedom and the case study. Comparative Political Studies, 8, 178-193. Campbell, J.D., & Fairey, P.J. (1985). Effects of self-esteem hypothetical explanations, and verbalizations of expectancies on future performance. Journal of Personality and Social Psychology, 48, 1097-1111. Campbell, J.D., Trapnell, P.D., Heine, S.J., Katz, I.M., Lavallee, L.F., & Lehman, D.R. (1996). Self-concept clarity: Measurement, personality correlates and cultural boundaries. Journal of Personality and Social Psychology, 70, 141-156. Cantor, N., & Norem, J.K. (1989). Defensive pessimism and stress and coping. Social Cognition, 7, 92-112. Cantor, N., Norem, J.K., Niedenthal, P.M., Langston, C.A., & Brower, A.M. (1987). Life tasks, self-concept ideals, and cognitive strategies in a life transition. Journal of Personality and Social Psychology, 53, 1178-1191. Carver, C.S., & Humphries, C. (1981). Havana daydreaming: A study of selfconsciousness and the negative reference group among Cuban Americans. Journal of Personality and Social Psychology, 40, 545-552. Cervone, D., & Peake, P.K. (1986). Anchoring, efficacy, and action: The influence of judgmental heuristics on self-efficacy judgments and behaviour. Journal of Personality and Social Psychology, 50, 492-501. 347

Cheek, J.M., & Briggs, S.R. (1982). Self-consciousness and aspects of identity. Journal of Research in Personality, 16, 401-408. Clark, J., & Tollefson, N. (1991). Differences in beliefs and attitudes toward the improvability of writing of gifted students who exhibit mastery-oriented and helpless behaviors. Journal for the Education of the Gifted, 14, 119-133. Cohen, J., & Cohen, P. (1983). Applied multiple regression/Correlation analysis for the behavioral sciences. Hillsdale, NJ. Erlbaum. Connell, J.P. (1985). A new multidimensional measure of childrens perceptions of control. Child Development, 56, 1018-1041. Constas, M.A. (1992). Qualitative analysis as a public event: The documentation of category development procedures. American Educational Research Journal, 29, 253-266. Conway, V.J., & Terry, D.J. (1992). Appraised controllability as a moderator of the effectiveness of different coping strategies: A test of the goodness of fit hypothesis. Australian Journal of Psychology, 44, 1-7. Cook, T.D., & Campbell, D.T. (1979). Quasi-experimentation: Design and analysis issues for field settings. Chicago: Rand McNally. Covington, M.V. (1984a). The motive for self-worth. In R. Ames & C. Ames (Eds.). Research on motivation in education. Vol 1. New York: Academic Press. Covington, M.V. (1984b). The motive for self-worth theory of achievement motivation: Findings and implications. Elementary School Journal, 85, 5-20. Covington, M.V. (1985). The role of self-processes in applied social psychology. Journal for the Theory of Social Behaviour, 15, 355-389. Covington, M.V. (1989). Self-esteem and failure in school: Analysis and policy implications. In A.M. Mecca., N.J. Smelser., & J. Vasconcellos (Eds.). The social importance of self-esteem. Berkeley: University of California Press. Covington, M.V. (1992). Making the grade: A self-worth perspective on motivation and school reform. Cambridge: Cambridge University Press. Covington, M.V. (1997). A motivational analysis of academic life in college. In R.P. Perry & J.C. Smart (Eds.). Effective teaching in higher education: Research and practice. New York: Agathon Press. Republished from Covington, M.V. (1993). A motivational analysis of academic life in college. Higher education: Handbook of theory and research. New York: Agathon Press. Covington, M.V., & Beery, R.G. (1976). Self-worth and school learning. New York: Holt, Rinehart, and Winston. 348

Covington, M.V., & Omelich, C.L. (1979a). Are causal attributions causal? A path analysis of the cognitive model of achievement motivation. Journal of Personality and Social Psychology, 37, 1487-1504. Covington, M.V., & Omelich, C.L. (1979b). Effort: The double-edged sword in school achievement. Journal of Educational Psychology, 71, 169-182. Covington, M.V., & Omelich, C.L. (1979c). Its best to be able and virtuous too: Student and teacher evaluative responses to successful effort. Journal of Educational Psychology, 71, 688-700. Covington, M.V., & Omelich, C.L. (1981). As failures mount: Affective and cognitive consequences of ability demotion in the classroom. Journal of Educational Psychology, 73, 796-808. Covington, M.V., & Omelich, C.L. (1984a). An empirical examination of Weiners critique of attribution research. Journal of Educational Psychology, 76, 1214-1225. Covington, M.V., & Omelich, C.L. (1984b). Controversies or consistencies? A reply to Brown and Weiner. Journal of Educational Psychology, 76, 159-168. Covington, M.V., & Omelich, C.L. (1984c). Task-oriented versus competitive learning structures: Motivational and performance consequences. Journal of Educational Psychology, 76, 1038-1050. Covington, M.V., & Omelich, C.L. (1985). Ability and effort valuation among failure-avoiding and failure-accepting students. Journal of Educational Psychology, 77, 446-459. Covington, M.V., & Omelich, C.L. (1991). Need achievement revisited: Verification of Atkinsons original 2 x 2 model. In C.D. Spielberger., I.G. Sarason., Z. Kulcsar., & G.L. Van Heck. (Eds.). Stress and emotion Vol 14. New York, NY: Hemisphere. Covington, M.V., & Roberts, B.W. (1994). Self-worth and college achievement: Motivational and personality correlates. In P.R. Pintrich., D.R. Brown., & C.E. Weinstein (Eds.). Student motivation, cognition, and learning: Essays in honor of Wilbert J. McKeachie. Hillsdale, NJ: Erlbaum. Covington, M.V., Spratt, M.F., & Omelich, C.L. (1980). Is effort enough, or does diligence count too? Student and teacher reactions to effort stability in failure. Journal of Educational Psychology, 72, 717-729. Craven, R.G., Marsh, H.W., & Debus, R.L. (1991). Effects of internally focused feedback and attributional feedback on the enhancement of academic self-concept. Journal of Educational Psychology, 83, 17-26.

349

Cronbach, L.J. (1975). Beyond the two disciplines of scientific psychology. American Psychologist, 30, 116-127. Davis, S.F., Hanson, H., Edson, R., & Ziegler, C. (1992). The relationship between optimism-pessimism, loneliness, and level of self-esteem in college students. College Student Journal, 26, 244-247. Deci, E.L. (1975) Intrinsic motivation. New York: Plenum Press. Deci, E.L., & Ryan, R.M. (1980). The empirical exploration of intrinsic motivational processes. In L. Berkowitz (Ed.). Advances in experimental social psychology. New York: Academic Press. DeGree, C.E., & Snyder, C.R. (1985). Adlers psychology (of use) today: Personal history of traumatic life events as a self-handicapping strategy. Journal of Personality and Social Psychology, 48, 1512-1519. Deppe, R.K., & Harackiewicz, J.M. (1996). Self-handicapping and intrinsic motivation: Buffering intrinsic motivation from the threat of failure. Journal of Personality and Social Psychology, 70, 868-876. Diener, C.I., & Dweck, C.S. (1978). An analysis of learned helplessness: Continuous changes in performance, strategy, and achievement cognitions following failure. Journal of Personality and Social Psychology, 36, 451-462. Diener, C.I., & Dweck, C.S. (1980). An analysis of learned helplessness II. The processing of success. Journal of Personality and Social Psychology, 39, 940-952. Drexler, L.P., Ahrens, A.H., & Haaga, D.A.P. (1995). The affective consequences of self-handicapping. Journal of Social Behavior and Personality, 10, 861-870. Drwal. L & Wiechnik, R. (1984). The effect of locus of control and self-esteem on attributions and expectancies after success and failure. Polish Psychological Bulletin, 15, 257-266. Duda, J.L. (1992). Motivation in sport settings: A goal perspective approach. In G.C. Roberts (Ed.). Motivation in sport and exercise. Champaign, IL: Human Kinetics Books. Duda, J.L. (1993). Goals: A social-cognitive approach to the study of achievement motivation in sport. In R.N. Singer., M. Murchey., & L. K. Tennant (Eds.). Handbook of research on sport psychology. New York: Macmillan. Duda, J.L., Fox, K.R., Biddle, S.J.H., & Armstrong, N. (1992). Childrens achievement goals and beliefs about success in sport. British Journal of Educational Psychology, 62, 313-323.

350

Duda, J.L., & Nicholls, J.G. (1992). Dimensions of achievement motivation in schoolwork and sport. Journal of Educational Psychology, 84, 290-299. Dweck, C.S. (1975). The role of expectations and attributions in the alleviation of learned helplessness. Journal of Personality and Social Psychology, 31, 674-685. Dweck, C.S. (1986). Motivational processes affecting learning. American Psychologist, 41, 1040-1048. Dweck, C.S. (1991). Self-theories and goals: Their role in motivation, personality, and development. In R.A. Dienstbier (Ed.). Perceptions on motivation: Nebraska Symposium on Motivation. Vol 38. Lincoln: University of Nebraska Press. Dweck, C.S., & Bempechat, J. (1983). Childrens theories of intelligence: Consequences for learning. In S.G. Paris, G.M. Olson, & H.W Stevenson. (Eds.). Learning and motivation in the classroom. NJ: Erlbaum. Dweck, C.S., Hong, Y., & Chiu, C. (1993). Implicit theories: Individual differences in the likelihood and meaning of dispositional inference. Personality and Social Psychology Bulletin, 19, 644-656. Dweck, C, S., & Leggett, E.L. (1988). A social-cognitive approach in motivation and personality. Psychological Review, 95, 256-273. Eccles, J. S., Midgley, C., & Adler, T. (1984). Grade-related changes in the school environment: Effects on achievement motivation. In J. G. Nicholls (Ed.), The development of achievement motivation. Greenwich, CT: JAI. Eccles, J.S., Wigfield, A., Harold, R., & Blumenfeld, P.B. (1993). Age and gender differences in childrens self- and task perceptions during elementary school. Child Development, 64, 830-847. Elliot, A.J., & Church, M.A. (1997). A hierarchical model of approach and avoidance achievement motivation. Journal of Personality and Social Psychology, 72, 218-232. Elliot, E.S., & Dweck, C.S. (1988). Goals: An approach to motivation and achievement. Journal of Personality and Social Psychology, 54, 5-12. Elliot, A.J., & Harackiewicz, J.M. (1994). Goal setting, achievement orientation, and intrinsic motivation: A motivational analysis. Journal of Personality and Social Psychology, 66, 968-980. Elliot, A.J., & Harackiewicz, J.M. (1996). Approach and avoidance achievement goals and intrinsic motivation: A mediational analysis. Journal of Personality and Social Psychology, 70, 968-980.

351

Epstein, J.A., & Harackiewicz, J.M. (1992). Winning is not enough: The effects of competition and achievement orientation on intrinsic interest. Personality and Social Psychology Bulletin, 18, 128-138. Eronen, S., Nurmi, J., & Salmela-Aro, K. (1998). Optimistic, defensive pessimistic, impulsive and self-handicapping strategies in university environments. Learning and Instruction, 8, 159-177. Feather, N.T. (1967). Level of aspiration and performance variability. Journal of Personality and Social Psychology, 6, 37-46. Feather, N.T. (1969). Attribution of responsibility and valance of success and failure in relation to initial confidence and task performance. Journal of Personality and Social Psychology, 7, 129-144. Federoff, N.A., & Harvey, J.H. (1976). Focus of attention, self-esteem, and the attribution of causality. Journal of Research in Personality, 10, 336-345. Felsenthal, H.M. (1970). Sex differences in teacher-pupil interactions and their relationship with teacher attitudes and pupil reading achievement. Dissertation Abstracts, 30, 3781-3782. Fenigstein, A. (1979). Self-consciousness, self-attention, and social interaction. Journal of Personality and Social Psychology, 37, 75-86. Fenigstein, A., Scheier, M.F., & Buss, A.H. (1979). Public and private selfconsciousness: Assessment and theory. Journal of Consulting and Clinical Psychology, 43, 522-527. Ferrari, J.R. (1991). Self-handicapping by procrastinators: Protecting selfesteem, social-esteem, or both. Journal of Research in Personality, 25, 245-261. Ferrari, J.R. (1992). Procrastinators and perfect behavior: An exploratory factoranalysis of self-presentation, self-awareness, and self-handicapping components. Journal of Research in Personality, 26, 75-84. Filstead, W.J. (1970). Qualitative methodology. Chicago: Markham. Firestone, W. (1987). Meaning in method: The rhetoric of quantitative and qualitative research. Educational Researcher, 16, 16-21. Fischer, M., & Leitenberg, H. (1986). Optimism and pessimism in elementary school-aged children. Child Development. 57, 241-248. Forsyth, D.R., & McMillan, J.H. (1981). Attributions, affect, and expectations: A test of Weiners three-dimensional model. Journal of Educational Psychology, 73, 393-403.

352

Frankel, A., & Snyder, M.L. (1978). Poor performance following unsolvable problems: Learned helplessness or egotism? Journal of Personality and Social Psychology, 36, 1415-1423. Frey, D. (1978). Reaction to success and failure in public and in private conditions. Journal of Experimental Social Psychology, 14, 172-179. Fried-Buchalter, S. (1992). Fear of success, fear of failure, and the imposter phenomenon: A factor analytic approach to convergent and discriminant validity. Journal of Personality Assessment, 58, 368-379. Froming, W.J., & Carver, C.S. (1981). Divergent influences of private and public self-consciousness in a compliance paradigm. Journal of Research in Personality, 15, 159-171. Gage, N. (1989). The paradigm wars and their aftermath. Educational Researcher, 18, 4-10. Galloway, D., Leo, E.L., Rogers, C., & Armstrong, D. (1996). Maladaptive motivational style: The role of domain specific task demand in English and mathematics. British Journal of Educational Psychology, 66, 197-207. Gamble, W.C. (1994). Perceptions of controllability and other stressor event characteristics as determinants of coping among young adolescents and young adults. Journal of Youth and Adolescence, 23, 65-84. Garcia, T., Matula, J.S., Harris, C.L., Egan-Dowdy, K., Lissi, M.R., & Davila, C. (1995). Worriers and procrastinators: Differences in motivation, cognitive engagement, and achievement between defensive pessimists and self-handicappers. Paper presented at Annual American Educational Research Association Conference, San Francisco, CA. Garcia, T., & Pintrich, P.R. (1994). Regulating motivation and cognition in the classroom: The role of self-schemas and self-regulatory strategies. In D.H. Schunk & B.J. Zimmerman (Eds.). Self-regulation of learning and performance: Issues and educational applications. Hillsdale NJ: Erlbaum. Gibbons, F.X., & Gaeddert, W.P. (1984). Focus of attention and placebo utility. Journal of Experimental Social Psychology, 20, 159-176. Gill, J. (1996). Beyond the baby and the bath water: Reflections on current debates in educational research. Australian Educational Researcher, 23, 31-46. Glesne, C., & Peshkin, A. (1992). Becoming qualitative researchers: An introduction. New York: Longman. Goffman, E. (1959). The presentation of self in everyday life. Garden City, NY: Doubleday Anchor. 353

Goldstein, H. (1995). Multilevel statistical models. London: Edward Arnold. Goodhart, D.E. (1986). The effects of positive and negative thinking on performance in an achievement situation. Journal of Personality and Social Psychology, 51, 117-124. Gottfried, A.E. (1982). Relations between academic intrinsic motivation and anxiety in children and young adolescents. Journal of School Psychology, 20, 205-215. Gottfried, A. E., Fleming, J. S., & Gottfried, A. W. (1994). Role of parental motivational practices in childrens academic intrinsic motivation and achievement. Journal of Educational Psychology, 86, 104-113. Graham, S., & Golan, S. (1991). Motivational influences on cognition: Task involvement, ego involvement, and depth of information processing. Journal of Educational Psychology, 83, 187-194. Greenberg, J. (1985). Unattainable goal choice as a self-handicapping strategy. Journal of Applied Social Psychology, 15, 140-152. Greenberg, J., Pyszczynski, T., & Paisley, C. (1985). Effect of extrinsic incentives on use of test anxiety as an anticipatory attributional defence: Playing it cool when the stakes are high. Journal of Personality and Social Psychology, 47, 11361145. Greenwald, A.G. (1982). Ego task analysis: An integration of research on egoinvolvement and self-awareness. In A.H. Hastorf & A.M. Isen (Eds.). Cognitive social psychology. New York: Elsevier. Guba, E.G., & Lincoln, Y.S. (1981). Effective evaluation: Improving the usefulness of evaluation results through responsive and naturalistic approaches. San Francisco: Jossey-Bass. Hair, J.F., Anderson, R.E., Tatham, R.L., & Black, W.C. (1995). Multivariate data analysis with readings. Englewood Cliffs, NJ: Prentice-Hall. Harackiewicz, J.M., Barron, K.E., Carter, S.M., Lehto, A.T., & Elliot, A.J. (1997). Predictors and consequences of achievement goals in the college classroom: Maintaining interest and making the grade. Journal of Personality and Social Psychology, 73, 1284-1295. Harackiewicz, J.M., Barron, K.E., & Elliot, A.J. (1998). Rethinking achievement goals: When are they adaptive for college students and why? Educational Psychologist, 33, 1-21. Harackiewicz, J.M., & Elliot, A.J. (1993). Achievement goals and intrinsic motivation. Journal of Personality and Social Psychology, 65, 904-915.

354

Harackiewicz, J.M., & Manderlink, G. (1984). A process analysis of the effects of performance-contingent rewards on intrinsic motivation. Journal of Experimental Social Psychology, 20, 531-551. Harackiewicz, J.M., & Sansone, C. (1991). Goals and intrinsic motivation: You can get there from here. In M.L. Maehr & P.R. Pintrich (Eds.). Advances in motivation and achievement. Greenwich, CT: JAI. Harackiewicz, J.M., Sansone, C., & Manderlink, G. (1985). Competence, achievement orientation, and intrinsic motivation: A process analysis. Journal of Personality and Social Psychology, 48, 493-508. Harari, O., & Covington, M.V. (1981). Reactions to achievement behavior from a teacher and student perspective: A developmental analysis. American Educational Research Journal, 18, 15-28. Harris, R.N., & Snyder, C.R. (1986). The role of uncertain self-esteem in selfhandicapping. Journal of Personality and Social Psychology, 51, 451-458. Harris, R.N., Snyder, C.R., Higgins, R.L., & Schrag, J.L. (1986). Enhancing the prediction of self-handicapping. Journal of Personality and Social Psychology, 51, 1191-1199. Harter, S., & Connell, J.P. (1984). A model of childrens achievement and related self-perceptions of competence, control, and motivation orientation. In J. Nicholls (Ed.). The development of achievement motivation. London: JAI Press. Harter, S., & Jackson, B. K. (1992). Trait vs. nontrait conceptualizations of intrinsic/extrinsic motivational orientation. Motivation & Emotion, 16, 209-230. Harter, S., Whitesell, N.R., & Kowalski, P. (1992). Individual differences in the effects of educational transitions on young adolescents perceptions of competence and motivational orientation. American Educational Research Journal, 29, 777-807. Hattie, J. (1992). Self-concept. Hillsdale, NJ: Erlbaum. Hattie, J., & Marsh, H.W. (1996). Future directions in self-concept research. In B.A. Bracken (Ed.). Handbook of self-concept: Developmental, social, and clinical considerations. New York: John Wiley & Sons. Hattie, J., & Watkins, D. (1988). Preferred classroom environment and approach to learning. British Journal of Educational Psychology, 58, 345-349. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Hewitt, P.L., Flett, G.L., Turnbull-Donovan, W., & Mikail, S.F. (1991). The Multidimensional Perfectionism Scale: Reliability, validity, and psychometric properties in psychiatric samples. Psychological Assessment, 3, 464-468. 355

Higgins, R.L. (1990). Self-handicapping: Historical roots and contemporary branches. In R.L. Higgins, C.R. Snyder., & S. Berglas (Eds.). Self-handicapping: The paradox that isnt. New York: Plenum Press. Higgins, R.L., & Berglas, S. (1990). The maintenance and treatment of selfhandicapping: From risk-taking to face-saving - and back. In R.L. Higgins, C.R. Snyder., & S. Berglas (Eds.). Self-handicapping: The paradox that isnt. New York: Plenum Press. Higgins, R.L., & Harris, R.N. (1988). Strategic alcohol use: Drinking to selfhandicap. Journal of Social and Clinical Psychology, 6, 191-202. Hirt, E.R., Deppe, R.K., & Gordon, L.J. (1991). Self-reported versus behavioral self-handicapping: Empirical evidence for a theoretical distinction. Journal of Personality and Social Psychology, 61, 981-991. Hobden, K., & Pliner, P. (1995). Self-handicapping and dimensions of perfectionism: Self-presentation vs self-protection. Journal of Research in Personality, 29, 461-474. Howe, K. (1988). Against a quantitative-qualitative incompatibility thesis, or dogmas die hard. Educational Researcher, 17, 10-16. Hu, L.T., & Bentler, P.M. (1995). Evaluating model fit. In R.H. Hoyle (Ed.). Structural equation modeling: Concepts, issues, and applications. Thousand Oaks, CA: Sage. Illingworth, K.S., & Norem, J.K. (1991). Convergent and divergent correlates of optimism and defensive pessimism. Paper presented at Midwestern Psychological Association Convention, Chicago, Illinois. Ingram, R.E., & Smith, T.W. (1984). Depression and internal versus external focus of attention. Cognitive Therapy and Research, 8, 139-152. Isleib, R.A., Vuchinich, R.E., & Tucker, J.A. (1988). Performance attributions and changes in self-esteem following self-handicapping with alcohol consumption. Journal of Social and Clinical Psychology, 6, 88-103. Jaccard, J., & Wan, C.K. (1995). Measurement error in the analysis of interaction effects between continuous predictors using multiple regression: Multiple indicator and structural equation approaches. Psychological Bulletin, 117, 348-357. Jaccard, J., & Wan, C.K. (1996). LISREL approaches to interaction effects in multiple regression. Thousand Oaks, CA: Sage.

356

Jagacinski, C.M., & Nicholls, J.G. (1984). Conceptions of ability and related effects in task involvement and ego involvement. Journal of Educational Psychology, 76, 909-919. Jagacinski, C.M., & Nicholls, J.G. (1987). Competence and affect in taskinvolvement and ego-involvement: The impact of social comparison information. Journal of Educational Psychology, 79, 107-114. Johnson, D.W., & Johnson, R.T. (1989). Cooperation and competition: Theory and research. Minnesota: Interaction. Jones, E.E., & Berglas, S. (1978). Control of attributions about the self through self-handicapping strategies: The appeal of alcohol and the role of underachievement. Personality and Social Psychology Bulletin, 4, 200-206. Jones, E.E., & Rhodewalt, F. (1982). The Self-Handicapping Scale. Princeton University. Jones, C.H., Slate, J.R., Blake, P.C., & Sloas, S. (1995). Relationship of study skills, conceptions of intelligence, and grade level in secondary school students. The High School Journal, 79, 25-32. Jones, C.H., Slate, J.R., Marini, I., & DeWater, B.K. (1993). Academic skills and attitudes toward intelligence. Journal of College Student Development, 34, 422-424. Joreskog, K.G. (1971). Statistical analysis of sets of congeneric tests. Psychometrika, 36, 109-133. Joreskog, K.G. (1979). Statistical estimation of structural models in longitudinal developmental investigations. In J.R. Nesselroade & P.B. Baltes (Eds.). Longitudinal research in the study of behavior and development. New York: Academic Press. Joreskog, K.G., & Sorbom, D. (1988). PRELIS. Chicago, IL: SPSS. Joreskog, K.G & Sorbom, D. (1989a). LISREL 7: Users reference guide. Chicago, IL: Scientific Software Inc. Joreskog, K.G & Sorbom, D. (1989b). LISREL 7: A guide to the program and applications. Chicago: SPSS Inc. Joreskog, K.G., & Yang, F. (1996). Nonlinear structural equation models: The Kenny-Judd model with interaction effects. In G.A. Marcoulides & R.E. Schumacker. (Eds.). Advanced structural equation modeling: Issues and techniques. Mahwah, NJ: Erlbaum.

357

Juvonen, J., & Murdock, T.B. (1993). How to promote social approval: Effects of audience and achievement outcome on publicly communicated attributions. Journal of Educational Psychology, 85, 365-376. Kelley, H.H. (1972). Causal schemata and the attribution process. Morristown NJ: General Learning Press. Kenny, D., & Judd, C.M. (1984). Estimating the nonlinear and interactive effects of latent variables. Psychological Bulletin, 96, 201-210. Kernis, M.H., Cornell, D.P., Sun, C.R., Berry, A., & Harlow, T. (1993). Theres more to self-esteem than whether it is high or low: The importance of stability of selfesteem. Journal of Personality and Social Psychology, 65, 1190-1204. Kernis, M.H., Grannemann, B.D., & Barclay, L.C. (1989). Stability and level self-esteem as predictors of anger arousal and hostility. Journal of Personality and Social Psychology, 56, 1013-1022. Kernis, M.H., Grannemann, B.D., & Barclay, L.C. (1992). Stability of selfesteem: Assessment, correlates, and excuse making. Journal of Personality, 60, 621644. Kernis, M.H., & Waschull, S.B. (1995). The interactive roles of stability and level of self-esteem: Research and theory. In M.P. Zanna (Ed.). Advances in Experimental Social Psychology. (Vol 27, pp. 93-141). San Diego: Academic Press. Kimble, C.E., Funk, S.C., & DaPolito, K.L. (1990). The effects of self-esteem certainty on behavioral self-handicapping. Journal of Social Behavior and Personality, 5, 137-149. Kinch, J.W. (1963). A formalised theory of self-concept. Journal of Social Psychology, 115, 25-29. Kleinhammer-Tramill, P.J., Tramill, J.L., Schrepel, S.N., & Davis, S.F. (1983). Learned helplessness in learning disabled adolescents as a function of noncontingent rewards. Learning Disability Quarterly, 6, 61-66. Koestner, R., Aube, J., Ruttner, J., & Breed, S. (1995). Theories of ability and the pursuit of challenge among adolescents with mild mental retardation. Journal of Intellectual Disability Research, 39, 57-65. Koestner, R., Zuckerman, M., & Koestner, J. (1987). Praise, involvement, and intrinsic motivation. Journal of Personality and Social Psychology, 53, 383-390. Kolditz, T.A., & Arkin, R.M. (1982). An impression management interpretation of the self-handicapping strategy. Journal of Personality and Social Psychology, 43, 492-502.

358

Leary, M.R. (1986). The impact of interactional impediments on social anxiety and self-presentation. Journal of Experimental Social Psychology, 22, 122-135. Leary, M.R., & Shepperd, J.A. (1986). Behavioral self-handicaps versus selfreported handicaps: A conceptual note. Journal of Personality and Social Psychology, 51, 1265-1268. Lefcourt, H.M., Von Baeyer, C.L., Ware, E.E., & Cox, D.J. (1979). The Multidimensional-Multiattributional Causality Scale: The development of a goal specific locus of control scale. Canadian Journal of Behavioural Science, 11, 286-304. Lennox, R. (1988). The problem with self-monitoring: A two-sided scale and a one-sided theory. Journal of Personality Assessment, 52, 58-73. Lepore, S., Kiely, M., Bempechat, J., & London, P. (1989). Childrens perceptions of social ability: Social cognitions and behavioral outcomes in the face of social rejection. Child Study Journal, 19, 259-271. Lincoln, Y., & Guba, E. (1985). Naturalistic inquiry. Beverly Hills: Sage. Lochbaum, M.R., & Roberts, G.C. (1993). Goal orientations and perceptions of the sport experience. Journal of Sport and Exercise Psychology, 15, 160-171. Luginbuhl, J., & Palmer, R. (1991). Impression management aspects of selfhandicapping: Positive and negative effects. Personality and Social Psychology Bulletin, 17, 655-662. Maehr, M. L. (1984). Meaning and motivation. In R. Ames & C. Ames (Eds.), Research in motivation: Vol. 1. Student motivation (pp. 115-144). Lexington, MA: D.C. Heath. Maehr, M. L., & Braskamp, L. A. (1986). The motivation factor: A theory of personal investment. Lexington, MA: Lexington Books. Marsh, H.W. (1984). Relationships among dimensions of self-attribution, dimensions of self-concept, and academic achievements. Journal of Educational Psychology, 76, 1291-1308. Marsh, H.W. (1986). Verbal and math self-concepts: An internal/external frame of reference model. American Educational Research Journal, 23, 129-149. Marsh, H.W. (1988a). Academic Self Description Questionnaire II. University of Western Sydney, Macarthur. Marsh, H.W. (1988b). The content specificity of math and English anxieties: The High School and Beyond Study. Anxiety Research, 1, 137-149.

359

Marsh, H.W. (1990). A multidimensional, hierarchical model of self-concept: Theoretical and empirical justification. Educational Psychology Review, 2, 77-172. Marsh, H.W. (1992). The content specificity of relations between academic achievement and academic self-concept. Journal of Educational Psychology, 84, 3542. Marsh, H. W. (1993a). Academic self-concept: Theory measurement and research. In J. Suls (Ed.), Psychological perspectives on the self (Vol. 4, pp. 59-98). Hillsdale, NJ: Erlbaum. Marsh, H.W. (1993b). Multitrait-multimethod analyses: Inferring each traitmethod combination with multiple indicators. Applied Measurement in Education, 6, 49-81. Marsh, H.W. (1993c). Stability of individual differences in multiwave panel studies: Comparison of simplex models and one-factor models. Journal of Educational Measurement, 30, 157-183. Marsh, H.W., Balla, J.R., & Hau, K.T. (1996). An evaluation of incremental fit indices: A clarification of mathematical and empirical processes. In G.A. Marcoulides & R.E. Schumacker (Eds.). Advanced structural equation modeling techniques. Hillsdale, NJ. Erlbaum. Marsh, H.W., Byrne, B.M., & Shavelson, R. (1988). A multifaceted academic self-concept: Its hierarchical structure and its relation to academic achievement. Journal of Educational Psychology, 80, 366-380. Marsh, H.W., Cairns, L., Relich, J.D., Barnes, J., & Debus, R. (1984). The relationship between dimensions of self-attribution and dimensions of self-concept. Journal of Educational Psychology, 76, 3-32. Marsh, H. W., & Craven, R. (1997). Academic self-concept: Beyond the dustbowl. In G. Phye (Ed.), Handbook of classroom assessment: Learning, achievement, and adjustment. Orlando, FL : Academic Press. Marsh, H. W., & Hau, K-T. (1996). Assessing goodness of fit: Is parsimony always desirable? Journal of Experimental Education, 64, 364-390. Marsh, H.W., & ONiell, R. (1984). Self Description Questionnaire III: The construct validity of multidimensional self-concept ratings by late adolescents. Journal of Educational Measurement, 21, 153-174. Marsh, H.W., Roche, L.A., Pajares, F., & Miller, D. (1997). Item-specific efficacy judgements in mathematical problem solving: The downside of standing too close to trees in a forest. Contemporary Educational Psychology, 22, 363-377.

360

Martin, A.J. & Debus, R.L. (1998). Self-reports of mathematics self-concept and educational outcomes: The roles of ego-dimensions and self-consciousness. British Journal of Educational Psychology, 68, 517-535. Mayerson, N.H., & Rhodewalt, F. (1988). Role of self-protective attributions in the experience of pain. Journal of Social and Clinical Psychology, 6, 203-218. McClelland, D.C. (1965). Toward a theory of motive acquisition. American Psychologist, 20, 321-333. McDonald, R.P., & Marsh, H.W. (1990). Choosing a multivariate model: Noncentrality and goodness of fit. Psychological Bulletin, 107, 247-255. McFarlin, D.B., & Blascovich, J. (1981). Effects of self-esteem and performance feedback on future affective preferences and cognitive expectations. Journal of Personality and Social Psychology, 40, 521-531. Meece, J.L., Blumenfeld, P.C., & Hoyle, R.H. (1988). Students goal orientation and cognitive engagement in classroom activities. Journal of Educational Psychology, 80, 514-523. Meece, J.L., & Holt, K. (1993). A pattern analysis of students achievement goals. Journal of Educational Psychology, 85, 582-590. Meichenbaum, D. (1977). Cognitive-behavior modification: An integrative approach. New York: Plenum Press. Mello-Goldner, D., & Wurf, E. (1997). The self in self-handicapping: Differential effects of public and private internal audiences. Current Psychology: Developmental, Learning, Personality, Social. 15, 319-331. Middleton, M.J., & Midgley, C. (1997). Avoiding the demonstration of lack of ability: An unexplored aspect of goal theory. Journal of Educational Psychology, 89, 710-718. Midgley, C., Arunkumar, R., & Urdan, T.C. (1996). If I dont do well tomorrow, theres a reason: Predictors of adolescents use of academic selfhandicapping strategies. Journal of Educational Psychology, 88, 423-434. Midgley, C., Kaplan, A., Middleton, M., & Maehr, M.L. (1998). The development and validation of scales assessing students achievement goal orientations. Contemporary Educational Psychology, 23, 113-131. Midgley, C., & Urdan, T. (1995). Predictors of middle school students use of self-handicapping strategies. Journal of Early Adolescence, 15, 389-411.

361

Miller, R.B., Greene, B.A., Montalvo, G.P., Ravindran, B., & Nichols, J.D. (1996). Engagement in academic work: The role of learning goals, future consequences, pleasing others, and perceived ability. Contemporary Educational Psychology, 21, 388-422. Montgomery, R.L., Haemmerlie, F.M., Zoellner, S. (1996). The imaginary audience, self-handicapping, and drinking patterns among college students. Psychological Reports, 79, 783-786. Moriarty, B., Douglas, G., Punch, K., & Hattie, J. (1995). The importance of self-efficacy as a mediating variable between learning environments and achievement. British Journal of Educational Psychology, 65, 73-84. Murray, C.B., & Warden, M.R (1992). Implications of self-handicapping strategies for academic achievement: A reconceptualization. Journal of Social Psychology, 132, 23-37. Musson, R.F., & Alloy, L.B. (1988). Depression and self-directed attention. In L.B. Alloy (Ed.). Cognitive processes in depression. New York: Guilford Press. Nasby, W. (1989). Private and public self-consciousness and articulation of selfschema. Journal of Personality and Social Psychology, 56, 117-123. Newman, L.S., & Wadas, R.F. (1997). When stakes are higher: Self-esteem instability and self-handicapping. Journal of Social Behavior and Personality, 12, 217232. Newton, M., & Duda, J.L. (1993). The relationship of task and ego orientation to performance cognitive content, affect, and attributions in bowling. Journal of Sport Behavior, 16, 209-220. Nicholls, J.G. (1976). Effort is virtuous, but its better to have ability: Evaluative response to perceptions of effort and ability. Journal of Research in Personality, 10, 306-315. Nicholls, J.G. (1989). The competitive ethos and democratic education. Cambridge: Harvard University Press. Nisbett, R.E., & Wilson, T.O. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231-259. Nolen, S. B., & Haladyna, T. M. (1990). Motivation and studying in high school science. Journal of Research in Science Teaching, 27, 115-126. Norem, J. K., & Cantor, N. (1986a). Anticipating and post hoc cushioning strategies: Optimism and defensive pessimism in risky situations. Cognitive Therapy and Research, 10, 347-362.

362

Norem, J.K., & Cantor, N. (1986b). Defensive pessimism: Harnessing anxiety as motivation. Journal of Personality and Social Psychology, 51, 1208-1217. Norem, J.K., & Cantor, N. (1990a). Capturing the flavor of behavior: Cognition, affect, and interpretation. In B.S. Moore & A.M. Isen (Eds.). Affect and social behavior. Cambridge: Cambridge University Press. Norem, J.K., & Cantor, N. (1990b). Cognitive strategies, coping and perceptions of competence. In R.J. Sternberg & J. Kolligian Jr. (Eds.). Competence considered. New Haven CT: Yale University Press. Norem, J.K., & Illingworth, K.S.S. (1993). Strategy-dependent effects of reflecting on self and tasks: Some implications of optimism and defensive pessimism. Journal of Personality and Social Psychology, 65, 822-835. Norwich, B. (1987). Self-efficacy and mathematics achievement. Journal of Educational Psychology, 79, 384-387. Pajares, F., & Urdan, T.C. (1996). Exploratory factor analysis of the Mathematics Anxiety Scale. Measurement and Evaluation in Counseling and Development, 29, 35-47. Passer, M.W. (1983). Fear of failure, fear of evaluation, perceived competence and self-esteem in competence-trait-anxious children. Journal of Sport Psychology, 5, 172-188. Patrick, B.C., Skinner, E.A., & Connell, J.P. (1993). What motivates childrens behavior and emotion? Joint effects of perceived control and autonomy in the academic domain. Journal of Personality and Social Psychology, 65, 781-791. Patton, M.Q. (1987). How to use qualitative methods in evaluation. Newbury Park, CA: Sage. Patton, M. (1990). Qualitative evaluation and research methods. Newbury Park, CA: Sage. Peake, P.K., & Cervone, D. (1989). Sequence anchoring and self- efficacy: Primary effects in the consideration of possibilities. Social cognition, 7, 31-50. Perry, R.P., & Dickens, W.J. (1984). Perceived control in the college classroom: Response-outcome contingency training and instructor expressiveness effects on student achievement and causal attributions. Journal of Educational Psychology, 76, 966-981. Perry, R.P., Magnusson, J.L., Parsonson, K.L., & Dickens, W.J. (1986). Perceived control in the college classroom: Limitations in instructor expressiveness due to noncontingent feedback. Journal of Educational Psychology, 78, 96-107.

363

Ping, R.A. (1996a). Latent variable interaction and quadratic effect estimation: A two-step technique using structural equation analysis. Psychological Bulletin, 119, 166-175. Ping, R.A. (1996b). Latent variable regression: A technique for estimating interaction and quadratic coefficients. Multivariate Behavioral Research, 31, 95-120. Pintrich, P. R. (1988). A process-oriented view of student motivation and cognition. In J. Stark & L. Mets (Eds.), Improving teaching and learning through research: New directions for instructional research. San Francisco: Jossey-Bass. Pintrich, P.R., & DeGroot, E.V. (1990). Motivational and self-regulated learning components of classroom academic performance. Journal of Educational Psychology, 82, 33-40. Pintrich, P. R., & Garcia, T. (1991). Student goal orientation and self-regulation in the college classroom. In M. L. Maehr & P. R. Pintrich (Eds.), Advances in motivation and achievement. Greenwich, CN: JAI Press. Pintrich, P. R., Smith, D. A., Garcia, T., & McKeachie, W. J. (1991). A manual for the use of the Motivated Strategies for Learning Questionnaire (MLSQ). Ann Arbor, MI: University of Michigan. Plant, R.W., & Ryan, R.M. (1985). Intrinsic motivation and the effects of selfconsciousness, self-awareness, and ego-involvement: An investigation of internally controlling styles. Journal of Personality, 53, 435-449. Polak, R.L., & Prokop, C.K. (1989). Defensive pessimism: A protective selfmonitoring strategy? Journal of Social Behavior and Personality, 4, 285-289. Pruzinsky, T., & Borkovec, T.D. (1990). Cognitive and personality characteristics of worriers. Behavior Research and Therapy, 28, 507-512. Pyszczynski, T. (1982). Cognitive strategies for coping with uncertain outcomes. Journal of Research in Personality, 16, 386-399. Pyszczynski, T., & Greenberg, J. (1983). Determinants of reduction in intended effort as a strategy for coping with anticipated failure. Journal of Research in Personality, 17, 412-422. Qin, Z., Johnson, D.W., & Johnson, R.T. (1995). Cooperative versus competitive efforts and problem solving. Review of Educational Research, 65, 129-144. Rhodewalt, F. (1990). Self-handicappers: Individual differences in the preference for anticipatory, self-protective acts. In R.L. Higgins, C.R. Snyder., & S. Berglas (Eds.). Self-handicapping: The paradox that isnt. New York: Plenum Press.

364

Rhodewalt, F. (1994). Conceptions of ability, achievement goals, and individual differences in self-handicapping behavior: On the application of implicit theories. Journal of Personality, 62, 67-85. Rhodewalt, F., & Davison, J. (1986). Self-handicapping and subsequent performance: Role of outcome valance and attributional certainty. Basic and Applied Social Psychology, 7, 307-322. Rhodewalt, F., & Fairfield, M. (1991). Claimed self-handicaps and the selfhandicapper: The relation of reduction in intended effort to performance. Journal of Research in Personality, 25, 402-417. Rhodewalt, F., & Hill, S.K. (1995). Self-handicapping in the classroom: The effects of claimed self-handicaps on responses to academic failure. Basic and Applied Social Psychology, 16, 397-416. Rhodewalt, F., Morf, C., & Hazlett, S., & Fairfield, M. (1991). Selfhandicapping: The role of discounting and augmentation in the preservation of selfesteem. Journal of Personality and Social Psychology, 61, 122-131. Rhodewalt, F., Saltzman, A.T., & Wittmer, J. (1984). Self-handicapping among competitive athletes: The role of practice in self-esteem protection. Basic and Applied Social Psychology, 5, 197-209. Rhodewalt, F., Sanbonmatsu, D.M., Tschanz, B., Feick, D.L., & Waller, A. (1995). Self-handicapping and interpersonal tradeoffs: The effects of claimed selfhandicaps on observers performance evaluations and feedback. Personality and Social Psychology Bulletin, 21, 1042-1050. Richards, T., Richards, L., McGalliard, J., & Sharrock, B. (1993). NUDIST 2.3: Users Manual. Melbourne: Replee. Riggs, J.M. (1992). Self-handicapping and achievement. In A.K. Boggiano & T.S. Pittman (Eds.). Achievement and motivation: A social-developmental perspective. Cambridge University Press. Roeser, R.W., Midgley, C., & Urdan, T.C. (1996). Perceptions of the school psychological environment and early adolescents psychological and behavioral functioning in school: The mediating role of goals and belonging. Journal of Educational Psychology, 88, 408-422. Rosenberg, M. (1965). Society and the adolescent child. Princeton: Princeton University Press. Rosenberg, M. (1986). Self-concept from middle childhood through adolescence. In J. Suls & A.G. Greenwald (Eds.). Psychological perspectives on the self. Hillsdale, NJ: Erlbaum.

365

Roth, D.L., Snyder, C.R., & Pace, L.M. (1986). Dimensions of favorable selfpresentation. Journal of Personality and Social Psychology, 51, 867-874. Rowe, K.J., & Hill, P.W. (1997). Simultaneous estimation of multilevel structural equations to model students educational progress. Paper presented at 10th International Congress for School Effectiveness and Improvement, Memphis, Tennessee, USA. Rowe, K.J., Hill, P.W., & Holmes-Smith, P. (1995). Methodological issues in educational performance and school effectiveness research: A discussion with worked examples. Australian Journal of Education, 39, 217-248. Ryan, R.M. (1982). Control and information inb the intrapersonal sphere: An extension of cognitive evaluation theory. Journal of Personality and Social Psychology, 43, 450-461. Sanna, L.J., & Mark, M.M. (1995). Self-handicapping, expected evaluation, and performance: Accentuating the positive and attenuating the negative. Organizational Behavior and Human Decision Processes, 64, 84-102. Scheier, M.F., & Carver, C.S. (1985). The Self-Consciousness Scale: A revision for use with general populations. Journal of Applied Social Psychology, 15, 687-699. Scheier, M.F., Carver, C.S., & Bridges, M.N. (1994). Distinguishing optimism from neuroticism (and test anxiety, self-mastery, and self-esteem): A reevaluation of the Life Orientation Test. Journal of Personality and Social Psychology, 67, 10631078. Schill, T. (1990). A measure of self-defeating personality. Psychological Reports, 66, 1343-1346. Schill, T., Morales, J., Beyler, J., Tatter, T., & Swigert, L. (1991). Correlations between self-handicapping and self-defeating personality. Psychological Reports, 69, 655-657. Schlenker, B.R. (1975). Self-presentation: Managing the impression of consistency when reality interferes with self-enhancement. Journal of Personality and Social Psychology, 32, 1030-1037. Schlenker, B.R., & Leary, M.R. (1982). Social anxiety and self-presentation: A conceptualisation and model. Psychological Bulletin, 92, 641-669. Schmitz, B., & Skinner, E. (1993). Perceived control, effort, and academic performance: Interindividual, intraindividual, and multivariate time series analysis. Journal of Personality and Social Psychology, 64, 1010-1028.

366

Schneider, W.J., & Nevid, J.S. (1993). Overcoming math anxiety: A comparison of stress inoculation training and systematic desensitization. Journal of College Student Development, 34, 283-288. Schneider, D.J., & Turkat, D. (1975). Self-presentation following success or failure: Defensive self-esteem models. Journal of Personality, 43, 127-135. Schunk, D.H. (1981). Modeling and attributional effects on childrens achievement: A self-efficacy analysis. Journal of Educational Psychology, 73, 93-105. Schunk, D.H. (1983). Ability versus effort attributional feedback: Differential effects on self-efficacy and achievement. Journal of Educational Psychology, 75, 848856. Self, E.A. (1990). Situational influences on self-handicapping. In R.L. Higgins, C.R. Snyder., & S. Berglas (Eds.). Self-handicapping: The paradox that isnt. New York: Plenum Press. Shepperd, J.A., & Arkin, R.M. (1989a). Self-handicapping: The moderating roles of public self-consciousness and task importance. Personality and Social Psychology Bulletin, 15, 252-265. Shepperd, J.A., & Arkin, R.M. (1989b). Determinants of self-handicapping: Task importance and the effects of preexisting handicaps on self-generated handicaps. Personality and Social Psychology Bulletin, 15, 101-112.. Sherman, S.J., Skov, R.B., Hervitz, E.F., & Stock, C.B. (1981). The effects of explaining hypothetical future events: From possibility to probability to actuality and beyond. Journal of Experimental Social Psychology, 17, 142-158. Showers, C. (1988). The effects of how and why thinking on perceptions of future negative events. Cognitive Therapy and Research, 12, 225-240. Showers, C. (1992). The motivational and emotional consequences of considering positive or negative possibilities for an upcoming event. Journal of Personality and Social Psychology, 63, 474-484. Showers, C., & Ruben, C. (1990). Distinguishing defensive pessimism from depression: Negative expectations and positive coping mechanisms. Cognitive Therapy and Research, 14, 385-399. Skaalvik, E.M. (1990). Attribution of perceived academic results and relations with self-esteem in senior high school students. Scandinavian Journal of Educational Research, 34, 259-269. Skaalvik, E.M. (1997a). Goal orientation in mathematics: Relations with achievement, self-perceptions, motivation, and anxiety. Paper presented at the 7th European conference for Research on Learning and Instruction. Athens, Greece. 367

Skaalvik, E.M. (1997b). Self-enhancing and self-defeating ego orientation: Relations with task and avoidance orientation, achievement, self-perceptions, and anxiety. Journal of Educational Psychology, 89, 71-81. Skaalvik. E.M. & Rankin, R.J. (1990). Math, verbal, and general academic selfconcept: The internal/external frame of reference model and gender differences in selfconcept structure. Journal of Educational Psychology, 82, 546-554. Skaalvik, E.M., & Rankin, R.J. (1992). Math and verbal achievement and selfconcepts: Testing the internal/external frame of reference model. Journal of Early Adolescence, 12, 267-279. Skaalvik, E.M., & Rankin, R.J. (1995). A test of the internal/external frame of reference model at different levels of math and verbal self-perception. American Educational Research Journal, 32, 161-184. Skaalvik, E.M., Valans, H., & Sletta, O. (1994). Task involvement and ego involvement: Relations with academic achievement, academic self-concept and selfesteem. Scandinavian Journal of Educational Research, 38, 231-243. Skinner, E.A., Wellborn, J.G., & Connell, J.P. (1990). What it takes to do well in school and whether Ive got it: A process model of perceived control and childrens engagement and achievement in school. Journal of Educational Psychology, 82, 22-32. Sletta, O., Valas, H., Skaalvik, E., & Sobstad, F. (1996). Peer-relations, loneliness, and self-perceptions in school-age children. British Journal of Educational Psychology, 66, 431-445. Smith, T.W., Snyder, C.R., & Handelsman, M.M. (1982). On the self-serving function of an academic wooden leg: Test anxiety as a self-handicapping strategy. Journal of Personality and Social Psychology, 42, 314-321. Smith, T.W., Snyder, C.R., & Perkins, S.C. (1983). The self-serving function of hypochondriacal complaints: Physical symptoms as self-handicapping strategies. Journal of Personality and Social Psychology, 44, 787-797. Smith, D.S., & Strube, M.J. (1991). Self-protective tendencies as moderators of self-handicapping impressions. Basic and Applied Social Psychology, 12, 63-80. Snyder, C.R. (1990). Self-handicapping: Processes and sequelae - On the taking of a psychological dive. In R.L. Higgins, C.R. Snyder., & S. Berglas (Eds.). Selfhandicapping: The paradox that isnt. New York: Plenum Press. Snyder, C.R., & Smith, T.W. (1982). Symptoms as self-handicapping strategies: The virtues of old wine in a new bottle. In G. Weary & H.L. Mirels (Eds.). Integrations of clinical and social psychology. New York: Oxford University Press.

368

Snyder, C.R., Smith, T.W., Augelli, R.W., & Ingram, R.E. (1985). On the selfserving function of social anxiety: Shyness as a self-handicapping strategy. Journal of Personality and Social Psychology, 48, 970-980. Snyder, M.L., Smoller, B., Strenta, A., & Frankel, A. (1981). A comparison of egotism, negativity, and learned helplessness as explanations for poor performance after unsolvable problems. Journal of Personality and Social Psychology, 40, 24-30. Snyder, M., & Swann, W.B. (1978). Hypothesis testing processes in social interaction. Journal of Personality and Social Psychology, 36, 1202-1212. Solomon, L.J., & Rothbaum, E.D. (1984). Academic procrastination: Frequency and cognitive-behavioral correlates. Journal of Counseling Psychology, 31, 503-509. Stake, R.E., (1975). Evaluating the arts in education: A responsive approach. Columbus, OH: Charles E Merrill. Stipek, D., & Gralinski, J.H. (1996). Childrens beliefs about intelligence and school performance. Journal of Educational Psychology, 88, 397-407. Strauss, A., & Corbin, J. (1990). Basics of qualitative research: Grounded theory procedures and techniques. Newbury Park, CA: Sage. Strube, M.J. (1986). An analysis of the Self-Handicapping Scale. Basic and Applied Social Psychology, 7, 211-224. Swann, W.B. (1987). Identity negotiation: Where two roads meet. Journal of Personality and Social Psychology, 52, 1038-1051. Swann, W.B., Pellham, B.W., & Krull, D.S. (1989). Agreeable fancy or disagreeable truth? Reconciling self-enhancement and self-verification. Journal of Personality and Social Psychology, 57, 782-791. Swann, W.B., & Read, S.J. (1981). Self-verification processes: How we sustain our self-conceptions. Journal of Experimental Social Psychology, 17, 351-372. Swann, W.B., Stein-Seroussi, A., & Giesler, R.B. (1992). Why people selfverify. Journal of Personality and Social Psychology, 62, 392-401. Tabachnick, B.G., & Fidell, L.S. (1996). Using multivariate statistics. New York: Harper Collins. Teevan, R.C., & Smith, B.D. (1975). Relationships of fear-of-failure and need achievement motivation to a confirming-interval measure of aspirational levels. Psychological Reports, 36, 967-976.

369

Thompson, T. (1993). Characteristics of self-worth protection in achievement behaviour. British Journal of Educational Psychology, 63, 469-488. Thompson, T. (1994). Self-worth protection: Review and implications for the classroom. Educational Review, 46, 259-274. Tice, D.M. (1991). Esteem protection or enhancement? Self-handicapping motives and attribution differ by self-esteem. Journal of Personality and Social Psychology, 60, 711-725. Tice, D.M., & Baumeister, R.F. (1990). Self-esteem, self-handicapping, and selfpresentation: The strategy of inadequate practice. Journal of Personality, 58, 443-464. Treasure, D.C., & Roberts, G.C. (1994). Cognitive and affective concomitants of task and ego goal orientations during the middle school years. Journal of Sport and Exercise Psychology, 16, 15-28. Tseng, M.S., & Carter, A.R. (1970). Achievement motivation and fear of failure as determinants of vocational choice, vocational aspiration, and perception of vocational prestige. Journal of Counseling Psychology, 17, 150-156. Tucker, J.A., Vuchinich, R.E., & Sobell, M. (1981). Alcohol consumption as a self-handicapping strategy. Journal of Abnormal Psychology, 90, 220-230. Urdan, T.C., Midgley, C., & Anderman, E.M. (1998). The role of classroom goal structure in students use of self-handicapping strategies. American Educational Research Journal, 35, 101-122. Vallacher, R.R., & Solodky, M. (1979). Objective self-awareness, standards of evaluation, and moral behavior. Journal of Experimental Social Psychology, 15, 254262. Vance, W.R., & Watson, S.T. (1994). Comparing management training and systematic restructuring for reducing mathematics anxiety in college students. Journal of College Student Development, 35, 261-266. Vispoel, W. P.; & Austin, J. R. (1995). Success and failure in junior high school: A critical incident approach to understanding students attributional beliefs. American Educational Research Journal, 32, 377-412. Walling, M.P., & Duda, J.L. (1995). Goals and their associations with beliefs about success in and perceptions of the purposes of physical education. Journal of Teaching in Physical Education, 14, 140-156. Watkins, D., & Hattie, J. (1985). A longitudinal study of the approaches to learning of Australian tertiary students. Human Learning, 4, 127-141.

370

Weary, G., & Williams, J.P. (1990). Depressive self-presentation: Beyond selfhandicapping. Journal of Personality and Social Psychology, 58, 892-898. Weidner, G. (1980). Self-handicapping following learned helplessness treatment and the Type-A coronary-prone behavior pattern. Journal of Psychosomatic Research, 24, 319-325. Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92, 548-573. Weiner, B. (1994). Integrating social and personal theories of achievement striving. Review of Educational Research, 64, 557-573. Weiner, B., Frieze, L., Kukla, A., Reed, L., Rest, S., & Rosenbaum, R. (1971). Perceiving the causes of success and failure. In E.E. Jones., D.E. Kanouse., H.H. Kelley., R.E. Nisbett., S. Valin., & B. Weiner (Eds.). Attribution: Perceiving the causes of behavior. Morristown, NJ: General Learning Press. Weiner, B., & Kukla, A. (1970). An attributional analysis of achievement motivation. Journal of Personality and Social Psychology, 15, 1-20. Weinglert, L. & Rosen, A.S. (1995). Optimism, self-esteem, mood and subjective health. Personality & Individual Differences, 18, 653-661. Weinstein, N.D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39, 806-820. Young, F.W., & Harris, D.F. (1994). Multidimensional scaling. In M.J. Norusis (Ed.). SPSS Professional Statistics 6.1. Chicago, IL: SPSS Inc. Zackay, D. (1984). The influence of perceived events controllability on its subjective occurrence probability. Psychological Record, 34, 233-240. Zuckerman, M., Kieffer, S.C., & Knee, C.R. (1998). Consequences of selfhandicapping: Effects on coping, academic performance, and attainment. Journal of Personality and Social Psychology, 74, 1619-1628.

371

APPENDICES

372

APPENDIX A ACADEMIC PROCESS QUESTIONNAIRE - FIRST ADMINISTRATION

INFORMATION SHEET
(Please tear this page off to keep for your information)

Dear Student I would like to invite you to assist me in a study I am running as part of my PhD in Educational Psychology that looks at the goals and strategies university students use when engaging in their studies as well as their view of themselves as students. When I am finished I would like to combine all the answers together in order to (a) get a broad picture of how university students describe themselves, (b) see what strategies students use when going about their studies and (c) look at the effects of these strategies on such things as future plans and motivation. THIS IS NOT A TEST. There are no right or wrong answers and everyone will often have different answers from each other. Just make sure that your answers show how you really feel about yourself. If you would prefer not to answer the questionnaire, that is OK - just quietly leave the lecture room while the others are answering their questionnaire. If in the middle of the questionnaire you want to stop, that is also OK again, just leave the lecture room quietly. There will be no disadvantage to you if you would prefer not to participate in the survey. Please note that your participation in the survey is considered to be your consent to participate. Importantly, however, in accordance with Macarthur Ethics Review Committee (Human subjects) guidelines, if you are under 18 years of age, parental consent is required before you can participate. The parental consent form is at the back of the questionnaire. If you are under 18 years please take this form home for a parents signature and return the consent form with the completed questionnaire to my pigeon hole in the Faculty of Education. When you have finished the questionnaire, just hand it in to me. Your answers will be confidential and under no circumstances will the answers that you give to the questionnaire be known to anyone other than the researcher. All the questionnaires will be stored in a secured location that only I can access and answers entered into a computer will be accessed using a password known only to me. When answering the questionnaire, if you want to change an answer just cross it out and circle the answer that you prefer. You should have ONLY ONE answer for each question. Do not leave out any questions. If you are not sure which answer to circle, just circle the one that you think is the closest to what you think. QUESTIONS ARE ON BOTH SIDES OF THE PAGES. If you have any questions do not hesitate to ask me. Otherwise turn over the page and begin.

Thank you for your assistance

Andrew Martin (PhD Student. Ph: 047 360 104) Professor Herb Marsh (Supervisor. Ph: 02 9772 9633)
This research project has been approved by the University of Western Sydney Macarthur Ethics Review Committee (Human Subjects). Any complaints or reservations about this research may be directed to the Ethics Committee through the Executive Officer, phone (046) 203 641. All complaints will be treated in confidence.

373

ACADEMIC PROCESS QUESTIONNAIRE

1. SURNAME FIRST NAME

(Note: This is only to link this questionnaire with a similar one that I may give to you next year)

2. SEX (Please circle one)

FEMALE

MALE

3. AGE

years

4. What degree/diploma are you enrolled in? (eg. BEd Prim; BEd Early Child; Dip Ed Prim; Adult Ed)

5. Are you an Overseas Student? 1 2

YES Country NO

6. In total, how many semesters INCLUDING THIS ONE have you done at uni? (include semesters from other degrees) semesters

7. If applicable, what level of English did you do in your final school year (eg. 2 Unit)

8. If applicable, what level of Maths did you do in your final school year (eg. 2 Unit; Maths in Society)

9a. If you received a TER at the end of school, what was it? OR 9b. If you attended school before TERs, what was your HSC (or equivalent final) mark? out of (/100)

To get a clearer picture of students motivation and learning strategies, I am hoping to talk to a few students about the way they go about their studies. If you wouldnt mind talking to me about these things (at uni and at a time of your convenience) I would really appreciate it if you would leave a contact phone number. If I do contact you, and then you would like to change your mind, thats no problem. Contact Phone: ( )

374

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. Some of my bad grades have been a result of bad luck 2. When I do an assignment, I first get it clear in my head what I am trying to accomplish 3. I feel really successful when I work really hard 4. I often worry about making a good impression academically 5. I feel really successful when I can do better than my friends 6. Others thinking that I am clever is very important to how I feel about myself 7. I avoid thinking about possible bad outcomes before tests and assignments 8. I care about how others view me academically 9. I feel really successful when others get things wrong and I don't 10. I feel really successful when I know more than other people 11. I tell others that Im more run-down (havent been looking after myself) than I actually am when assignments or exams are due, so if I dont do as well as I had hoped, I can say that is the reason 12. When I get a bad grade Im not sure how Im going to avoid getting that grade again 13. Low marks cause me to question my academic ability 14. When I do well Im not sure what I can do to do well again 15. It is very important to how I feel about myself that I dont feel stupid 16. No matter how well I have done in the past, I tell others that I expect less good things to go my way than I really think will

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1 1

2 2 2

3 3 3

4 4 4

5 5 5

6 6 6

7 7 7

1 1 1

2 2 2

3 3 3

4 4 4

5 5 5

6 6 6

7 7 7

375

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. Any student could get smarter if they worked hard 2. When I dont do as well as expected, it is often due to a lack of effort on my part 3. My successes are due to my ability 4. When I do poorly it is because of my low ability 5. When I get a good grade Im not sure how Im going to get that grade again 6. Others thinking that I am proficient is very important to how I feel about myself 7. If I were asked to describe my capabilities in this area, my description might end up being different from one day to another day 8. It is very important to how I feel about myself that others dont think that Im incompetent 9. I feel really successful when I'm the smartest 10. A student who works really hard could be one of the smartest in the class 11. I get good grades mostly because the course material was easy to learn 12. I can overcome all obstacles in the path of academic success if I work hard enough 13. As a students knowledge increases, he or she becomes smarter 14. Further education with a lot of subjects in this area does not appeal to me 15. Im concerned how others view the way I go about my studies 16. Im often unsure how I can avoid future failure

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

376

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. Often I get poorer grades in courses because the teacher has failed to make them interesting 2. No matter how well I have done in the past, I let others think that I expect less things to go my way than I think will 3. When I finish working on tasks I check my work for errors 4. When my performance is not so good, Im often unsure about what I can do to improve it 5. Others thinking that I am skilful is very important to how I feel about myself 6. It is very important to how I feel about myself that others dont think that Im stupid 7. I'm usually aware of how other people view me academically 8. When I do poorly Im not sure what to do to minimise the chances of it happening again 9. Some low grades Ive received seem to be because some teachers are just stingy with marks 10. If I have trouble solving a homework problem, Im more likely to just copy the answer from the book than to keep working at it trying to figure it out 11. I feel that some of my good grades depend to a considerable extent on luck, such as having the right questions show up on an exam 12. If I have trouble understanding a problem or task, I keep going over it until I understand it 13. I feel really successful when I solve a problem by working hard 14. Some of my lower grades have seemed to be due to bad luck 15. I feel really successful when I do my very best

377

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. I care a lot about how I present myself academically 2. Feeling I am skilful is very important to how I feel about myself 3. My beliefs about how capable I am in this area seem to change very frequently 4. Often the main reason I do my uni work is because I do not want people to say bad things about me 5. Whenever I receive good grades, it is because I have studied hard 6. No matter how well I have done in the past, I generally have negative expectations about how I will do in the future 7. Considering what can go wrong in tests and assignments helps me to prepare 8. No matter how well I have done in the past, I tell people that Im more pessimistic about future tests and assignments than I really am 9. When I do well Im unsure as to how to repeat that success 10. In the future I would like to take on more of these subjects 11. It is very important to how I feel about myself that others dont think that Im incapable 12. When I do poorly it is because I havent worked hard 13. It is very important to how I feel about myself that I dont feel incapable 14. I let others think that my friends are keeping me from paying attention in class or during study time more than they actually are, so if I dont do as well as I had hoped, I can say that is the reason

378

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. I'm self-conscious about how I appear academically 2. No matter how well I have done in the past, I tell people that if something can go wrong for me it will, more than I think it actually will 3. When I try to study, I let others think that I am distracted by things more than I really am, so if I dont do as well as I had hoped, I can say that is the reason 4. Often the main reason I do my uni work is because I do not want to have to repeat my course 5. When I study I take note of the material I have or have not mastered 6. My successes are due to my hard work 7. Feeling I am clever is very important to how I feel about myself 8. I carefully consider all possible outcomes before tests and assignments 9. I tell others that I fool around the night before an exam more than I really do, so if I dont do as well as I had hoped, I can say that is the reason 10. Sometimes my success on exams depends on luck 11. Often the main reason I do my uni work is because I do not want people to think that Im not good enough to be at uni 12. Often the main reason I do my uni work is because I do not want to get bad marks 13. I learn quickly in these subjects 14. No matter how well I have done in the past, I tell others that my expectations for future performances are more negative than they actually are

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

379

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. Before taking an exam or quiz, I plan out how I will study the material 2. When I finish working on a task I check my work to see if it is reasonable 3. I feel that my good grades reflect directly on my academic ability 4. I feel really successful when something I learn makes me think about things 5. Students can learn new things but how smart they are doesnt change 6. Feeling I am smart is very important to how I feel about myself 7. I try to organise an approach in my mind before I actually start working on tasks 8. My hard work is the reason I do well 9. I feel really successful when I learn something interesting 10. I find that on one day I have one opinion of myself in this area and on another day I have a different opinion 11. Often the main reason I do my uni work is because I do not want to do poorly on tests and assignments 12. I'm concerned about what other people think of my academic capabilities 13. I often try to figure out how likely it is that I will do very well in tests and assignments 14. My ability is the reason I do well 15. Some students will never be smart, no matter how hard they try 16. Others thinking that I am capable is very important to how I feel about myself

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

380

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. It is very important to how I feel about myself that I dont feel dumb 2. Before doing an assignment, I clearly plan out how I am going to do it 3. Often the main reason I do my uni work is because I do not want my tutor or lecturer to think less of me 4. When I do better than I expect Im usually unclear about what to do to increase my chances of doing as well again 5. My opinion of myself tends to change a good deal instead of always remaining the same 6. When I do worse than I expect Im usually unsure about what to do to minimise the chances of it happening again 7. I spend a lot of time thinking through possible outcomes when a test or assignment is coming up 8. I have noticed that my ideas about myself in this area seem to change very quickly 9. I dont look forward to doing any more subjects in this area 10. There isnt much some students can do to make themselves smarter 11. I feel really successful when others can't do as well as me 12. How smart a student is doesnt change 13. No matter how well I have done in the past, in upcoming tests or assignments, I think if something can go wrong for me, it will 14. I often think about how I will feel if I do very poorly in tests and assignments 15. I'm concerned about how Im seen to be performing academically

381

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN THESE SUBJECTS . . .
1. When I read something in the book that doesnt make sense, I skip it without spending time to try to understand it 2. Im concerned about the way others see my academic abilities 3. No matter how well I have done in the past, I tell others that I expect Ill do more poorly in the future than I actually think Ill do 4. No matter how well I have done in the past, in academic situations I let others think that Im more negative about being able to do what is required than I really am 5. A smart student will always be smart 6. If I have trouble working at a problem or task, Im more likely to write the first thing that comes to mind than to keep working at it and try to figure it out 7. Not doing so well in a course is because I lacked skill in that area 8. Others thinking that I am competent is very important to how I feel about myself 9. Feeling I am capable is very important to how I feel about myself 10. As a student learns new things, he or she gets smarter 11. When I get good grades it is because of my academic competence 12. I let my friends keep me from paying attention in tutorials or from doing my study or assignments so I have an excuse if I dont do as well as I hoped 13. The good grades I receive are always the direct result of my efforts

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

382

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN THESE SUBJECTS . . .
1. I often fool around the night before a test or exam so I have an excuse if I dont do as well as I hoped 2. No matter how well I have done in the past, I hardly ever expect things to go my way in the future 3. I think about the academic image I project to others 4. When I work on a task, I make sure I know what I am asked to do before I begin 5. My poorer results are because of my low ability 6. If I have trouble working at a problem, Ill stick at it and try to work it out rather than get someone else to tell me how to do it 7. Compared to other students I am good in these subjects 8. I tell people that I put assignments and study off until the last moment more than I actually do, so if I dont do as well as I had hoped, I can say that is the reason 9. Some students wont be smart no matter what 10. Poor grades tell me that I havent worked hard enough 11. It is very important to how I feel about myself that others dont think that Im dumb 12. In my further education, I dont want to do any more subjects in this area 13. I dont mind doing subjects in this area in my further education 14. Some of the times that I have gotten a good grade was because of the teachers easy grading scheme

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

383

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. When I get poor grades it leads me to assume that I lacked ability to succeed in those courses 2. Im more interested in completing homework and assignments as fast as possible than spending time to do a good job 3. Often the main reason I do my uni work is because I do not want to fail 4. I let myself get run-down (dont look after myself) when assignments or exams are due so I have an excuse if I dont do as well as I hoped 5. I tend to not try hard at assignments so I have an excuse if I dont do as well as I hoped 6. No matter how well I have done in the past, in future performances, I rarely expect good things to go my way 7. I feel really successful when what I learn really makes sense 8. I look forward to teaching subjects in this area in the future 9. If some assessment is coming up at university, I am inclined to tell people that I am more depressed or down than I really am, so if I dont do as well as I had hoped, they will think that is the reason 10. When I receive poor results it is because I havent studied enough 11. I want to avoid subjects in this area in further education 12. When I dont do well Im unsure about what to do to avoid that happening again 13. Often the main reason I do my uni work is because I do not want to let my parents or friends down

384

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .
1. No matter how well I have done in the past, when going into academic situations, I let others think that I expect to do worse than I really think Ill do 2. Any student can get smarter by learning more 3. No matter how well I have done in the past, in upcoming performances, I expect more bad things to happen to me than good 4. When I do well it is because I have high ability 5. I feel really successful when I beat others 6. I let people think that I sometimes dont study as hard for exams as I actually do, so if I dont do as well as I had hoped, I can say that is the reason 7. I establish goals for learning 8. Work in these subjects is easy for me 9. I organise my study time 10. I feel really successful when something I learn makes me want to find out more 11. When assignments or exams are due, I am likely to let people think that Im in a worse mood than I really am, so if I dont do as well as I had hoped, they will think that is the reason 12. I often get involved in a lot of activities when exams or assignments are coming up so I have an excuse if I dont do as well as I hoped 13. Im often unsure how I can follow up on one success with another success 14. Feeling I am competent is very important to how I feel about myself 15. I feel really successful when I'm the only one who can answer questions

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1 1 1

2 2 2 2

3 3 3 3

4 4 4 4

5 5 5 5

6 6 6 6

7 7 7 7

1 1 1 1

2 2 2 2

3 3 3 3

4 4 4 4

5 5 5 5

6 6 6 6

7 7 7 7

385

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. I feel really successful when I get a new idea about how things work 2. It is very important to how I feel about myself that I dont feel incompetent 3. I'm hopeless when it comes to these subjects 4. Lack of effort is the reason I have not done well at times in the past 5. I let people think that I dont try as hard at assignments as I actually do, so if I dont do as well as I had hoped, I can say that is the reason 6. No matter how well I have done in the past, Im pessimistic about my future performances 7. I let people think that I am involved in a lot more activities than I really am when exams or assignments are due, so if I dont do as well as I had hoped, I can say that is the reason 8. Feeling I am proficient is very important to how I feel about myself 9. When I run into a difficult homework or assignment task, I keep working at it until I think Ive solved it 10. No matter how well I have done in the past, I often expect I will do more poorly in the future 11. When I receive a poor grade, I usually feel that the main reason is that I havent studied enough 12. I often think about other things, allow myself to be distracted, or daydream when I try to study so I have an excuse if I dont do as well as I hoped 13. When I do well it is because I have worked hard 14. When my performance is good Im not sure what I can do to maintain it

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

386

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN THESE SUBJECTS . . .
1. No matter how well I have done in the past, I often expect that in upcoming tests or assignments I wont be able to do what is required of me 2. When an exam or assignment is coming up, I am likely to tell others that Im feeling more hassled than I really am, so if I dont do as well as I had hoped, they will think that is the reason 3. Some of my good grades in courses may simply reflect that they were easier courses than most 4. Often the main reason I do my uni work is because I do not want people to think that I am dumb 5. I tend to not study very hard before exams so I have an excuse if I dont do as well as I hoped 6. No matter how well I have done in the past, I go into academic situations expecting to do worse 7. I tend to put assignments and study off until the last moment so I have an excuse if I dont do as well as I hoped 8. When I run into a difficult homework or assignment task, I usually give up quickly and go on to the next problem or task 9. When an exam or assignment is coming up, I am inclined to tell others that Im more anxious than I really am, so if I dont do as well as I had hoped, they will think that is the reason 10. Others thinking that I am smart is very important to how I feel about myself 11. I often think about how I will feel if I do very well in tests and assignments 12. Any student could improve their ability if he or she worked really hard 13. Sometimes I feel that I should count myself lucky for the good grades I get 14. I feel that nothing can change the opinion I currently hold about myself in this area

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

387

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .

Subjects that mainly involve maths and numbers

1 2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

1. The most important ingredient in getting good grades is my academic ability 2. I often try to figure out how likely it is that I will do very poorly in tests and assignments 3. My academic low points sometimes make me think I was just unlucky 4. When some assessment or exam is coming up, I am inclined to tell people that things arent going as well at home as they really are, so if I dont do as well as I had hoped, they will think that is the reason 5. I have always done well in these subjects 6. I change from a very good opinion of myself in this area to a very poor opinion 7. I get good marks in these subjects 8. In my experience once a teacher gets the idea youre a poor student, your work is much more likely to get poor a grade than if someone else handed it in 9. Students who are not smart cant do anything to change that 10. My low ability is the reason I have not done well at times in the past 11. I feel really successful when I have the highest test scores 12. I often become occupied on things other than study the night before an exam, so I have an excuse if I dont do as well as I hoped 13. I purposely dont get enough sleep or rest before upcoming exams and assignments, so I have an excuse if I dont do as well as I hoped

1 1

2 2

3 3

4 4

5 5

6 6

7 7

1 1

2 2

3 3

4 4

5 5

6 6

7 7

388

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN THESE SUBJECTS . . .
1. I dont attend all the classes in this area so I have an excuse if I dont do as well as I hoped 2. I let people think that I become more occupied on things other than study the night before an exam more than I actually do, so I can say that is the reason if I dont do as well as I hoped 3. I let people think that I dont get as much sleep or rest before upcoming exams and assignments than I actually do, so I can say that is the reason if I dont do as well as I hoped 4. I let people think that I attend less classes in this area than I really do so I have an excuse if I dont do as well as I hoped 1

Subjects that mainly involve maths and numbers

2 3 4 5 6 7 1

Subjects that mainly involve reading and writing

2 3 4 5 6 7

THANKYOU FOR YOUR PARTICIPATION

389

APPENDIX B ITEM WORDING AND SUBSCALE STATISTICS

Results in the main body of the tables are based on FINAL models having removed some items due to poor loadings (see Chapter 6 for further details). Items removed are indicated at the bottom of the table along with their associated statistics obtained in the FIRST one factor models

B1. SELF-HANDICAPPING Active self-handicapping Adapted from Midgley et al., (1996); Strube (1986)
Coefficient of determination .92; Cronbachs alpha .92 Chi square=61.97 DF=20 TLI=.94 RNI=.96 Item Item-total correlation
.77 .75 .73 .65 .66 .79 .77 .76 .64 .58 .70

Alpha with item removed

.91 .91 .91 .91 .92 .90 .91 .91 .92 .93 .92

1 Factor congeneric loading

.79 .77 .73 .70 .68 .81 .78 .79 .68 .59 .73

- I let myself get run-down (dont look after myself) when assignments or exams are due so I have an excuse if I dont do as well as I hoped - I tend to not try hard at assignments so I have an excuse if I dont do as well as I hoped - I often fool around the night before a test or exam so I have an excuse if I dont do as well as I hoped - I let my friends keep me from paying attention in tutorials or from doing my study or assignments so I have an excuse if I dont do as well as I hoped - I tend to not study very hard before exams so I have an excuse if I dont do as well as I hoped - I often become occupied on things other than study the night before an exam, so I have an excuse if I dont do as well as I hoped - I purposely dont get enough sleep or rest before upcoming exams and assignments, so I have an excuse if I dont do as well as I hoped - I dont attend all the classes in this area so I have an excuse if I dont do as well as I hoped *I tend to put assignments and study off until the last moment so I have an excuse if I dont do as well as I hoped *I often think about other things, allow myself to be distracted, or daydream when I try to study so I have an excuse if I dont do as well as I hoped *I often get involved in a lot of activities when exams or assignments are coming up so I have an excuse if I dont do as well as I hoped

* Item removed following first CFA

390

Self-presented self-handicapping
Coefficient of determination .92; Cronbachs alpha .90 Chi square=103 DF=20 TLI=.88 RNI=.92 Item Item-total correlation
.75 .62 .71 .68 .68 .79 .75 .79 .74 .61 .71

Alpha with item removed

.90 .90 .90 .90 .90 .89 .90 .90 .92 .93 .92

1 Factor congeneric loading

.75 .63 .71 .65 .69 .87 .82 .87 .77 .63 .74

- I tell others that Im more run-down (havent been looking after myself) than I actually am when assignments or exams are due, so if I dont do as well as I had hoped, I can say that is the reason - I let people think that I dont try as hard at assignments as I actually do, so if I dont do as well as I had hoped, I can say that is the reason - I tell others that I fool around the night before an exam more than I really do, so if I dont do as well as I had hoped, I can say that is the reason - I let others think that my friends are keeping me from paying attention in class or during study time more than they actually are, so if I dont do as well as I had hoped, I can say that is the reason - I let people think that I sometimes dont study as hard for exams as I actually do, so if I dont do as well as I had hoped, I can say that is the reason - I let people think that I become more occupied on things other than study the night before an exam more than I actually do, so I can say that is the reason if I dont do as well as I hoped - I let people think that I attend less classes in this area than I really do so I have an excuse if I dont do as well as I hoped - I let people think that I dont get as much sleep or rest before upcoming exams and assignments than I actually do, so I can say that is the reason if I dont do as well as I hoped *When I try to study, I let others think that I am distracted by things more than I really am, so if I dont do as well as I had hoped, I can say that is the reason *I let people think that I am involved in a lot more activities than I really am when exams or assignments are due, so if I dont do as well as I had hoped, I can say that is the reason *I tell people that I put assignments and study off until the last moment more than I actually do, so if I dont do as well as I had hoped, I can say that is the reason

* Item removed following first CFA

Self-presented affective self-handicapping

Coefficient of determination .89; Cronbachs alpha .88 Chi square=17.35 DF=5 TLI=.95 RNI=.98 Item Item-total correlation
.71 .70 .76 .72 .66

Alpha with item removed

.85 .86 .84 .85 .87

1 Factor congeneric loading

.77 .76 .83 .79 .72

- When an exam or assignment is coming up, I am inclined to tell others that Im more anxious than I really am, so if I dont do as well as I had hoped, they will think that is the reason - When an exam or assignment is coming up, I am likely to tell others that Im feeling more hassled than I really am, so if I dont do as well as I had hoped, they will think that is the reason - If some assessment is coming up at university, I am inclined to tell people that I am more depressed or down than I really am, so if I dont do as well as I had hoped, they will think that is the reason - When assignments or exams are due, I am likely to let people think that Im in a worse mood than I really am, so if I dont do as well as I had hoped, they will think that is the reason - When some assessment or exam is coming up, I am inclined to tell people that things arent going as well at home as they really are, so if I dont do as well as I had hoped, they will think that is the reason

391

B2. DEFENSIVE PESSIMISM Defensive expectations Adapted from Norem & Cantor (1986b); Norem (personal communication, December 12, 1996); Scheier, Carver, & Bridges (1994)
Coefficient of determination .92; Cronbachs alpha .90 Chi square=46.39 DF=20 TLI=.98 RNI=.99 Item Item-total correlation
.62 .80 .67 .61 .73 .56 .78 .75 .78

Alpha with item removed

.90 .88 .89 .90 .89 .90 .88 .88 .90

1 Factor congeneric loading

.63 .84 .74 .61 .80 .55 .85 .82 .83

- No matter how well I have done in the past, I go into academic situations expecting to do worse - No matter how well I have done in the past, I often expect I will do more poorly in the future - No matter how well I have done in the past, I generally have negative expectations about how I will do in the future - No matter how well I have done in the past, I often expect that in upcoming tests or assignments I wont be able to do what is required of me - No matter how well I have done in the past, in upcoming tests or assignments, I think if something can go wrong for me, it will - No matter how well I have done in the past, Im pessimistic about my future performances - No matter how well I have done in the past, I hardly ever expect things to go my way in the future - No matter how well I have done in the past, in future performances, I rarely expect good things to go my way *No matter how well I have done in the past, in upcoming performances, I expect more bad things to happen to me than good

* Item removed following first CFA

Self-presented defensive expectations

Coefficient of determination .93; Cronbachs alpha .92 Chi square=75.62 DF=20 TLI=.92 RNI=.95 Item Item-total correlation
.65 .82 .77 .79 .66 .76 .70 .75

Alpha with item removed

.92 .91 .91 .91 .92 .91 .92 .91

1 Factor congeneric loading

.65 .87 .80 .85 .69 .78 .73 .77

- No matter how well I have done in the past, when going into academic situations, I let others think that I expect to do worse than I really think Ill do - No matter how well I have done in the past, I tell others that I expect Ill do more poorly in the future than I actually think Ill do - No matter how well I have done in the past, I tell others that my expectations for future performances are more negative than they actually are - No matter how well I have done in the past, in academic situations I let others think that Im more negative about being able to do what is required than I really am - No matter how well I have done in the past, I tell people that if something can go wrong for me it will, more than I think it actually will - No matter how well I have done in the past, I tell people that Im more pessimistic about future tests and assignments than I really am - No matter how well I have done in the past, I let others think that I expect less things to go my way than I think will - No matter how well I have done in the past, I tell others that I expect less good things to go my way than I really think will

392

Reflectivity Norem (personal communication, December 12, 1996)

Coefficient of determination .79; Cronbachs alpha .78 Chi square=41.56 DF=14 TLI=.86 RNI=.91 Item Item-total correlation
.46 .55 .46 .51 .48 .61 .42 .15

Alpha with item removed

.75 .74 .76 .75 .75 .72 .76 .78

1 Factor congeneric loading

.52 .63 .52 .59 .58 .71 .44 .02

- I carefully consider all possible outcomes before tests and assignments - I often think about how I will feel if I do very poorly in tests and assignments - I often think about how I will feel if I do very well in tests and assignments - I often try to figure out how likely it is that I will do very poorly in tests and assignments - I spend a lot of time thinking through possible outcomes when a test or assignment is coming up - I often try to figure out how likely it is that I will do very well in tests and assignments - Considering what can go wrong in tests and assignments helps me to prepare *I avoid thinking about possible bad outcomes before tests and assignments

* Item removed following first CFA

B3. ESTEEM-RELEVANT COMPETENCE VALUATION Private Enhancing

Coefficient of determination .90; Cronbachs alpha .89 Chi square=45.04 DF=9 TLI=.91 RNI=.94

Item
- Feeling I am competent is very important to how I feel about myself - Feeling I am capable is very important to how I feel about myself - Feeling I am smart is very important to how I feel about myself - Feeling I am skilful is very important to how I feel about myself - Feeling I am proficient is very important to how I feel about myself - Feeling I am clever is very important to how I feel about myself

Item-total correlation
.74 .74 .76 .72 .58 .72

Alpha with item removed

.87 .87 .86 .87 .89 .87

1 Factor congeneric loading

.77 .79 .83 .79 .61 .76

Private Protective
Coefficient of determination .82; Cronbachs alpha .79 Chi square=20.98 DF=2 TLI=.77 RNI=.92 Item Item-total correlation
.47 .53 .72 .66

Alpha with item removed

.79 .76 .66 .70

1 Factor congeneric loading

.54 .65 .83 .74

- It is very important to how I feel about myself that I dont feel incompetent - It is very important to how I feel about myself that I dont feel stupid - It is very important to how I feel about myself that I dont feel dumb - It is very important to how I feel about myself that I dont feel incapable

393

Public Enhancing
Coefficient of determination .93; Cronbachs alpha .93 Chi square=18.45 DF=9 TLI=.98 RNI=.99 Item Item-total Alpha correlation with item removed
.78 .79 .78 .78 .83 .79 .92 .92 .92 .92 .91 .92

1 Factor congeneric loading

.82 .82 .82 .82 .85 .82

- Others thinking that I am competent is very important to how I feel about myself - Others thinking that I am capable is very important to how I feel about myself - Others thinking that I am smart is very important to how I feel about myself - Others thinking that I am skilful is very important to how I feel about myself - Others thinking that I am proficient is very important to how I feel about myself - Others thinking that I am clever is very important to how I feel about myself

Public Protective
Coefficient of determination .90; Cronbachs alpha .88 Chi square=6.42 DF=2 TLI=.97 RNI=.99 Item Item-total correlation
.73 .76 .79 .71

Alpha with item removed

.86 .85 .84 .87

1 Factor congeneric loading

.77 .84 .88 .76

- It is very important to how I feel about myself that others dont think that Im incompetent - It is very important to how I feel about myself that others dont think that Im stupid - It is very important to how I feel about myself that others dont think that Im dumb - It is very important to how I feel about myself that others dont think that Im incapable

B4. ATTRIBUTIONS Adapted from Lefcourt et al. (1979) Ability attributions for success
Coefficient of determination .78; Cronbachs alpha .77 Chi square=19.01 DF=9 TLI=.93 RNI=.96 Item Item-total correlation
.50 .41 .61 .49 .53 .53

Alpha with item removed

.74 .76 .71 .74 .73 .73

1 Factor congeneric loading

.57 .46 .73 .61 .61 .59

- The most important ingredient in getting good grades is my academic ability - I feel that my good grades reflect directly on my academic ability - When I get good grades it is because of my academic competence - When I do well it is because I have high ability - My successes are due to my ability - My ability is the reason I do well

394

Ability attributions for failure

Coefficient of determination .78; Cronbachs alpha .75 Chi square=12.45 DF=9 TLI=.98 RNI=.99 Item Item-total correlation
.39 .45 .53 .45 .63 .48

Alpha with item removed

.74 .72 .70 .72 .67 .71

1 Factor congeneric loading

.45 .52 .59 .58 .77 .55

- Low marks cause me to question my academic ability - Not doing so well in a course is because I lacked skill in that area - When I get poor grades it leads me to assume that I lacked ability to succeed in those courses - When I do poorly it is because of my low ability - My poorer results are because of my low ability - My low ability is the reason I have not done well at times in the past

Effort attributions for success

Coefficient of determination .84; Cronbachs alpha .78 Chi square=17.28 DF=9 TLI=.96 RNI=.97 Item Item-total correlation
.53 .60 .35 .49 .68 .57

Alpha with item removed

.75 .73 .79 .76 .72 .74

1 Factor congeneric loading

.58 .65 .37 .59 .82 .71

- The good grades I receive are always the direct result of my efforts - Whenever I receive good grades, it is because I have studied hard - I can overcome all obstacles in the path of academic success if I work hard enough - When I do well it is because I have worked hard - My successes are due to my hard work - My hard work is the reason I do well

Effort attributions for failure

Coefficient of determination .83; Cronbachs alpha .82 Chi square=24.82 DF=9 TLI=.93 RNI=.96 Item Item-total correlation
.61 .62 .61 .51 .65 .53

Alpha with item removed

.78 .78 .78 .81 .78 .81

1 Factor congeneric loading

.71 .69 .68 .56 .74 .61

- When I receive a poor grade, I usually feel that the main reason is that I havent studied enough - When I dont do as well as expected, it is often due to a lack of effort on my part - Poor grades tell me that I havent worked hard enough - When I do poorly it is because I havent worked hard - When I receive poor results it is because I havent studied enough - Lack of effort is the reason I have not done well at times in the past

395

External attributions for success

Coefficient of determination .75; Cronbachs alpha .73 Chi square=45.98 DF=9 TLI=.72 RNI=.83

Item

Item-total correlation
.54 .48 .44 .38 .47 .48

Alpha with item removed

.67 .69 .70 .72 .69 .69

1 Factor congeneric loading

.73 .64 .53 .39 .50 .48

- Sometimes my success on exams depends on luck - I feel that some of my good grades depend to a considerable extent on luck, such as having the right questions show up on an exam - Sometimes I feel that I should count myself lucky for the good grades I get - Some of the times that I have gotten a good grade was because of the teachers easy grading scheme - Some of my good grades in courses may simply reflect that they were easier courses than most - I get good grades mostly because the course material was easy to learn

External attributions for failure

Coefficient of determination .67 Cronbachs alpha .66 Chi square=27.9 DF=9 TLI=.78 RNI=.87 Item Item-total correlation
.45 .28 .48 .36 .34 .41

Alpha with item removed

.59 .65 .58 .62 .63 .60

1 Factor congeneric loading

.73 .64 .53 .39 .50 .48

- Some of my lower grades have seemed to be due to bad luck - My academic low points sometimes make me think I was just unlucky - Some of my bad grades have been a result of bad luck - In my experience once a teacher gets the idea youre a poor student, your work is much more likely to receive poor grades than if someone else handed it in - Often I get poorer grades in courses because the teacher has failed to make them interesting - Some low grades Ive received seem to be because some teachers are just stingy with marks

B5. PUBLIC SELF-CONSCIOUSNESS Adapted from Scheier & Carver (1985)

Coefficient of determination .93; Cronbachs alpha .92 Chi square=9.7 DF=9 TLI=.99 RNI=.99 Item Item-total correlation
.75 .83 .76 .73 .85 .72 .58 .52 .64 .67

Alpha with item removed

.91 .90 .91 .91 .89 .91 .92 .93 .92 .92

1 Factor congeneric loading

.78 .87 .80 .75 .91 .74 .61 .54 .68 .66

- I'm self-conscious about how I appear academically - I'm concerned about what other people think of my academic capabilities - I'm concerned about how Im seen to be performing academically - I care about how others view me academically - Im concerned about the way others see my academic abilities - I think about the academic image I project to others *Im concerned how others view the way I go about my studies *I'm usually aware of how other people view me academically *I care a lot about how I present myself academically *I often worry about making a good impression academically

* Item removed following first CFA

396

B6. MOTIVATION ORIENTATION Adapted from Duda & Nicholls (1992); Harter et al. (1992); Nicholls (1989)

Ego-orientation
Coefficient of determination .91; Cronbachs alpha .90 Chi square=38.43 DF=9 TLI=.93 RNI=.96 Item Item-total correlation
.75 .75 .73 .67 .75 .75 .60 .64

Alpha with item removed

.88 .88 .89 .89 .88 .88 .91 .90

1 Factor congeneric loading

.81 .81 .74 .71 .79 .81 .61 .68

- I feel really successful when I know more than other people - I feel really successful when others get things wrong and I don't - I feel really successful when I'm the smartest - I feel really successful when I beat others - I feel really successful when I do better than my friends - I feel really successful when others can't do as well as me *I feel really successful when I have the highest test scores *I feel really successful when I'm the only one who can answer questions

* Item removed following first CFA

Task-orientation
Coefficient of determination .86 Cronbachs alpha .86 Chi square=22.78 DF=14 TLI=.97 RNI=.98 Item Item-total correlation
.69 .59 .60 .65 .60 .57 .67 .40

Alpha with item removed

.83 .84 .84 .83 .84 .84 .83 .86

1 Factor congeneric loading

.73 .64 .65 .72 .66 .62 .74 .43

- I feel really successful when what I learn really makes sense - I feel really successful when I solve a problem by working hard - I feel really successful when something I learn makes me want to find out more - I feel really successful when something I learn makes me think about things - I feel really successful when I get a new idea about how things work - I feel really successful when I do my very best - I feel really successful when I learn something interesting *I feel really successful when I work really hard

* Item removed following first CFA

397

Avoidance-oriented performance
Coefficient of determination .87; Cronbachs alpha .84 Chi square=39.21 DF=14 TLI=.92 RNI=.95 Item Item-total correlation
.51 .49 .75 .59 .66 .64 .56 .55 .55

Alpha with item removed

.84 .84 .80 .82 .81 .82 .83 .85 .85

1 Factor congeneric loading

.58 .50 .85 .61 .76 .70 .60 .55 .55

- Often the main reason I do my uni work is because I do not want people to think that I am dumb - Often the main reason I do my uni work is because I do not want to get bad marks - Often the main reason I do my uni work is because I do not want people to think that Im not good enough to be at uni - Often the main reason I do my uni work is because I do not want to have to repeat my course - Often the main reason I do my uni work is because I do not want people to say bad things about me - Often the main reason I do my uni work is because I do not want to let my parents or friends down - Often the main reason I do my uni work is because I do not want my tutor or lecturer to think less of me *Often the main reason I do my uni work is because I do not want to do poorly on tests and assignments *Often the main reason I do my uni work is because I do not want to fail

* Item removed following first CFA

B7. VIEWS OF INTELLIGENCE Adapted from Stipek & Gralinski (1996) Entity beliefs
Coefficient of determination .87; Cronbachs alpha .85 Chi square=52.37 DF=9 TLI=.85 RNI=.91 Item Item-total correlation
.67 .74 .65 .51 .58 .67 .49

Alpha with item removed

.82 .81 .82 .85 .83 .82 .85

1 Factor congeneric loading

.75 .81 .75 .55 .62 .72 .54

- Some students will never be smart, no matter how hard they try - There isnt much some students can do to make themselves smarter - Some students wont be smart no matter what - Students who are not smart cant do anything to change that - Students can learn new things but how smart they are doesnt change - How smart a student is doesnt change *A smart student will always be smart

* Item removed following first CFA

Incremental beliefs
Coefficient of determination .80; Cronbachs alpha .79 Chi square=18.3 DF=9 TLI=.95 RNI=.97 Item Item-total correlation
.51 .52 .59 .57 .47 .60

Alpha with item removed

.76 .76 .74 .75 .77 .74

1 Factor congeneric loading

.61 .55 .68 .67 .50 .70

- A student who works really hard could be one of the smartest in the class - Any student could get smarter if they worked hard - Any student can get smarter by learning more - As a student learns new things he or she gets smarter - Any student could improve their ability if he or she worked really hard - As a students knowledge increases, he or she becomes smarter

398

B8. PERCEPTIONS OF FUTURE CONTROL Adapted from Connell (1985)

Low control over future success

Coefficient of determination .91; Cronbachs alpha .90 Chi square=20.43 DF=9 TLI=.97 RNI=.98 Item Item-total correlation
.80 .72 .77 .75 .73 .62

Alpha with item removed

.87 .89 .88 .88 .88 .90

1 Factor congeneric loading

.83 .76 .83 .80 .75 .64

- When I do well Im unsure as to how to repeat that success - When I get a good grade Im not sure how Im going to get that grade again - When I do well Im not sure what I can do to do well again - When I do better than I expect Im usually unclear about what to do to increase my chances of doing as well again - When my performance is good Im not sure what I can do to maintain it - Im often unsure how I can follow up on one success with another success

Low control over future failure

Coefficient of determination .89; Cronbachs alpha .88 Chi square=44.53 DF=9 TLI=.90 RNI=.94 Item Item-total correlation
.68 .68 .74 .64 .72 .68

Alpha with item removed

.86 .86 .85 .87 .85 .86

1 Factor congeneric loading

.73 .72 .81 .68 .79 .75

- When I dont do well Im unsure about what to do to avoid that happening again - When I get a bad grade Im not sure how Im going to avoid getting that grade again - When I do poorly Im not sure what to do to minimise the chances of it happening again - Im often unsure how I can avoid future failure - When I do worse than I expect Im usually unsure about what to do to minimise the chances of it happening again - When my performance is not so good, Im often unsure about what I can do to improve it

B9. LEVEL OF SELF-CONCEPT Adapted from Marsh (1988a)

Coefficient of determination .90; Cronbachs alpha .89 Chi square=39.64 DF=9 TLI=.92 RNI=.95 Item Item-total correlation
.73 .68 .71 .71 .80 .65

Alpha with item removed

.87 .88 .87 .87 .86 .88

1 Factor congeneric loading

.76 .75 .77 .75 .85 .67

- I learn quickly in these subjects - I have always done well in these subjects - Compared to other students I am good in these subjects - Work in these subjects is easy for me - I get good marks in these subjects - I'm hopeless when it comes to these subjects

399

B10. STABILITY OF SELF-CONCEPT Adapted from Campbell et al. (1996); Rosenberg (1965)
Coefficient of determination .82; Cronbachs alpha .85 Chi square=9.52 DF=9 TLI=.99 RNI=.99 Item Item-total correlation
.45 .66 .47 .65 .56 .53 .16

Alpha with item removed

.79 .74 .78 .74 .76 .77 .84

1 Factor congeneric loading

.52 .75 .51 .73 .64 .61 .15

- My opinion of myself tends to change a good deal instead of always remaining the same - I find that on one day I have one opinion of myself in this area and on another day I have a different opinion - I change from a very good opinion of myself in this area to a very poor opinion - I have noticed that my ideas about myself in this area seem to change very quickly - My beliefs about how capable I am in this area seem to change very frequently - If I were asked to describe my capabilities in this area, my description might end up being different from one day to another day *I feel that nothing can change the opinion I currently hold about myself in this area

* Item removed following first CFA

B11. SELF-REGULATION Adapted from Miller et al. (1996)

Coefficient of determination .84; Cronbachs alpha .83 Chi square=53.76 DF=14 TLI=.86 RNI=.90 Item Item-total correlation
.59 .53 .55 .47 .62 .59 .54 .66 .42 .45

Alpha with item removed

.81 .82 .82 .82 .81 .81 .82 .80 .84 .84

1 Factor congeneric loading

.65 .60 .61 .52 .67 .65 .59 .72 .50 .49

- Before taking an exam or quiz, I plan out how I will study the material I organise my study time - When I do an assignment, I first get it clear in my head what I am trying to accomplish - When I work on a task, I make sure I know what I am asked to do before I begin - When I finish working on a task I check my work to see if it is reasonable - I try to organise an approach in my mind before I actually start working on tasks - When I finish working on tasks I check my work for errors - Before doing an assignment, I clearly plan out how I am going to do it * I establish goals for learning *When I study I take note of the material I have or have not mastered

* Item removed following first CFA

400

B12. PERSISTENCE Adapted from Miller et al., (1996)

Coefficient of determination .77; Cronbachs alpha .75 Chi square=39.2 DF=14 TLI=.86 RNI=.91 Item Item-total correlation
.47 .42 .56 .38 .52 .50 .45 .34

Alpha with item removed

.72 .74 .70 .74 .71 .72 .73 .76

1 Factor congeneric loading

.58 .44 .68 .45 .59 .61 .52 .39

- If I have trouble understanding a problem or task, I keep going over it until I understand it - Im more interested in completing homework and assignments as fast as possible than spending time to do a good job (reversed) - If I have trouble solving a homework problem, Im more likely to just copy the answer from the book than to keep working at it trying to figure it out (reversed) - If I have trouble working at a problem, Ill stick at it and try to work it out rather than get someone else to tell me how to do it - When I read something in the book that doesnt make sense, I skip it without spending time to try to understand it - When I run into a difficult homework or assignment task, I keep working at it until I think Ive solved it - When I run into a difficult homework or assignment task, I usually give up quickly and go on to the next problem or task (reversed) *If I have trouble working at a problem or task, Im more likely to write the first thing that comes to mind than to keep working at it and try to figure it out (reversed)

* Item removed following first CFA

B13. FUTURE PLANS Adapted from Skaalvik & Rankin (1995)

Coefficient of determination .91; Cronbachs alpha .91 Chi square=39.12 DF=14 TLI=.95 RNI=.97 Item Item-total correlation
.77 .72 .78 .70 .66 .69 .76

Alpha with item removed

.89 .89 .89 .90 .90 .90 .89

1 Factor congeneric loading

.80 .78 .82 .73 .67 .73 .81

- I dont look forward to doing any more subjects in this area - In my further education, I dont want to do any more subjects in this area - I dont mind doing subjects in this area in my further education - I look forward to teaching subjects in this area in the future - Further education with a lot of subjects in this area does not appeal to me - In the future I would like to take on more of these subjects - I want to avoid subjects in this area in further education

401

APPENDIX C

29 factor CFA loadings (Nepean, n=204) Chi square=5687.32, df=3421; TLI=.81, RNI=.84, RMSEA=.057 Item parcel
Self-handicapping 1 Self-handicapping 2 Self-handicapping 3 Self-handicapping 4 Self-pres self-handicapping 1 Self-pres self-handicapping 2 Self-pres self-handicapping 3 Self-pres self-handicapping 4 Affective self-handicapping 1 Affective self-handicapping 2 Affective self-handicapping 3 Defensive expectation 1 Defensive expectation 2 Defensive expectation 3 Defensive expectation 4 Self-pres defensive expect 1 Self-pres defensive expect 2 Self-pres defensive expect 3 Self-pres defensive expect 4 Reflectivity 1 Reflectivity 2 Reflectivity 3 Private enhance comp val 1 Private enhance comp val 2 Private enhance comp val 3 Public enhance comp val 1 Public enhance comp val 2 Public enhance comp val 3 Private protect comp val 1 Private protect comp val 2 Public protect comp val 1 Public protect comp val 2 Ability - success 1 Ability - success 2 Ability - success 3 Ability - failure 1 Ability - failure 2 Ability - failure 3 Effort - failure 1 Effort - failure 2 Effort - failure 3 Effort - success 1 Effort - success 2 Effort - success 3

Factor loading
.78 .86 .85 .81 .80 .81 .88 .80 .92 .79 .72 .87 .88 .80 .69 .79 .85 .90 .88 .55 .73 .78 .87 .88 .74 .89 .89 .88 .80 .74 .90 .90 .67 .80 .71 .72 .69 .71 .65 .84 .77 .66 .73 .74

Item parcel
External - success 1 External - success 2 External - success 3 External - failure 1 External - failure 2 External - failure 3 Public self-consciousness 1 Public self-consciousness 2 Public self-consciousness 3 Ego-orientation 1 Ego-orientation 2 Ego-orientation 3 Task-orientation 1 Task-orientation 2 Task-orientation 3 Avoidance orientation 1 Avoidance orientation 2 Avoidance orientation 3 Entity view 1 Entity view 2 Entity view 3 Incremental view 1 Incremental view 2 Incremental view 3 Control future fail 1 Control future fail 2 Control future fail 3 Control future success 1 Control future success 2 Control future success 3 Self-concept 1 Self-concept 2 Self-concept 3 Self-concept instability 1 Self-concept instability 2 Self-concept instability 3 Self-regulation 1 Self-regulation 2 Self-regulation 3 Persistence 1 Persistence 2 Persistence 3 Future plans 1 Future plans 2 Future plans 3

Factor loading
.53 .76 .69 .43 .65 .65 .87 .90 .90 .85 .91 .86 .77 .74 .82 .83 .84 .76 .78 .86 .83 .74 .74 .75 .77 .77 .87 .88 .90 .78 .87 .86 .79 .74 .75 .73 .71 .78 .76 .73 .60 .76 .77 .87 .90

402

APPENDIX D ACADEMIC PROCESS QUESTIONNAIRE (REVISED)

Dear Student I would like to invite you to assist me in a study I am running as part of my PhD in Educational Psychology that looks at the goals and strategies university students use when engaging in their studies as well as their view of themselves as students. When I am finished I would like to combine all the answers together in order to (a) get a broad picture of how university students describe themselves, (b) see what strategies students use when going about their studies and (c) look at the effects of these strategies on such things as future plans and motivation. THIS IS NOT A TEST. There are no right or wrong answers and everyone will often have different answers from each other. Just make sure that your answers show how you really feel about yourself. If you would prefer not to answer the questionnaire, that is OK - just quietly go on with some work while the others are answering their questionnaire. If in the middle of the questionnaire you want to stop, that is also OK again, just go on with some work. There will be no disadvantage to you if you would prefer not to participate in the survey. Please note that your participation in the survey is considered to be your consent to participate. Importantly, however, in accordance with Macarthur Ethics Review Committee (Human subjects) guidelines, if you are under 18 years of age, parental consent is required before you can participate. The parental consent form is at the back of the questionnaire. If you are under 18 years please take this form home for a parents signature and return the consent form with the completed questionnaire to me in the Faculty of Education. Your answers will be confidential and under no circumstances will the answers that you give to the questionnaire be known to anyone other than the researcher. All the questionnaires will be stored in a secured location that only I can access and answers entered into a computer will be accessed using a password known only to me. When answering the questionnaire, if you want to change an answer just cross it out and circle the answer that you prefer. You should have ONLY ONE answer for each question. Do not leave out any questions. If you are not sure which answer to circle, just circle the one that you think is the closest to what you think. QUESTIONS ARE ON BOTH SIDES OF THE PAGES. Thank you for your assistance Andrew Martin (PhD Student. Ph: 047 360 104) and Professor Herb Marsh (Supervisor. Ph: 02 9772 9633) This research project has been approved by the University of Western Sydney Macarthur Ethics Review Committee (Human Subjects). Any complaints or reservations about this research may be directed to the Ethics Committee through the Executive Officer, phone (046) 203 641. All complaints will be treated in confidence. STUDENT ID NUMBER (Note: This is only to link this questionnaire with a similar one that I may give to you next year) SEX 1 FEMALE 2 MALE AGE years

What degree/diploma are you enrolled in? Are you an Overseas Student? 1 YES Country 2 NO semesters

In total, how many semesters INCLUDING THIS ONE have you done at uni? If applicable, what level of Maths did you do in your final school year If you received a TER at the end of school, what was it? (/100) OR If you attended school before TERs, what was your FINAL year mark?

out of

403

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN SUBJECTS THAT MAINLY INVOLVE MATHS AND NUMBERS . .

Some of my bad grades have been a result of bad luck If I were asked to describe my capabilities in this area, my description might end up being different from one day to another day I feel really successful when I'm the smartest A student who works really hard could be one of the smartest in the class I get good grades mostly because the course material was easy to learn As a students knowledge increases, he or she becomes smarter Further education with a lot of subjects in this area does not appeal to me Im often unsure how I can avoid future failure Often I get poorer grades in courses because the teacher has failed to make them interesting

When I do an assignment, I first get it clear in my head what I am trying to accomplish I feel really successful when I can do better than my friends Others thinking that I am clever is very important to how I feel about myself I care about how others view me academically

I feel really successful when others get things wrong and I don't I feel really successful when I know more than other people I tell others that Im more run-down (havent been looking after myself) than I actually am when assignments or exams are due, so if I dont do as well as I had hoped, I can say that is the reason When I get a bad grade Im not sure how Im going to avoid getting that grade again

No matter how well I have done in the past, I let others think that I expect less things to go my way than I think will When I finish working on tasks I check my work for errors When my performance is not so good, Im often unsure about what I can do to improve it Others thinking that I am skilful is very important to how I feel about myself When I do poorly Im not sure what to do to minimise the chances of it happening again Some low grades Ive received seem to be because some teachers are just stingy with marks

Low marks cause me to question my academic ability No matter how well I have done in the past, I tell others that I expect less good things to go my way than I really think will Any student could get smarter if they worked hard

When I do poorly it is because of my low ability

Others thinking that I am proficient is very important to how I feel about myself

404

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN SUBJECTS THAT MAINLY INVOLVE MATHS AND NUMBERS . .

If I have trouble solving a homework problem, Im more likely to just copy the answer from the book than to keep working at it trying to figure it out I feel that some of my good grades depend to a considerable extent on luck, such as having the right questions show up on an exam I'm self-conscious about how I appear academically

No matter how well I have done in the past, I tell people that if something can go wrong for me it will, more than I think it actually will Often the main reason I do my uni work is because I do not want to have to repeat my course Feeling I am clever is very important to how I feel about myself I carefully consider all possible outcomes before tests and assignments I tell others that I fool around the night before an exam more than I really do, so if I dont do as well as I had hoped, I can say that is the reason Sometimes my success on exams depends on luck Often the main reason I do my uni work is because I do not want people to think that Im not good enough to be at uni Often the main reason I do my uni work is because I do not want to get bad marks

If I have trouble understanding a problem or task, I keep going over it until I understand it I feel really successful when I solve a problem by working hard Some of my lower grades have seemed to be due to bad luck I feel really successful when I do my very best

Feeling I am skilful is very important to how I feel about myself My beliefs about how capable I am in this area seem to change very frequently

Often the main reason I do my uni work is because I do not want people to say bad things about me No matter how well I have done in the past, I generally have negative expectations about how I will do in the future Considering what can go wrong in tests and assignments helps me to prepare

I learn quickly in these subjects

No matter how well I have done in the past, I tell people that Im more pessimistic about future tests and assignments than I really am In the future I would like to take on more of these subjects I let others think that my friends are keeping me from paying attention in class or during study time more than they actually are, so if I dont do as well as I had hoped, I can say that is the reason

No matter how well I have done in the past, I tell others that my expectations for future performances are more negative than they actually are Before taking an exam or quiz, I plan out how I will study the material

When I finish working on a task I check my work to see if it is reasonable I feel really successful when something I learn makes me think about things

405

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN SUBJECTS THAT MAINLY INVOLVE MATHS AND NUMBERS . .

Students can learn new things but how smart they are doesnt change Feeling I am smart is very important to how I feel about myself I try to organise an approach in my mind before I actually start working on tasks I feel really successful when I learn something interesting I find that on one day I have one opinion of myself in this area and on another day I have a different opinion I'm concerned about what other people think of my academic capabilities I have noticed that my ideas about myself in this area seem to change very quickly I dont look forward to doing any more subjects in this area There isnt much some students can do to make themselves smarter I feel really successful when others can't do as well as me How smart a student is doesnt change

No matter how well I have done in the past, in upcoming tests or assignments, I think if something can go wrong for me, it will I often think about how I will feel if I do very poorly in tests and assignments I'm concerned about how Im seen to be performing academically When I read something in the book that doesnt make sense, I skip it without spending time to try to understand it Im concerned about the way others see my academic abilities No matter how well I have done in the past, I tell others that I expect Ill do more poorly in the future than I actually think Ill do No matter how well I have done in the past, in academic situations I let others think that Im more negative about being able to do what is required than I really am Not doing so well in a course is because I lacked skill in that area Others thinking that I am competent is very important to how I feel about myself

I often try to figure out how likely it is that I will do very well in tests and assignments Some students will never be smart, no matter how hard they try Others thinking that I am capable is very important to how I feel about myself

Before doing an assignment, I clearly plan out how I am going to do it Often the main reason I do my uni work is because I do not want my tutor or lecturer to think less of me My opinion of myself tends to change a good deal instead of always remaining the same

When I do worse than I expect Im usually unsure about what to do to minimise the chances of it happening again I spend a lot of time thinking through possible outcomes when a test or assignment is coming up

406

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN SUBJECTS THAT MAINLY INVOLVE MATHS AND NUMBERS . .

Feeling I am capable is very important to how I feel about myself When I get poor grades it leads me to assume that I lacked ability to succeed in those courses Im more interested in completing homework and assignments as fast as possible than spending time to do a good job I let myself get run-down (dont look after myself) when assignments or exams are due so I have an excuse if I dont do as well as I hoped I tend to not try hard at assignments so I have an excuse if I dont do as well as I hoped No matter how well I have done in the past, in future performances, I rarely expect good things to go my way I feel really successful when what I learn really makes sense I look forward to teaching subjects in this area in the future I want to avoid subjects in this area in further education When I dont do well Im unsure about what to do to avoid that happening again

As a student learns new things, he or she gets smarter

I let my friends keep me from paying attention in tutorials or from doing my study or assignments so I have an excuse if I dont do as well as I hoped I often fool around the night before a test or exam so I have an excuse if I dont do as well as I hoped No matter how well I have done in the past, I hardly ever expect things to go my way in the future I think about the academic image I project to others When I work on a task, I make sure I know what I am asked to do before I begin My poorer results are because of my low ability

If I have trouble working at a problem, Ill stick at it and try to work it out rather than get someone else to tell me how to do it Compared to other students I am good in these subjects Some students wont be smart no matter what

In my further education, I dont want to do any more subjects in this area I dont mind doing subjects in this area in my further education Some of the times that I have gotten a good grade was because of the teachers easy grading scheme

Often the main reason I do my uni work is because I do not want to let my parents or friends down No matter how well I have done in the past, when going into academic situations, I let others think that I expect to do worse than I really think Ill do Any student can get smarter by learning more I feel really successful when I beat others I let people think that I sometimes dont study as hard for exams as I actually do, so if I dont do as well as I had hoped, I can say that is the reason

407

Disagree Strongly 1

Disagree 2

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree 6

Agree Strongly 7

IN SUBJECTS THAT MAINLY INVOLVE MATHS AND NUMBERS . .

Work in these subjects is easy for me No matter how well I have done in the past, I often expect that in upcoming tests or assignments I wont be able to do what is required of me Some of my good grades in courses may simply reflect that they were easier courses than most Often the main reason I do my uni work is because I do not want people to think that I am dumb I tend to not study very hard before exams so I have an excuse if I dont do as well as I hoped No matter how well I have done in the past, I go into academic situations expecting to do worse When I run into a difficult homework or assignment task, I usually give up quickly and go on to the next problem or task Others thinking that I am smart is very important to how I feel about myself I often think about how I will feel if I do very well in tests and assignments Any student could improve their ability if he or she worked really hard Sometimes I feel that I should count myself lucky for the good grades I get

I organise my study time

I feel really successful when something I learn makes me want to find out more

Feeling I am competent is very important to how I feel about myself

I often become occupied on things other than study the night before an exam, so I have an excuse if I dont do as well as I hoped I purposely dont get enough sleep or rest before upcoming exams or assignments, so I have an excuse if I dont do as well as I hoped

I dont attend all the classes in these subjects so I have an excuse if I dont do as well as I hoped I feel really successful when I get a new idea about how things work I'm hopeless when it comes to these subjects

I let people think that I dont try as hard at assignments as I actually do, so if I dont do as well as I had hoped, I can say that is the reason No matter how well I have done in the past, Im pessimistic about my future performances Feeling I am proficient is very important to how I feel about myself When I run into a difficult homework or assignment task, I keep working at it until I think Ive solved it No matter how well I have done in the past, I often expect I will do more poorly in the future

I often try to figure out how likely it is that I will do very poorly in tests and assignments My academic low points sometimes make me think I was just unlucky I have always done well in these subjects I change from a very good opinion of myself in this area to a very poor opinion

408

Disagree Strongly 1

Disagree

Disagree Somewhat 3

Neither Disagree nor Agree 4

Agree Somewhat 5

Agree

Agree Strongly 7

IN SUBJECTS THAT MAINLY INVOLVE MATHS AND NUMBERS . .

I get good marks in these subjects I let people think that I become more occupied on things other than study the night before an exam more than I actually do, so I can say that is the reason if I dont do as well as I hoped I let people think that I dont get as much sleep or rest before upcoming exams and assignments than I actually do, so I can say that is the reason if I dont do as well as I hoped I let people think that I attend less classes in this area than I really do so I have an excuse if I dont do as well as I hoped

In my experience once a teacher gets the idea youre a poor student, your work is much more likely to get poor a grade than if someone else handed it in

Students who are not smart cant do anything to change that

My low ability is the reason I have not done well at times in the past

To get a clearer picture of students motivation and learning strategies, I am hoping to talk to a few students about the way they go about their studies. If you wouldnt mind talking to me about these things (at uni and at a time of your convenience) I would really appreciate it if you would leave a contact phone number. NOTE: THIS IS COMPLETELY OPTIONAL. If I do contact you, and then you would like to change your mind, thats no problem. First Name: Contact Phone: ( )

THANKYOU FOR YOUR PARTICIPATION

409

APPENDIX E 22 factor CFA for Nepean and Macarthur (n=393)

Chi square=3794.08 (df=2114), TLI=.89, RNI=.91, RMSEA=.04 Item parcel

Self-handicapping 1 & 2 Self-handicapping 3 & 4 Self-handicapping 7 & 9 Self-handicapping 10 & 11 Self-pres self-handicapping 1 & 2 Self-pres self-handicapping 4 & 6 Self-pres self-handicapping 7 & 9 Self-pres self-handicapping 10 & 11 Defensive pessimism 1 & 2 Defensive pessimism 3 & 4 Defensive pessimism 6 & 7 Defensive pessimism 8 & 9 Self-pres defensive pessimism 1 & 2 Self-pres defensive pessimism 3 & 4 Self-pres defensive pessimism 5 & 6 Self-pres defensive pessimism 7 & 8 Reflectivity 2 & 3 Reflectivity 4 & 5 Reflectivity 6, 7, & 8 Private enhancing 1 & 2 Private enhancing 3 & 4 Private enhancing 5 & 6 Public enhancing 1 & 2 Public enhancing 3 & 4 Public enhancing 5 & 6 Ability - failure 1 & 2 Ability - failure 3 & 4 Ability - failure 5 & 6 External - failure 1 & 2 External - failure 3 & 4 External - failure 5 & 6 External - success 1 & 2 External - success 3 & 4 External - success 5 & 6 Ego-orientation 1 & 2

Factor loading
.79 .83 .86 .82 .76 .81 .81 .87 .80 .84 .87 .78 .74 .83 .90 .90 .59 .78 .70 .85 .90 .74 .84 .89 .88 .67 .66 .74 .54 .64 .67 .45 .70 .70 .84

Item parcel
Ego-orientation 3 & 4 Ego-orientation 5 & 6 Public self-conscious 2 & 6 Public self-conscious 7 & 8 Public self-conscious 9 & 10 Task-orientation 2 & 3 Task-orientation 4 & 5 Task-orientation 6, 7, & 8 Avoid-orientation 1 & 2 Avoid-orientation 3 & 4 Avoid-orientation 6, 8, 9 Entity view 1 & 2 Entity view 3 & 4 Entity view 6 & 7 Increment view 1 & 2 Increment view 3 & 4 Increment view 5 & 6 Control future fail 1 & 2 Control future fail 3 & 4 Control future fail 5 & 6 Self-concept 1 & 2 Self-concept 3 & 4 Self-concept 5 & 6 Stability 1 & 2 Stability 3 & 4 Stability 5 & 7 Self-regulation 1 & 2 Self-regulation 4 & 5 Self-regulation 6, 7, & 8 Persistence 1 & 2 Persistence 3 & 5 Persistence 6, 7, & 8 Future plans 1 & 2 Future plans 3 & 4 Future plans 5, 6, & 7

Factor loading
.88 .87 .83 .89 .89 .77 .78 .82 .77 .81 .78 .74 .81 .82 .72 .74 .79 .83 .83 .84 .84 .88 .81 .73 .82 .78 .70 .84 .79 .65 .73 .79 .80 .89 .91

410

Correlations among strategies and both predictors and outcomes for Nepean and Macarthur (Time 1)

Active SH

Self-presented SH

DE

Selfpresented DE

Reflect

Active SH .91 Self-presented SH .38 .47 DE .56 .70 Self-presented DE .08 .19 Reflectivity .30 .04 Ability - failure .44 .44 External - failure .41 .48 External - success .30 .29 Entity view -.01 .01 Incremental view .24 .37 Ego-orientation -.32 -.24 Task-orientation .38 .45 Avidance orientation -.19 -.16 Future plans .32 .36 Control future failure -.52 -.48 Persistence .21 .31 Public self-consciousness -.05 .04 Private enhancing s-wth .19 .26 Public enhancing s-wth -.05 -.09 Self-concept .22 .29 Instability self-concept -.37 -.34 Self-regulation SH=Self-handicapping DE=Defensive expectations

.73 .36 .62 .47 .47 .22 -.11 .31 -.06 .49 -.17 .64 -.37 .37 .26 .34 -.25 .52 -.14

.42 .48 .50 .47 .27 -.04 .39 -.02 .50 -.19 .48 -.36 .48 .28 .42 -.09 .37 -.08

.53 .34 .33 .08 .32 .36 .47 .43 .01 .38 .04 .60 .55 .49 -.01 .39 .43

411

Correlations among predictors and outcomes for Nepean and Macarthur (Time 1)

1 1.Ability - failure 2.External - failure 3.External - success 4.Entity view 5.Incremental view 6.Ego-orientation 7.Task-orientation 8.Avoidance orient 9.Future plans 10.Control fut fail 11.Persistence 12.Pub self-cons 13.Private s-wth 14.Public s-wth 15.Self-concept 16.Instability s-c 17.Self-regulation Decimal omitted
52 49 38 -09 27 09 45 -31 65 -39 48 34 39 -49 45 -04

2
81 36 05 38 -02 48 -12 50 -39 27 23 28 02 37 -17

10

11

12

13

14

15

16

33 20 43 -05 55 -12 44 -55 31 21 32 -10 41 -19

-47 13 -26 22 -13 23 -26 05 -03 05 -02 07 -18

25 35 10 24 -05 12 15 27 18 22 15 22

17 44 03 21 -31 60 56 66 08 20 -07

07 25 03 39 35 63 33 11 15 47

-07 43 -39 61 45 58 -08 35 -10

-16 35 -04 04 -04 62 -06 16

-33 31 24 26 -23 44 -19

-17 24 26 -23 44 67

76 94 -03 36 10

80 02 33 23

01 33 05

-14 10

-02

412

APPENDIX F

MODELLING LATENT INTERACTION EFFECTS

Appendix F1. Technical issues involved in testing latent interaction effects Interaction effects in structural equation modelling have only in the last few years been given empirical attention. Whilst a seminal paper by Kenny and Judd in 1984 (see also Busemeyer & Jones, 1983) was published, this approach was difficult to implement and it is only recently that researchers have incorporated their suggestions into modelling packages such as LISREL. Indeed, it is only later versions of LISREL which can readily accommodate non-linear constraints in any straight-forward manner. With these recent advances in statistical software, a number of researchers have demonstrated how interaction effects can be estimated (see Jaccard & Wan, 1995, 1996; Joreskog & Yang, 1996; Ping, 1996a, 1996b). Of these approaches, the most parsimonious, requiring less specification tedium, and which encounters the least estimation difficulties is that recommended by Ping (1996a, 1996b). This involves a number of steps. First, items are mean centred (see Aiken & West, 1991). Second, these mean-centred items are estimated in a measurement model (CFA) and the loadings and error variances are estimated. Third, the product of the indicators is computed to create interaction indicators. Fourth, the loadings for the interaction indicators (a) are calculated by computing the product of the two main effect loadings derived from the measurement model (Equation 1). a=xz

(1)

Where x refers to the loadings on the first factor and z refers to loadings on the second factor.

Fifth, the error variances of the interaction indicators (Var[b]) are calculated under the Kenny and Judd (1984) normality assumptions using Equation 2. Var[b]=2xVar[X]Var[z]+2zVar[Z]Var[x]+Var[z]Var[x] 413

(2)

Where Var[X] refers to the variance of the latent factor, X; Var[Z] refers to the variance of the latent factor, Z; Var[x] refers to the variance of the uniqueness (error) associated with each item loading on factor X; Var[Z] refers to the variance of the uniqueness (error) associated with each item loading on factor Z.

Sixth, in the structural equation model, loadings and error variances for the latent interaction terms are fixed at their respective a and Var[b] values and the complete model estimated.

Appendix F2. Following up the significant Time 2 interaction between defensive expectations and reflectivity As outlined in Chapter 9, the interaction of defensive expectations and reflectivity significantly predicted self-regulation. Follow-up analyses to explore the precise nature of this interaction effect used the approach recommended by Aiken and West (1991). Values of self-regulation (Y) can be determined by the formula in Equation 3 in which represents the unstandardised path coefficients generated by LISREL in the SEM, X represents defensive expectations, Z represents reflectivity, and XZ represents the interaction of defensive expectations and reflectivity.

Y = BX + BZ + BXZ

(3)

Substituting B with coefficients generated by LISREL gives the formula presented in Equation 4.

Y = -1.32(X) + 1.78(Z) + .24(XZ)

(4)

Adapting recommendations of Aiken and West (1991), values of Y (self-regulation) were determined by inserting high and low values of X and Z into Equation 4. According to Aiken and West (see also Cohen & Cohen, 1983), high and low values can be assigned at one standard deviation above and below the mean (in this case 0 following mean centring). The standard deviation for X (defensive expectations) is 1.21 and for Z (reflectivity) is .86 yielding values for four groups: high defensive 414

expectations (1.21), low defensive expectations (-1.21), high reflectivity (.86), and low reflectivity (-.86). Substituting these values into Equation 4 yields four values for Y which are plotted in Figure 1. This pattern of results indicates that the interaction effect lies largely in the distinctiveness of the Low defensive expectations/High reflectivity group from the High defensive expectations/Low reflectivity group and these two from the other two groups (see Chapter 9 for substantive interpretations).

4 3 Self-regulation 2 1 0 -1 -2 -3 -4 -3.38 Strategy Profile Lo DE/Lo Ref Lo DE/Hi Ref Hi DE/Lo Ref Hi DE/Hi Ref 0.32 0.18 2.88

Figure 1 Graphing the effect on self-regulation of the Time 2 interaction between defensive expectations and reflectivity

415

APPENDIX G INTERVIEW CONSENT FORM - TIME 1 AND TIME 2 QUALITATIVE STUDIES

ACADEMIC PROCESS QUESTIONNAIRE

CONSENT FORM

I understand that the interview with Andrew Martin is for research purposes contributing to his PhD in Education. I understand and consent to him audio taping the interview and that when reporting extracts from the interview, a pseudonym will be used and not my real name. I have been informed of the purposes of the research. I agree that participation is voluntary and that I can withdraw from the interview at any stage with no penalties to me whatsoever. I understand that the audio tape will be stored in a secure location accessible only to Andrew Martin

Name

Signature

Date

Any person with concerns or complaints about the conduct of the research study can contact the Secretary of the Human Ethics Committee at University of Sydney, on (02) 9351 4811 or at UWS Macarthur, on (046) 203 641.

416

APPENDIX H SEMI-STRUCTURED INTERVIEW SCHEDULE - TIME 1 QUALITATIVE STUDY Have you come straight from school? If not, what did you do between school and university? What degree are you enrolled in (Primary, Secondary education etc.)? What subjects are you majoring in? So now youre at university, what do you think? Is university what you expected? Are you enjoying university? What made you decide to study Education? What do you think of the course? Have you found it demanding? Many assignments/exams? How have you performed in your assignments and exams up to now? What do you think of yourself as a student? Would you say that your view of yourself as a student is fairly consistent or would you say it can fluctuate from day to day? When you do well in your studies, what is usually the main reason?
(If focus needed give choice of ability, effort, or external factors and ask which is the main cause)

Do you ever do more poorly than you thought youd do? When this happens what is generally the main reason?
(If focus needed give choice of ability, effort, or external factors and ask which is the main cause)

How confident are you in your ability to avoid doing poorly in the future? What steps can you take to avoid doing poorly in the future? Some students are motivated at university because they want to strive for success while others are motivated because they want to avoid the negative consequences of not doing well. Would you relate to either of those two students? How? Would you say you feel more successful when you outperform others or when you learn/master something new? Why do you say that? To what extent are you concerned about how others view your academic performance? Why?

417

 How important is it to you that youre seen by others to be academically competent? Why? How important is it that you see yourself as competent? Why? How would you go about preparing for an exam or assignment? To what extent would you be organised, plan, or monitor your progress? When youre presented with a challenging academic task or problem, would you be more inclined to stick at it until you solved it or would you be more inclined to move on to something else?

For LODE and HIDE students If I ask you to look down the track to your next assignment or exam, do you think youll do the same, better, or worse than youve done before? Why do you say that? How does it make you feel when you go into an exam thinking more negatively/positively than previous experience would predict? Does thinking this way have any advantages/disadvantages? (If respondent reports defensive expectations) Would you tell people that youre feeling negative? (If respondent reports defensive expectations) Would you tell others that youre feeling more negative than you really are?

For LOSH and HISH students Some students do things that seem to get in the way of their success. For example, they might procrastinate before an upcoming assignment, become occupied on other things before the exams, or fool around the night before an exam. Other students get onto the job of doing their assignments and studying for the exam and maximise their chances of success. Would you relate to either of these two students? If so, in what ways do you do some of these things? Why do you think you do these things? What advantages might there be in doing these things? (If respondent reports self-handicapping) Would you tell other people that you do these things?

418

 (If respondent reports self-handicapping) Would you tell others that you do these things more than is actually the case? Thats all the questions I have to ask you. Is there anything you would like to add? Is there anything you feel I havent touched or focused on? Is there anything you would like to emphasise? Is there a particular way you go about your studies that I havent addressed?

419

APPENDIX I CATEGORIES AND NUDIST NODE ADDRESSES - TIME 1 QUALITATIVE STUDY

1 AFFECTIVE AND MOTIVATIONAL PREDICTORS 1.1 Performance orientation 1.1.1 Academic public self-consciousness 1.1.1.1 Concerned with how academic performance viewed by others 1.1.1.2 Not concerned with how viewed 1.1.1.3 Some concern but not dominated by it 1.1.1.4 Thats interesting Ego-orientation/competitiveness 1.1.2.1 Feelings of success through outperforming others 1.1.2.2 Feelings of success not dependent on outperforming others 1.1.2.3 Feelings of success only partly determined by outperforming others 1.1.2.4 Thats interesting Avoidance-oriented performance 1.1.3.1 Motivated to avoid negative consequences 1.1.3.2 Motivated to strive for success 1.1.3.3 Motivated by both aspiring to success and avoiding failure 1.1.3.4 Thats interesting Private esteem-relevant competence valuation 1.1.4.1 Important to see myself as competent 1.1.4.2 Not important to see myself as competent 1.1.4.3 Only somewhat important to see myself as competent 1.1.4.4 Thats interesting Public esteem-relevant competence valuation 1.1.5.1 Important that others see me as competent 1.1.5.2 Not important that others see me as competent 1.1.5.3 Only somewhat important that others see me as competent 1.1.5.4 Thats interesting

1.1.2

1.1.3

1.1.4

1.1.5

1.2 Attributional orientation 1.2.1 Causes of success 1.2.1.1 External causes of success eg. good luck, good teaching, easy marking 1.2.1.2 Internal causes of success eg. ability or effort 1.2.1.3 Both external and internal 1.2.1.4 Thats interesting Causes of failure 1.2.2.1 1.2.2.2 1.2.2.3 1.2.2.4

1.2.2

External causes of failure eg. bad luck, poor teaching, hard marking Internal causes of failure eg. ability or effort Both external and internal Thats interesting

420

1.3 Uncertain personal control 1.3.1 Control over avoiding poor future performance 1.3.1.1 Little control over ability to avoid doing poorly in future 1.3.1.2 Control over ability to avoid doing poorly in future 1.3.1.3 Some factors in my control, others out of my control 1.3.1.4 Thats interesting Stability of academic self-concept 1.3.2.1 View of myself stable and consistent from day to day 1.3.2.2 View of myself fluctuates somewhat from day to day 1.3.2.3 Extent to which it fluctuates depends 1.3.2.4 Thats interesting

1.3.2

1.4 Task-orientation 1.4.1 1.4.2 1.4.3 1.4.4 1.4.5 Feelings of success through mastery and new learning Feelings of success not dependent on mastery or new learning Feelings of success only partly dependent on mastery Thats interesting Both mastery and competitive concerns

2 ACADEMIC STRATEGIES 2.1 Self-handicapping 2.1.1 Students who self-handicap 2.1.1.1 Typical behaviour 2.1.1.2 Why do I do it? 2.1.1.3 What advantages does it offer? 2.1.1.4 Self-presented/public aspects 2.1.1.5 Thats interesting Students who do not self-handicap 2.1.2.1 Typical behaviour 2.1.2.2 Why do some students self-handicap? 2.1.2.3 What advantages does it offer? 2.1.2.4 Thats interesting

2.1.2

2.2

Defensive expectations 2.2.1 Students who think in a defensively expectant fashion 2.2.1.1 Typical defensive expectations 2.2.1.2 Why do I think this way? 2.2.1.3 What advantages does it offer? 2.2.1.4 Self-presented/public aspects 2.2.1.5 Thats interesting Students who think optimistically 2.2.2.1 Typical optimism 2.2.2.2 Why do I think this way? 2.2.2.3 Thats interesting

2.2.2

421

3 ACADEMIC OUTCOMES 3.1 Self-Regulation 3.1.1 3.1.2 3.1.3 3.1.4 3.2 I organise my approach, set goals, monitor progress No particular organisation or monitoring when going about study Only somewhat of a self-regulatory approach Thats interesting

Persistence 3.2.1 3.2.2 3.2.3 3.2.4 I stick at challenging tasks and persist I have a go at challenging tasks but soon move on to something else Extent to which I persist depends Thats interesting

4. Background details
4.1 Status 4.1 4.2 4.3 4.4

High self-handicapping Low self-handicapping High defensive expectations Low defensive expectations

4.2

University 5.1 Nepean 5.2 Macarthur 5.3 Sydney Sex 6.1 6.2

4.3

Male Female

422

APPENDIX J SUBSCALES ADDED TO QUESTIONNAIRE AT TIME 2

Pintrich and DeGroot (1990) anxiety items

1. I am so nervous during a test that I cannot remember facts I have learned. 2. I worry a great deal about tests. 3. When I take a test I think about how poorly I am doing. 4. I have an uneasy, upset feeling when I take a test.

Elliot and Church (1996) performance-avoidance goal items

1. I worry about the possibility of getting a bad grade. 2. My fear of performing poorly is often what motivates me 3. Im afraid that if I ask my lecturer a dumb question, they might not think Im very smart. 4. I just want to avoid doing poorly. 5. I wish these subjects were not graded. 6. I often think to myself, What if I do badly?

423

APPENDIX K

Assessing the Elliot and Church (1997) performance-avoidance subscale As noted, Time 1 had not incorporated an avoidance measure along the lines of that proposed by Elliot and Church (1997) and it was considered important to do so in order to clarify its role in the context of other predictors in the model generally and in relation to the three other motivation orientation scales (ego-orientation, taskorientation, and avoidance-oriented performance) more specifically. Accordingly, the Elliot and Church performance-avoidance orientation subscale was introduced at Time 2. In relation to this performance-avoidance subscale, however, concerns have been raised (Midgley et al., 1998) which suggest that it reflects worry and fear more than it does an avoidance orientation.

A procedure in LISREL (the AM - automatic modification - procedure) enables the researcher to leave target loadings for factors free to be estimated by the program (somewhat akin to exploratory factor analysis) whilst specifying the assignment of target loadings for factors whose loadings are not in question. In the context of the predictor subscales, this procedure was utilised to explore the factors on which the Elliot and Church performance-avoidance items loaded best. To ascertain whether this LISREL-determined item allocation is preferable to the theoretically/user-determined item allocation, the difference in chi square relative to the difference in degrees of freedom between the AM procedure and the user-determined procedure is compared. Consistent with contentions by Midgley et al. (1998) results from this procedure indicated that performance-avoidance items loaded more clearly on the anxiety factor (and also loaded quite strongly on other predictor factors) than they did on the Elliot and Church performance-avoidance factor. Moreover, this model (1916.57, df=902) fit the data better than one which specified avoidance items to load only on an avoidance factor (chi square=2027.23, df=898). In view of this, it was considered that the Elliot and Church performance-avoidance items did not adequately reflect an avoidance factor and were therefore excluded from subsequent analyses.

424

APPENDIX L
SEMI-STRUCTURED INTERVIEW SCHEDULE - TIME 2 QUALITATIVE STUDY

Did you come straight from school into first year university? If not, what did you do between school and first year? What degree are you enrolled in (Primary, Secondary education etc.)? Why did you decide to study Education? What did you think of first year? How were your grades in first year? So now youre in second year, what do you think? Is it what you expected? How were your grades in first semester? Thinking about your grades, to what extent do you feel they are due to factors such as teaching, luck, easy/hard marking? How does this compare to first year? Why do you say that? To what extent are you certain of your personal control over achieving success and avoiding failure? How does this compare to first year? Why do you say that? To what extent are you interested in mastering material, learning, gaining new knowledge and skills, and solving problems? How does this compare to first year? Why do you say that? To what extent are you interested in your performance and performance-related issues such as outperforming others, being seen to be competent, and avoiding failure? How does this compare to first year? Why do you say that? To what extent are you anxious when it comes to assignments and exams? How does this compare to first year? Why do you say that? 425

 What would you say is the major difference between first and second year for you? Has there been any significant event or series of events that marks first from second year? Have there been any major events which have changed the way you go about your studies in second year? It seems that there have been some shifts in the way you go about your studies between first and second year. In your opinion what are the main reasons for these shifts?

For students selected for upward or downward shifts in self-handicapping Some students do things that seem to get in the way of their success. For example, they might procrastinate before an upcoming assignment, become occupied on other things before the exams, or fool around the night before an exam. Other students get onto the job of doing their assignments and studying for the exam, maximising their chances of success. Would you relate to either of these two students? If so, in what ways do you do some of these things? How does this compare to first year? Why do you say that?

For students selected for upward or downward shifts in defensive expectations If I ask you to look down the track to your next assignment or exam, do you think youll do the same, better, or worse than youve done before? How does this compare to first year? Why do you say that? Thats all the questions I have to ask you. Is there anything you would like to add? Is there anything you feel I havent touched or focused on? Is there anything you would like to emphasise? Is there a particular way you go about your studies that I havent addressed?

426

APPENDIX M CATEGORIES AND NUDIST NODE ADDRESSES - TIME 2 QUALITATIVE STUDY

1. Background details 1.1 Status 1.1.1 1.1.2 1.1.3 1.1.4

Increase in self-handicapping Decrease in self-handicapping Increase in defensive expectations Decrease in defensive expectations

1.2

University 1.2.1 Nepean 1.2.2 Macarthur 1.2.3 Sydney Sex 1.3.1 1.3.2

1.3

Male Female

2. Academic strategies 2.1 Self-handicapping 2.1.1 Salience in second year 2.1.1.1 How/Describe 2.1.2 Increase from first to second year 2.1.2.1 How/Describe 2.1.3 Decrease from first to second year 2.1.3.1 How/Describe 2.1.4 Same in first as in second year 2.1.4.1 How/Describe 2.1.5 Thats interesting Defensive expectations 2.2.1 Salience in second year 2.2.1.1 How/Describe 2.2.2 Increase from first to second year 2.2.2.1 How/Describe 2.2.3 Decrease from first to second year 2.2.3.1 How/Describe 2.2.4 Same in first as in second year 2.2.4.1 How/Describe 2.2.5 Thats interesting

2.2

3. Affective and motivational predictors 3.1 External attributional orientation 3.1.1 Salience in second year 3.1.1.1 How/Describe 3.1.2 Increase from first to second year 3.1.2.1 How/Describe 3.1.3 Decrease from first to second year 3.1.3.1 How/Describe

427

3.1.4 3.1.5

Same in first as in second year 3.1.4.1 How/Describe Thats interesting

3.2

Uncertain personal control 3.2.1 Salience in second year 3.2.1.1 How/Describe 3.2.2 Increase from first to second year 3.2.2.1 How/Describe 3.2.3 Decrease from first to second year 3.2.3.1 How/Describe 3.2.4 Same in first as in second year 3.2.4.1 How/Describe 3.2.5 Thats interesting

3.3

Task-orientation 3.3.1 Salience in second year 3.3.1.1 How/Describe 3.3.2 Increase from first to second year 3.3.2.1 How/Describe 3.3.3 Decrease from first to second year 3.3.3.1 How/Describe 3.3.4 Same in first as in second year 3.3.4.1 How/Describe 3.3.5 Thats interesting

3.4

Performance orientation 3.4.1 Salience in second year 3.4.1.1 How/Describe 3.4.2 Increase from first to second year 3.4.2.1 How/Describe 3.4.3 Decrease from first to second year 3.4.3.1 How/Describe 3.4.4 Same in first as in second year 3.4.4.1 How/Describe 3.4.5 Thats interesting

3.5

Anxiety 3.5.1 3.5.2 3.5.3 3.5.4 3.5.5 Salience in second year 3.5.1.1 How/Describe Increase from first to second year 3.5.2.1 How/Describe Decrease from first to second year 3.5.3.1 How/Describe Same in first as in second year 3.5.4.1 How/Describe Thats interesting

4. Major differences between first and second year (nominated by respondents) 4.1 Differences between first and second year 4.1.1 Differences that yielded positive outcomes 4.1.2 Differences that did not yield positive outcomes Major events distinguishing first from second year

4.2

428

4.3

4.2.1 Events that yielded positive outcomes 4.2.2 Events that did not yield positive outcomes Thats interesting

429