Copyright 1998 Elsevier Science Ltd. All rights reserved.

8.01 Health Behavior

MARK CONNER University of Leeds, UK and PAUL NORMAN University of Sheffield, UK
8.01.1 INTRODUCTION 8.01.2 KEY BEHAVIORS INFLUENCING HEALTH, MORBIDITY, AND MORTALITY 8.01.2.1 8.01.2.2 8.01.2.3 8.01.2.4 8.01.2.5 Introduction Smoking Diet Exercise Other Behaviors 2 2 2 3 3 3 4 5 5 6 6 7 8 8 8 10 11 12 12 12 13 14 14 15 17 17 17 18 18 19 19 21 21 22 24 24 27

8.01.3 UNDERSTANDING THE BASIS OF HEALTH BEHAVIOR 8.01.3.1 Introduction 8.01.3.2 Health Belief Model 8.01.3.2.1 Model description 8.01.3.2.2 Review of research 8.01.3.2.3 Commentary 8.01.3.3 Theory of Planned Behavior 8.01.3.3.1 Model description 8.01.3.3.2 Review of research 8.01.3.3.3 Commentary 8.01.3.4 Health Locus of Control 8.01.3.4.1 Model description 8.01.3.4.2 Review of research 8.01.3.4.3 Commentary 8.01.3.5 Protection Motivation Theory 8.01.3.5.1 Model description 8.01.3.5.2 Review of research 8.01.3.5.3 Commentary 8.01.3.6 Self-efficacy 8.01.3.6.1 Model description 8.01.3.6.2 Review of research 8.01.3.6.3 Commentary 8.01.4 ENCOURAGING THE ADOPTION OF HEALTH BEHAVIORS 8.01.4.1 Intervention Studies 8.01.5 LIMITATIONS AND EXTENSIONS OF EXISTING MODELS 8.01.5.1 8.01.5.2 8.01.5.3 8.01.5.4 8.01.5.5 Limitations of Current Models Additional Theoretical Constructs Processes by which Cognitions Influence Behavior Stage Models of Health Behavior Theoretical Integrations and Future Directions

2
8.01.6 CONCLUSIONS 8.01.7 REFERENCES

Health Behavior
28 28

8.01.1 INTRODUCTION Interest in health behaviors is derived from two assumptions; that a substantial proportion of the mortality from the leading causes of death is attributable to the behavior of individuals, and that the behavior is modifiable (Stroebe & Stroebe, 1995). It is now widely recognized that individuals can influence their own health and well-being through the adoption of healthenhancing behaviors (e.g., exercise) and the avoidance of health-compromising behaviors (e.g., smoking). The identification of the factors which predict who engages in health behaviors has become a focus of research in health psychology and other health-related disciplines in recent years (e.g., Adler & Matthews, 1994; Conner & Norman, 1996a; Glanz, Lewis & Rimmer, 1990; Hockbaum & Lorig, 1992; Rodin & Salovey, 1989; Winett, 1985). The health behaviors examined have been many and varied; from health enhancing behaviors such as exercise and healthy eating, on the one hand, to avoidance of health harming behaviors such as smoking and excessive alcohol consumption, on the other. Each of these behaviors have immediate or long-term effects upon the individual's health and are to varying extents within the individual's control. Epidemiological studies reveal great variability in who performs these behaviors. The approaches taken to understanding such individual differences have been equally varied. One can classify these approaches into those which examine factors intrinsic to the individual (e.g., sociodemographic factors, personality, social support, cognitions) and those which examine factors extrinsic to the individual. This second group of approaches can be further divided into those based on incentive structures (e.g., taxing tobacco and alcohol, subsidizing sports facilities) and those based on legal restrictions (e.g., banning dangerous substances, financial penalties for not wearing seat-belts). The first approach (factors intrinsic to the individual) has received most attention from psychologists, with a particular focus on cognitive factors as the most important proximal determinants of behavior. A variety of models of how such cognitive factors produce various social behaviors have been developed and are commonly referred to as social cognition models (SCMs). SCMs are recognized to have made a valuable contribution to the understanding of both who performs health behaviors (Conner &

Norman, 1996b; Marteau, 1989) and how extrinsic factors produce behavior change (e.g., Rutter & Quine, 1996). This chapter examines the key health behaviors and ways in which their adoption might be encouraged. The chapter then considers the contribution psychology has made to understanding and changing health behaviors through the development of SCMs. The most widely applied SCMs are described and reviewed along with recent developments and future prospects for this field of research. 8.01.2 KEY BEHAVIORS INFLUENCING HEALTH, MORBIDITY, AND MORTALITY 8.01.2.1 Introduction We might define health behavior as any activity taken for the purpose of preventing or detecting disease or for improving general wellbeing (Conner & Norman, 1996b). The behaviors within this definition include medical service usage (e.g., physician visits, vaccination, screening), compliance with medical regimens (e.g., dietary, diabetic, antihypertensive regimens), and self-directed health behaviors (e.g., diet, exercise, smoking, alcohol consumption). Numerous studies have examined the relationship between health behaviors and health outcomes (e.g., Black Report, 1988; Blane, Smith, & Bartley, 1990; Blaxter, 1990; Cox, Huppert, & Whichelow, 1993; Doll, Peto, Wheatley, Gray, & Sutherland, 1994; Gottlieb & Green, 1984). Such studies demonstrate the role of a variety of behaviors for both morbidity and mortality. One of the first such studies in Alameda County identified seven features of lifestyle: not smoking, moderate alcohol intake, sleeping 78 hours per night, exercising regularly, maintaining a desirable body weight, avoiding snacks, and eating breakfast regularly which together were associated with lower morbidity and higher subsequent long-term survival (Belloc, 1973; Belloc & Breslow, 1972; Breslow & Enstrom, 1980). Such results have been replicated in a variety of different populations (e.g., Brock, Haefner, & Noble, 1988; Metzner, Carman, & House, 1983). The impact of health behaviors upon individuals' quality of life, via delaying the onset of chronic disease and extending active life span, should also be noted (Fries, Green, & Levine, 1989; Stroebe & Stroebe, 1995). Smoking, alcohol consumption,

Key Behaviors Influencing Health, Morbidity, and Mortality diet, gaps in primary care services and low screening uptake are all significant determinants of poor health (Amler & Dull, 1987). Such findings have led to a focus by those interested in health promotion on changing such behaviors in order to improve health. For example, in the USA, Healthy People 2000 (USDHHS, 1990) lists increased physical activity, changes in nutrition and reductions in tobacco, alcohol, and drug use as important for health promotion and disease prevention. Below we examine several health behaviors in more detail, focusing on their prevalence and relationship to health outcomes. 8.01.2.2 Smoking Smoking is the behavior most closely linked with long-term negative health outcomes. Both morbidity and mortality from coronary heart disease (CHD) are increased among smokers (Doll et al., 1994; Friedman, Dales, & Ury, 1979). Moreover, there is a strong positive relationship between the number of cigarettes smoked per day and the incidence of CHD (Friedman et al., 1979). In addition, smoking has been linked to a number of cancers including cancer of the lung, throat, stomach, and bowel as well as a number of more immediate negative health effects such as reduced lung capacity and bronchitis (Royal College of Physicians, 1983). Despite the array of negative health outcomes, smokers often report positive mood effects from smoking and the use of smoking as a coping strategy (Graham, 1987). The number of people smoking in the USA and UK has shown a steady decline over the past 20 years. Data from the General Household Survey (1992) showed that 28% of people over the age of 16 smoke in the UK. Smoking is more common among men and among unskilled manual workers (General Household Survey, 1994). A similar pattern is evident in the USA, with smoking more common among less educated, lower income, and minority groups (Rigotti, 1989). Those who quit smoking reduce the risk to their health, particularly if they quit before 35 years of age (Doll et al., 1994). 8.01.2.3 Diet The impact of various aspects of diet upon health, morbidity, and mortality are well established (USDHHS, 1988). Whilst in the Third World the problems related to diet and health are ones of undernutrition, in the First World, the problems are predominantly linked to overconsumption of food. In Western industrialized countries the major problems

are linked to excessive fat consumption and insufficient fiber, fruit, and vegetable consumption. In addition, excess consumption of calories combined with insufficient exercise has made obesity a major health problem. Diet has been implicated in cardiovascular diseases (CVDs), strokes and high blood pressure, cancer, diabetes, obesity, osteoporosis, and dental disease. It is generally agreed that elevated blood cholesterol level is a major risk factor for the development of CVD (Consensus Development Conference on Lowering Blood Cholesterol to Prevent Heart Disease, 1985). Nutbeam and Catford (1990) estimate that 26% of men and 25% of women in the UK have cholesterol levels greater than 6.5 mmol l71 (a level considered to be excessive). While in the USA, it is estimated that 50% of the adult population is at risk of CHD by virtue of elevated blood cholesterol levels (Sampos, Fulwood, Haines et al., 1989). International studies have clearly demonstrated an association between saturated fat consumption (one source of cholesterol) and blood cholesterol levels. For example, Keys (1970) reported a correlation of 0.89 across the seven countries studied. The reduction of blood cholesterol via dietary change is now widely accepted as an important way of tackling CHD (Expert Panel, 1993). Drug treatment to reduce blood cholesterol levels is generally seen as only advisable if dietary change is ineffective (Wardle, 1995). Dietary recommendations include reducing fat in the diet and increasing soluble fiber intake (Committee on Medical Aspects of Food Policy, 1991; Expert Panel, 1993). However, their impact upon cholesterol levels may be limited. 8.01.2.4 Exercise Engaging in regular exercise is seen to be another key component of a healthy lifestyle. The potential health benefits of engaging in regular exercise are many and include reduced cardiovascular morbidity and mortality (Oberman, 1985), lowered blood pressure (Blair, Goodyear, Gibbons, & Cooper, 1984) and the increased metabolism of carbohydrates (Lennon et al., 1983) and fats (Rosenthal, Haskell, Solomon, Widstrom, & Reavan, 1983), as well as a range of psychological benefits such as improved self-esteem (Sachs, 1984), positive mood states (Folkins & Sime, 1981), reduced life stress (Brown, 1991), and reduced levels of anxiety (Singer, 1992). However, despite the various health benefits of exercise, a significant proportion of the population lead a sedentary lifestyle. For

Health Behavior used cannabis at some stage, while the Leitner et al. (1993) study reported that 10% of the same age group had used it. In general, drug use in the UK does not differ substantially from the rest of the Western world. Reported lifetime usage in the USA is slightly higher than in the UK, although estimates for injecting use (12% of the general population), alcohol use (around 85%), and lifetime tobacco use (around 75%) are roughly similar. Sexual behaviors have also long been considered health behaviors because of their impact upon the spread of sexually transmitted diseases (STDs) such as gonorrhoea and syphilis. More recently, the role of sexual behaviors in the spread of the human immunodeficiency virus (HIV) has been a focus of attention (O'Leary & Raffaelli, 1996). Whilst early health education campaigns emphasized the need to reduce the number of sexual partners or avoid particular sexual practices (e.g., anal sex, penetrative sex), more recently the focus has been upon the use of condoms during penetrative sex to reduce the risk of HIV transmission (Reiss & Leik, 1989). Condom use is particularly recommended for those with multiple partners or those who do not know their partners' sexual history. For these reasons, much of the health advice concerning condom use has been focused on young people. There seems to be considerable variation in the use of condoms in response to the threat of HIV/AIDS. For example, among heterosexuals, Richard and van der Pligt (1991) reported that 50% of their sample of Dutch teenagers with multiple partners consistently used condoms. While other studies in the UK and USA report rates of between 24% and 58% (Fife-Schaw & Breakwell, 1992; Gerrard, Gibbons, & Bushman, 1996). Among homosexuals, Weatherburn, Hunt, Davies, Coxon, and McManus (1991) reported that 39% of their sample always used a condom during anal sex. Whilst among bisexuals, Boulton, Schram Evans, Fitzpatrick, and Hart (1991) report that 25% of their male sample used condoms with their current male partner and only 12% with their current female partner. The General Household Survey (1993) in the UK reported changes by age group in the use of condoms for the period 19831991. Among 1624-year-olds, condom use increase from around 6% to around 12% during this period, whilst among 4049-year-olds it dropped from around 18% to around 12% over the same period. Thus, health messages aimed at increasing condom use may be having some impact among younger people, although the overall rates of use are still worrying low. High alcohol consumption has been linked to a range of negative health outcomes including high blood pressure (Shaper et al. 1981), heart

example, the General Household Survey (1989) indicated that only one in three men and one in five women in the UK participate in any sport or recreational physical activity. Moreover, the Allied Dunbar Fitness Survey (1992) of 6000 English adults reported that one in six adults had done no exercise (i.e., for 20 minutes or more at a moderate or vigorous level) in the previous four weeks. Participation in regular exercise is strongly related to a number of sociodemographic variables. In particular, young people and males are more likely to engage in regular exercise. For example, the 1988 Welsh Heart Health Survey (Health Promotion Authority for Wales, 1990) reported that among 1834 year olds, 61% of men engaged in moderately vigorous exercise at least two times a week compared with only 35% of women. For 3564 year olds, the percentages drop to 37% for men and 17% for women. Overall, the typical exerciser is likely to be young, welleducated, affluent, and male (King et al., 1992). 8.01.2.5 Other Behaviors A number of other behaviors show clear links to health. Below we briefly consider recreational drug use, safe sex, alcohol use, and health screening as behaviors with important health consequences. The use of recreational drugs has long been recognized as a potential health problem which needs to be tackled (Aguirre-Molina & Gorman, 1996). The most commonly used drugs are alcohol and tobacco, however, here we consider what are usually referred to as psychoactive drugs (i.e., a drug that alters mood or behavior). There are a large number of such drugs including hypnotic drugs such as barbiturates which reduce anxiety and produce sedation; stimulants such as amphetamines which elevate mood, increase wakefulness, and give an enhanced sense of mental and physical energy; opiates such as opium and heroin which produce pleasant mood states; antipsychotic agents such as chlorpromazine which diminish the symptoms of psychoses; and psychedelics/ hallucinogens such as LSD, cannabis, and MDMA (3,4-methylenedioxymethamphetamine) which cause visual and auditory hallucinations. Most of these drugs have medical uses as well as being used recreationally. Data from various drug use surveys in the UK (e.g., Leitner, Shapland, & Wiles, 1993) suggest that opiate use has remained stable (at around 1%), but the use of cannabis and other nonopiates has been on the increase over the last 25 years. The 1968 OPCS survey reported that 2% of the 16 years and over age group had

Understanding the Basis of Health Behavior disease (Sherlock, 1982), and cirrhosis of the liver (Colliver & Malin, 1986), although there is some evidence to suggest that low levels of alcohol consumption may have slightly beneficial effects on health (Hennekens, 1983). High levels of alcohol consumption have also been associated with accidents, injuries, suicides, crime, domestic violence, rape, murder, and unsafe sex (British Medical Journal, 1982). While many of the adverse effects of high alcohol consumption are due to continued heavy drinking (e.g., cirrhosis of the liver, heart disease), others are more specifically related to excessive alcohol consumption in a single drinking session (e.g., accidents, violence) (Honkanen et al., 1983). The General Household Survey (1992) reported that the average weekly consumption of alcohol in the UK was 15.9 units (approximately 8 pints of beer) for men and 5.4 (approximately 2.5 pints of beer) for women. Of more interest was the finding that about 27% of men and 11% of women were drinking more than the recommended weekly sensible limits (21 units for men, 14 units for women). Heavy drinking is also more likely among younger age groups. In a survey of 12 000 Welsh adults, Moore, Smith, and Catford (1994) reported that 31.1% of drinkers aged 1824 engaged in binge drinking (i.e., drinking half the recommended weekly consumption of alcohol in a single session) at least once a week. Finally, individuals may seek to protect their health by participating in various screening programs which attempt to detect disease at an early, or asymptomatic, stage. In the UK, screening programs have been set up for various diseases including anemia (Ashworth, 1963), diabetes (Redhead, 1960), bronchitis (Gregg, 1966), cervical cancer (Freeling, 1965), and breast cancer (Forrest, 1986). Considering breast cancer, it has been estimated that breast screening programs which include mammograms can reduce breast cancer mortality by up to 40% among women aged 50 and over (Strax, 1984). However, participation rates in breast screening programs show great variability across different countries, ranging from 25% to 89% (Vernon, Laville, & Jackson, 1990). Participation tends to be negatively related to age and positively related to education level and socioeconomic status (Vernon et al., 1990). 8.01.3 UNDERSTANDING THE BASIS OF HEALTH BEHAVIOR 8.01.3.1 Introduction A clearer understanding of why individuals perform health behaviors might assist in the

development of interventions to help individuals gain the benefits of improved health and well-being. A variety of factors have been found to account for individual differences in the performance of various health behaviors, including demographic factors, social factors, emotional factors, perceived symptoms, factors relating to access to medical care, personality factors, and cognitive factors (Adler & Matthews, 1994; Rosenstock, 1974; Taylor, 1991). Demographic variables show reliable associations with the performance of health behaviors. For example, there is a curvilinear relationship between many health behaviors and age, with high incidences of many healthrisking behaviors such as smoking in young adults and much lower incidences in children and older adults (Blaxter, 1990). Such behaviors also vary by gender, with females being generally less likely to smoke, consume large amounts of alcohol, engage in regular exercise but more likely to monitor their diet, take vitamins, and engage in in dental care (Waldron, 1988). Differences by socioeconomic status and ethnic group are also apparent for behaviors such as diet, exercise, alcohol consumption, and smoking (e.g., Blaxter, 1990). Generally speaking, younger, wealthier, better educated individuals, under low levels of stress, with high levels of social support, are more likely to practice health enhancing behaviors. Higher levels of stress and/or fewer resources are associated with health-compromising behaviors such as smoking and alcohol abuse (Adler & Matthews, 1994). Social factors seem to be important in instilling health behaviors in childhood. Parent, sibling, and peer influences are important, for example, in the initiation of smoking (e.g., McNeil et al., 1988). Cultural values also have a major impact, for instance in determining the number of women exercising in a particular culture (e.g., Steptoe & Wardle, 1996). For example, Steptoe and Wardle (1992) report that between 34% and 95% of women in their European student sample had exercised in the past 14 days. Perceived symptoms control health habits when, for example, smokers regulate their smoking on the basis of sensations in the throat. Access to medical care has been found to influence the use of such health services (e.g., Black Report, 1988). Personality factors have also been associated with health behaviors (Adler & Matthews, 1994; Steptoe et al., 1994). Cognitive factors also determine whether or not an individual practices health behaviors. Knowledge about behaviorhealth links is an important factor in an informed choice concerning a healthy lifestyle. Various other cognitive variables have been studied including

Health Behavior that they do not necessarily provide an adequate description of the way in which individuals make decisions (e.g., Edwards, 1992; Feather, 1982; Frisch & Clemen, 1994; Jonas, 1993). In the section which follows, the most widely used of these models (HBM, TPB, HLOC, PMT, SE) are outlined, the research using them described, and their use reviewed. 8.01.3.2 Health Belief Model 8.01.3.2.1 Model description The HBM is probably the most widely used social cognition model in health psychology (Becker, 1974; Rosenstock, 1966; Sheeran & Abraham, 1996). It was originally developed by US public health researchers attempting to develop models upon which to base health education programs (Hochbaum, 1958; Rosenstock, 1966). The model attempts to conceptualize the health beliefs which make a behavior more or less attractive. In particular, the key health beliefs were seen to be the likelihood of experiencing a health problem, the severity of the consequences of the health problem, and the perceived costs and benefits of the health behavior. Thus, the HBM employs two aspects of individuals' representations of health behavior in response to threat of illness: perceptions of the threat of illness and evaluation of the effectiveness of behaviors to counteract this threat (see Figure 1). Threat perceptions depend upon two beliefs: the perceived susceptibility to the illness and the perceived severity of the consequences of the illness. Together these two variables determine the likelihood of the individual following a health-related action, although their effect is modified by individual differences in demographic variables, social pressure, and personality. The particular action taken is determined by evaluation of the possible alternatives. This behavioral evaluation depends upon beliefs concerning the benefits or efficacy of the health behavior and the perceived costs or barriers to performing the behavior. Hence, individuals are likely to follow a particular health action if they believe themselves to be susceptible to a particular condition or illness which they consider to be serious, and believe the benefits of the action taken to counteract the condition or illness outweigh the costs. Cues to action and health motivation are two other variables commonly included in the model. Cues to action include a diverse range of triggers to the individual taking action and are commonly divided into factors which are internal (e.g., physical symptom) or external (e.g., mass media campaign, advice from others such as physicians) to the individual (Janz & Becker, 1984).

perceptions of health risk, efficacy of behaviors in influencing this risk, social pressures to perform the behavior, and control over performance of the behavior. The relative importance of various cognitive factors in determining who performs various health behaviors constitutes the basis of different models. Such models have been labeled SCMs because of their focus on cognitive variables as the primary determinant of individual social behaviors. Two types of SCMs have been applied in health psychology, predominantly to explain health-related behaviors and response to treatment (Conner, 1993). The first type focus on individuals' understanding of the causes of health-related events and are best typified by attribution models (e.g., King, 1982). The second type are more diverse in nature and attempt to predict future health-related behaviors and outcomes. These include the health belief model (HBM; e.g., Becker, 1974; Janz & Becker, 1984; Sheeran & Abraham, 1996), health locus of control (HLOC; Norman & Bennett, 1996; Seeman & Seeman, 1983; Wallston, Wallston, & De Vellis, 1978), protection motivation theory (PMT; e.g., Boer & Seydel, 1996; Rogers, 1983; van der Velde & van der Pligt, 1991), theory of reasoned action/theory of planned behavior (TRA/TPB; e.g., Ajzen, 1988; 1991; Ajzen & Fishbein, 1980; Conner & Sparks, 1996), and self-efficacy (SE; e.g., Bandura, 1982, 1991; Schwarzer, 1992; Schwarzer & Fuchs, 1996). Other models include self-regulation theory (Leventhal, Nerenz, & Steele, 1984), the transtheoretical model of change (Prochaska & DiClemente, 1984), the precaution-adoption process (Weinstein, 1988), and the model of goal achievement (Bagozzi, 1992). However, none of these latter models have been widely applied to the prediction of health behaviors at present. These social cognition models provide a basis for understanding the determinants of behavior and behavior change. Each of these models emphasize the rationality of human behavior, although they do not assume that all behavior is based upon careful thought (Ajzen, 1996). Most assume that behavior and decisions are based upon elaborate, but subjective, cost/benefit analysis of the likely outcomes of differing courses of action. As such they have roots going back to expectancy-value theory (Peak, 1955) and subjective expected utility theory (SEU; Edwards, 1954). It is assumed that individuals generally aim to maximize utility and so prefer behaviors which are associated with the highest expected utility (Van der Pligt & de Vries, 1998). Whilst such considerations may well provide good predictions of which behaviors are selected, it has been noted by several authors

Understanding the Basis of Health Behavior

EXTERNAL VARIABLES
Demographic variables Age, sex, occupation, socioeconomic status, religion, education Personality traits Extraversion Agreeableness Conscientiousness Neuroticism Openness Other psychological factors Peer pressure Self-efficacy

Threat (motivation) Perceived susceptibility

Perceived severity

Health motivation

Behavior

Response effectiveness Perceived benefits Cues to action Perceived barriers

Figure 1 Health belief model.

Becker (1974) has argued that the HBM should also contain a measure of health motivation (readiness to be concerned about health matters) because certain individuals may be predisposed to respond to cues to action because of the value they place on their health. Other influences upon the performance of health behaviors, such as demographic factors or psychological characteristics (e.g., personality, peer pressure, perceived control over behavior), are assumed to exert their effect via changes in the six components of the HBM. This is a potentially important issue if the HBM is to claim to be a complete model of health behavior. However, this has not been widely addressed in empirical studies and where it has the evidence has been equivocal. Orbell, Crombie, and Johnson (1995), for example, reported that HBM components did mediate the effects of social class upon uptake of cervical screening, but did not mediate the effects of marital status or sexual experience. 8.01.3.2.2 Review of research The HBM has been applied to a very broad range of health behaviors and populations.

Sheeran and Abraham (1996) distinguish three broad areas of research. First, the HBM has been applied to various preventive health behaviors. These include health-risk behaviors such as smoking (Gianetti, Reynolds, & Rihen, 1985; Stacey & Lloyd, 1990) and alcohol use (K. H. Beck, 1981; Gottlieb & Baker, 1986), as well as health-promoting behaviors such as diet (Aho, 1979), exercise (Langlie, 1977), genetic (Becker, Kaback, Rosenstock, & Ruth, 1975) and health screening (Conner & Norman, 1994; King, 1982), vaccination (Oliver & Berger, 1979), breast self-examination (Champion, 1984; Ronis & Harel, 1989), contraceptive use (Hester & Macrina, 1985), and dental behaviors (Chen & Land, 1986). A second area the HBM has been applied to is various sick role behaviors which refer to compliance with professionally recommended medical regimens in response to illness. These include compliance with antihypertensive regimens (Taylor, 1979), diabetic regimens (Harris & Lynn, 1985), and renal disease regimens (Hartman & Becker, 1978), and regimens adhered to by parents for a child's condition (Becker, Radius, & Eveland, 1978). Third, the HBM has been applied to clinic use, which includes physician visits for a variety of

Health Behavior mance of health behavior partly account for the model's popularity. However, compared to other similar social cognitive models of health behaviors, the HBM suffers from a number of weaknesses. The way in which the variables in the HBM combine to produce behavior has not been precisely specified (but see Becker & Rosenstock, 1987) and so the HBM is thus frequently tested as six independent predictors of behavior. In addition, various researchers have used somewhat different operationalizations of the six constructs (see Becker & Maiman, 1983; Rosenstock, 1974). Together these factors have weakened the status of the HBM as a coherent SCM of health behavior (Conner, 1993; Sheeran & Abraham, 1996). Moreover, key social cognitive variables, found to be highly predictive of behavior in other models, are not incorporated in the HBM. For example, intentions to perform a behavior and social pressure are key components of the TRA/TPB which do not appear in the HBM. Also, perceptions of personal control over the performance of the behavior (self-efficacy beliefs) which have been found to be such powerful predictors of behavior in models based upon social cognitive theory (Bandura, 1982; Schwarzer & Fuchs, 1996) are not explicitly included in the HBM. In addition, in not specifying a causal ordering among the variables, as is done in other models, more powerful analysis of data and clearer indications of how interventions may have their effects are precluded in the HBM. Several authors have noted, for example, that threat is perhaps best seen as a more distal predictor of behavior acting via influences upon outcome expectancies. Finally, the model is static; there is no distinction between a motivational stage dominated by cognitive variables and a volitional phase where action is planned, performed, and maintained (Schwarzer, 1992). Such distinctions are thought to be important in understanding various health behaviors. Hence, while an extremely popular SCM for use in understanding health behavior, it is also in a number of ways limited and may receive relatively less attention in the future. 8.01.3.3 Theory of Planned Behavior 8.01.3.3.1 Model description The TPB was developed by social psychologists and has been widely applied to the understanding of a variety of behaviors including health behaviors (Ajzen, 1988, 1991; Conner & Sparks, 1996) (see Figure 2). The TPB details how the influences upon an individual determine that individual's decision to follow a

reasons including preventative (Aiken, West, Woodward, Reno, & Reynolds, 1994), psychiatric (Connelly, 1984), and parent and child conditions (Kirscht, Becker, & Eveland, 1976). There is no strong evidence that the HBM has been more predictive of behavior in any one of these behaviors compared with any other (Sheeran & Abraham, 1996). There have been two quantitative reviews of research with the HBM (Harrison, Mullen, & Green, 1992; Janz & Becker, 1984). The first, conducted by Janz and Becker (1984), examined the proportion of times each of the HBM's components showed a significant relationship with health behavior through the use of a significance ratio. Across the 46 studies reviewed, the barriers component was found to have the most consistent relationship with health behavior (89%), followed by the susceptibility (81%), benefits (78%), and severity (65%) components. However, while Janz and Becker (1984) suggest that the HBM components are consistent predictors of health behavior, they fail to estimate the strength of the relationships. This question was addressed in the second quantitative review, conducted by Harrison et al. (1992). Over 200 published studies on the HBM were identified although only 16 of these were found to measure each of the components adequately. Harrison et al's (1992) meta-analysis on these 16 studies produced a similar pattern of results to Janz and Becker's (1984) earlier review with the barriers components having the highest average correlation with health behavior (r = 70.21), followed by the susceptibility (r = 0.15), benefits (r = 0.13), and severity (r = 0.08) components. The predictive power of individual components is therefore relatively modest, accounting for only 0.54% of variance in behavior. However, it should be noted that it is the combined effects of the six health beliefs which is generally of interest and this is commonly in excess of the sum of the effects of the individual components. 8.01.3.2.3 Commentary The HBM has provided a useful framework for investigating health behaviors and identifying key health beliefs, has been widely used, and has met with moderate success in predicting a range of health behaviors (for reviews see Harrison et al., 1992; Janz & Becker, 1984; Sheeran & Abraham, 1996). The strength of the HBM lies in the fact that it was developed by researchers working directly with health behaviors and so many of the concepts possess facevalidity to those working in this area. This commonsense operationalization of a number of cognitive variables relevant to the perfor-

Understanding the Basis of Health Behavior particular behavior. This theory is an extension of the widely applied TRA (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975). The TPB suggests that the proximal determinants of behavior are intentions to engage in that behavior and perceived behavioral control over that behavior. Intentions represent a person's motivation in the sense of his or her conscious plan or decision to exert effort to perform the behavior. Perceived behavioral control is a person's expectancy that performance of the behavior is within his/her control. The concept is similar to Bandura's (1982) concept of selfefficacy (see Schwarzer & Fuchs, 1996). Control is seen as a continuum with easily-executed behaviors at one end and behavioral goals demanding resources, opportunities, and specialized skills at the other. Intentions are determined by three variables. The first is attitudes, which are the overall evaluations of the behavior by the individual. The second is subjective norms, which consist of a person's beliefs about whether significant others think he/she should engage in the behavior. The third is perceived behavioral control (PBC), which is the individual's perception of the extent to which performance of the behavior is within his/her control. In addition, to the extent that PBC reflects actual control, it is predicted to directly influence behavior. Just as intentions are held to have determinants, so the attitude, subjective norm, and

perceived behavioral control components are also held to have determinants. The attitude component is a function of a person's salient behavioral beliefs, which represent perceived likely consequences of the behavior. Following expectancy-value conceptualizations (Peak, 1955), the model quantifies consequences as being composed of the multiplicative combination of the judged likelihood that performance of the behavior will lead to a particular outcome and the evaluation of that outcome. These expectancy-value products are then summed over the salient consequences. It is not claimed that individuals perform such calculations each time they are faced with a decision about whether to perform a behavior or not, but rather the results of such considerations are maintained in memory and retrieved and used when necessary (Eagly & Chaiken, 1993). However, it is also possible for the individual to retrieve the relevant beliefs and evaluations when necessary. Subjective norm is a function of normative beliefs, which represent perceptions of specific salient others' preferences about whether one should or should not engage in a behavior. In the model, this is quantified as the subjective likelihood that specific salient groups or individuals (referents) think the person should or should not perform the behavior, multiplied by the person's motivation to comply with that referent's expectation. Motivation to comply is the extent to which the person wishes to comply

EXTERNAL VARIABLES
Demographic variables Age, sex, occupation, socioeconomic status, religion, education Personality traits Extraversion Agreeableness Conscientiousness Neuroticism Openness Perceived Perceived likelihood facilitating/ x of inhibiting occurrence power Perceived behavioral control Normative Motivation x beliefs to comply Subjective norm Behavioral intention Behavior Belief Evaluation about x of outcomes outcomes Attitude towards behavior

Figure 2 Theory of planned behavior.

10

Health Behavior Several studies examined condom use (Boldero, Moore, & Rosenthal, 1992; Wilson, Zenda, McMaster, & Lavelle, 1992). Nucifora, Gallois, and Kashima (1993), for example, examined undergraduates' use of condoms using the TPB. PBC was found to make a small but significant contribution to the predictions of intentions to use condoms and actual condom use. However, intentions appeared to be principally determined by attitudes and subjective norms, while behavior was mainly influenced by intentions. Exercise has also been examined in several studies (Dzewaltowski, Noble, & Shaw, 1990; Godin & Shepherd, 1987; Norman & Smith, 1995). Dzewaltowski et al. (1990) reported the application of the TPB to exercise participation. Intentions were based both upon attitudes and PBC, but not subjective norms, whilst actual behavior seemed to be principally determined by intentions. Breast or testicle self-examination has been the focus of a couple of studies (McCaul, Sandgren, O'Neill, & Hinsz, 1993; Young, Lierman, Powell-Cope, & Kasprzyk, 1991). McCaul et al. (1993) showed the TRA components to predict breast/testicle self-examination intentions and behaviors, with PBC adding significantly to predictions of intentions but not behavior. A range of other behaviors have been examined using the TPB, including health screening attendance (DeVellis, Blalock, & Sandler, 1990; Norman & Conner, 1993), food choices (Beale & Manstead, 1991; Sparks & Shepherd, 1992), kidney donation (Borgida, Conner, & Manteufel, 1992), drug compliance (Hounsa, Godin, Alihonou, & Valois, 1993), patient education (Kinket, Paans, & Verplanken, 1992), and weight control (Netemeyer, Burton, & Johnston, 1991; Schifter & Ajzen, 1985). The published studies applying the TRA have been reviewed by Sheppard, Hartwick, and Warshaw (1988) and van den Putte (1993), with Ajzen (1991) reviewing 16 studies using the TPB. The findings are generally supportive of the TRA/TPB. Ajzen (1991) reports the multiple correlation between intentions and attitude, subjective norm and PBC to be 0.71 across the 16 studies he reviewed. Van den Putte (1993) computes a value of r = 0.64, but notes the large variation in results between behaviors. Ajzen reports the mean correlation between intentions, PBC and behavior to be 0.51, while van den Putte computes a value of 0.46. Ajzen (1991) and Madden, Ellen, and Ajzen (1992) report empirical evidence that PBC significantly improves predictions of both intentions and behavior. Hence, in summary, the evidence is broadly supportive of the TPB in helping to understand and predict health behaviors. The relative importance of the different predictors is

with the specific wishes of the referent on this issue. These products are then summed across salient referents. Judgments of perceived behavioral control are influenced by beliefs concerning access to the necessary resources and opportunities to perform the behavior successfully, weighted by the perceived power of each factor (Ajzen, 1988, 1991). The perception of factors likely to facilitate or inhibit the performance of the behavior are referred to as control beliefs. These factors include both internal control factors (information, personal deficiencies, skills, abilities, emotions) and external control factors (opportunities, dependence on others, barriers). People who perceive they have access to the necessary resources and perceive that there are the opportunities (or lack of obstacles) to perform the behavior are likely to perceive a high degree of behavioral control (Ajzen, 1991). Ajzen (1991) has suggested that each control factor is weighted by its perceived power to facilitate or inhibit performance of the behavior. The model quantifies these beliefs by multiplying the frequency or likelihood of occurrence of the factor by the subjective perception of the power of the factor to facilitate or inhibit the performance of the behavior. So, according to the TPB, individuals are likely to follow a particular health action if they believe that the behavior will lead to particular outcomes which they value, if they believe that people whose views they value think they should carry out the behavior, and if they feel that they have the necessary resources and opportunities to perform the behavior. 8.01.3.3.2 Review of research The TPB has been applied to the prediction of a number of different behaviors including health-relevant behaviors with varying degrees of success (Ajzen, 1991; see Conner & Sparks, 1996 for a review of the application of the TPB to health behaviors). For example, smoking has been a focus of several studies (Babrow, Black, & Tiffany, 1990; Godin, Valois, Lepage, & Desharnais, 1992). Godin et al. (1992) looked at the prediction of the frequency of smoking in the general public over a six-month period. The prediction of intentions was significantly improved by the addition of the PBC component, and actual smoking behavior appeared to be primarily related to PBC. One study has examined drinking alcohol (Schlegel, D'Avernas, Zanna, & DeCourville, 1992) and found that PBC contributed to the predictions of intentions but not the frequency of getting drunk in nonproblem drinkers, while in problem drinkers, the PBC also contributed to predictions of frequency of getting drunk.

Understanding the Basis of Health Behavior largely an empirical matter. However, in terms of predictors of intentions, it has been argued that attitude may be more important than subjective norms for health behaviors performed in private (e.g., breast self-examination), while subjective norm may be more important than attitudes where the behavior is performed in public (e.g., safety helmet use) (Quine, Rutter, & Arnold, 1998). 8.01.3.3.3 Commentary The TPB has been widely tested and successfully applied to the understanding of a variety of behaviors (for reviews see Ajzen, 1991; Conner & Sparks, 1996; Sheppard et al., 1988). The theory incorporates a number of important cognitive variables which appear to determine health behaviors (intentions, expectancy values, perceived behavioral control). Also the role of social pressure from others is incorporated in the model in the form of subjective norms. However, perhaps because the model was developed outside the health arena, the model does not make an assessment of health threat as is included in models such as the HBM. Finally, the theory states a clear causal ordering among variables in how they relate to behavior, allowing sophisticated analysis techniques to be applied to assessing the model. Sheppard et al. (1988), in a review of the TRA, have argued for the need to consider both behavioral intentions and self-predictions when predicting behavior. Warshaw and Davis (1985) noted a number of different ways in which intentions had been measured, and distinguished measures of behavioral intentions (e.g., I intend to perform behavior x) and from measures of self-predictions (e.g., How likely is it that you will perform behavior x?). This distinction is important when considering the prediction of health behavior because while, for example, David might intend to quit smoking, he might also think that it is unlikely that he will do so. Sheppard et al. went on to argue that self-predictions should provide better predictions of behavior as they are likely to include a consideration of those factors which may facilitate or inhibit performance of a behavior as well as a consideration of the likely choice of other competing behaviors. Sheppard et al.'s meta-analysis supports this view; measures of self-predictions were found to have stronger relationships with behavior than behavioral intentions. However, Norman and Smith (1995) found no difference in the extent to which the two measures correlated with exercise behavior. Furthermore, the measures of behavioral intentions and self-predictions are strongly correlated, suggesting that the discri-

11

minant validity for the two concepts may be weak. Bagozzi (1992) notes that the causal path may begin with the formation of desires which then develop into intentions, which in turn inform self-predictions. However, Conner and Sparks (1996) note that while theoretically these concepts may be distinguishable, empirically there is little to distinguish the three concepts. Clearly, more work is needed to further disentangle these and other related constructs that have appeared in the literature such as planning and commitment (Bagozzi, 1992, 1993), need to change (Paisley & Sparks, 1998), and behavioral willingness (Gibbons, Gerrard, Ouelette, & Burzette, 1998). Self-efficacy is a powerful predictor of the performance of a range of health behaviors (Schwarzer & Fuchs, 1996). Ajzen (1991) argues that the PBC and self-efficacy constructs are interchangeable. However, several authors (e.g., Terry & O'Leary, 1995) have suggested that selfefficacy and PBC are not entirely synonymous. Bandura (1986), for example, argues that control and self-efficacy are quite different concepts. Although Bandura (1986) accepts that some external factors (e.g., task difficulty) will have an influence on self-efficacy, it may still be argued that self-efficacy is more concerned with perceptions of control based on internal control factors. In contrast, PBC is likely to reflect more external factors, and may be more usefully described as perceptions of control over the behavior (Armitage & Conner, in press). For example, Terry and O'Leary (1995) measure self-efficacy over exercising by items such as For me to exercise would be . . . easydifficult and perceived control over the behavior by How much control do you have over exercising? no controlcomplete control. De Vries, Dijkstra, and Kuhlman (1988) have advocated the use of measures of self-efficacy as opposed to PBC in the prediction of intentions and behavior. Further, Dzewaltowski et al. (1990), in a comparison of the TRA, TPB, and Bandura's (1986) Social Cognitive Theory, found that self-efficacy rather than PBC had a direct impact on behavior. Terry and O'Leary (1995) examined exercise behavior and found that self-efficacy only predicted intentions, while PBC had main and interactive effects on exercise behavior. Crucially, a combined measure of PBC and self-efficacy failed to moderate the effect of intention on behavior, suggesting that the two constructs are not synonymous. This issue warrants further empirical study. It seems plausible, however, that perceptions of control and self-efficacy are two separable constructs which not only have differential effects on intentions and behavior but may act differently for different behaviors.

12

Health Behavior three dimensions. These measure the extent to which individuals believe their health is a function of their own actions (i.e., internal HLOC), the actions of powerful others such as health professionals (i.e., powerful others HLOC), and the influence of chance or fate (i.e., chance HLOC). According to HLOC theory, individuals who have strong internal HLOC beliefs should be more likely to engage in health-promoting behaviors. Conversely, those who believe that their health is due to chance or fate should be less likely to engage in healthpromoting behaviors. The prediction for powerful others HLOC is less clear cut. Strong powerful others HLOC beliefs may reflect a receptivity to health messages endorsed by health professionals. Alternatively, strong powerful others HLOC beliefs may indicate a strong belief in the ability of health professionals to cure subsequent illnesses and may be unrelated or negatively related to the performance of health-promoting behaviors. According to social learning theory, the above relationships should only hold for individuals who place a high value on their health as behavior is a function of both expectancy beliefs (e.g., HLOC) and the value attached to certain outcomes (e.g., health value). As K. A. Wallston (1991) argues, individuals are unlikely to engage in health-promoting behaviors if they place a low value on their health, whatever their HLOC beliefs. Thus, health value should act as a moderator of the relationship between HLOC and health behavior. 8.01.3.4.2 Review of research The majority of the research using the HLOC construct has correlated HLOC beliefs with the performance of health behavior, without paying attention to the potential moderating role of health value. This may be due to a lack of appreciation of the complexity of social learning theory (Wallston, 1991) and/or an unchallenged assumption that all people value their health (Lau, Hartman, & Ware, 1986). Several studies have examined the relationship between HLOC beliefs and general indices of health behavior. Most of these have reported a positive correlation between internal HLOC beliefs and the performance of health-promoting behaviours (Duffy, 1988; Mechanic & Cleary, 1980; Seeman & Seeman, 1983; Waller & Bates, 1992; Weiss & Larsen, 1990), although other studies have failed to find such a relationship (Brown, Muhlenkamp, Fox, & Osborn, 1983; Muhlenkamp, Brown, & Sands, 1985; Norman, 1995; Steptoe et al., 1994; Wurtele, Britcher, & Saslawsky, 1985). Some studies have found a negative relationship

The TPB is correctly regarded as a theory of the proximal determinants of behavior. Indeed, Ajzen (1991) describes the model as open to further elaboration if further important proximal determinants are identified. A number of potential candidate variables for addition to the TRA/TPB have been suggested. In each case both theoretical and empirical justifications are necessary (Fishbein, 1993). Some of the most promising of these additional variables are considered in Section 8.01.5.2. 8.01.3.4 Health Locus of Control 8.01.3.4.1 Model description The HLOC construct is one of the most widely researched constructs in relation to the prediction of health behavior (K. A. Wallston, 1992). Its origins can be traced back to Rotter's (1954) social learning theory which states that the likelihood of a behavior occurring in a given situation is a function of the individual's expectancy that the behavior will lead to a particular reinforcement and the extent to which the reinforcement is valued. As well as being applied on a specific level, Rotter argued that social learning theory could be applied on a general level such that individuals may have generalized expectancy beliefs which cut across situations. It was at this generalized level that Rotter introduced the distinction between internal and external locus of control orientations, with internals believing that events are a consequence of their own actions and thereby under personal control and externals believing that events are unrelated to their actions and thereby beyond their personal control. Early work examining the relationship between locus of control and health behavior with Rotter's (1966) InternalExternal (IE) scale produced mixed results. However, the IE scale was criticized for being too generalized to predict health behavior and for conceptualizing locus of control as a unidimensional construct. In particular, Levenson (1974) argued that internal locus of control beliefs are orthogonal to external locus of control beliefs, and that within external locus of control a distinction can be made between external control exerted by powerful others and the influence of chance or fate. The development of the multidimensional health locus of control (MHLC) Scale (K. A. Wallston et al., 1978) addressed both these criticisms, and has since become the most popular locus of control measure in research on health behavior (Wallston & Wallston, 1981, 1982). The MHLC scale measures generalized expectancy beliefs with respect to health along

Understanding the Basis of Health Behavior between chance HLOC beliefs and health behavior indices (Brown et al., 1983; Duffy, 1988; Muhlenkamp et al., 1985; Steptoe et al., 1994). Finally, powerful others HLOC beliefs have rarely been found to predict the performance of health-promoting behaviors (Brown et al., 1983; Duffy, 1988; Muhlenkamp et al., 1985; Steptoe et al., 1994; Waller & Bates, 1992; Weiss & Larsen, 1990). The above results are generally in line with predictions, although the relationship between HLOC and health behavior is typically a weak one. A similar pattern of results is obtained for the relationship between HLOC beliefs and specific health behaviors. For example, studies have reported a link between internal HLOC beliefs and exercise (Carlson & Petti, 1989; O'Connell & Price, 1982; Slenker, Price, & O'Connell, 1985), while other studies have found only a weak link or no link (Burk & Kimiecik, 1994; Calnan, 1989; Liao, Hunter & Weinman, 1995; Norman, 1990, 1995; Speake, Cowart, & Stephens, 1991). In a large-scale representative sample, Calnan (1989) found significant negative correlations between the powerful others and chance HLOC dimensions and exercise. In relation to alcohol consumption, a number of early studies used the locus of control construct to compare alcoholics with nonalcoholics producing mixed results (Butts & Chotlas, 1973; Costello & Manders, 1974). More recent work with the HLOC construct has produced similar mixed results with a number of studies finding no relationship between HLOC beliefs and drinking behavior (Dean, 1991; Liao et al., 1995; Norman, 1990, 1995) and Calnan (1989) only finding weak negative correlations between powerful others and chance HLOC beliefs and alcohol consumption. A number of studies have applied the HLOC construct to the prediction of smoking cessation. A few of these studies have found internal HLOC beliefs to be related to smoking cessation (Horwitz, Hindi-Alexander, & Wagner, 1985; Rosen & Shipley, 1983). Other studies have failed to find a relationship between internal HLOC beliefs and smoking cessation (Kaplan & Cowles, 1978; Segall & Wynd, 1990; Wojcik, 1988). Smokers who believe that their health is under the control of powerful others might be expected to be more successful in giving up smoking after attending a formal smoking cessation program. However, a couple of studies suggest that strong powerful others HLOC beliefs are related to a greater likelihood of relapse following attendance at a smoking cessation program (Segall & Wynd, 1990; Wojcik, 1988). Studies examining the relationship between HLOC beliefs and the performance of breast

13

self-examination among women have found a positive relationship with internal HLOC beliefs (Redeker, 1989) and a negative relationship with powerful others HLOC beliefs (Hallal, 1982; Nemeck, 1990). The negative relationship with powerful others HLOC beliefs may reflect a belief that breast examination is the responsibility of health professionals. In support of this view, Bundek, Marks, and Richardson (1993) found a positive relationship between gynecological screening including physician breast examination and powerful others HLOC beliefs, and a positive relationship between self breast examination and internal HLOC beliefs. However, other studies have failed to find any relationship between HLOC beliefs and breast self-examination (Lau et al., 1986; Liao et al., 1995; Seeman & Seeman, 1983). Those studies that have tested the moderating role of health value have generally produced positive results. Considering the prediction of indices of health behavior first, Weiss and Larsen (1990) found a significant correlation between internal HLOC beliefs and a health behavior index among individuals placing a high value on their health, but a nonsignificant correlation among individuals placing a low value on their health. Similar results have been reported by a number of researchers (Lau, 1982; Lau et al., 1986; Seeman & Seeman, 1983; K. A. Wallston & Wallston, 1980), although other studies have failed to find evidence for the moderating role of health value (Norman, 1995; Wurtele et al., 1985). Studies looking at the interaction between health value and the powerful others and chance dimensions have generally produced nonsignificant results (Lau et al., 1986; Wurtele et al., 1985 ). A similar pattern of results emerges when the performance of specific health behaviors is considered. Evidence for an interaction between internal HLOC beliefs and health value has been found for a range of behaviors including dietary behavior (Hayes & Ross, 1987), smoking cessation (Kaplan & Cowles, 1978), breast self-examination (Lau et al., 1986), and information seeking (K. A. Wallston, Maiders, & Wallston, 1976). However, other studies have failed to find evidence for such an interaction when considering cancer-preventive behavior (McCusker & Morrow, 1979), exercise (Burk & Kimiecik, 1994), attendance at health checks (Norman, 1991), and information seeking (De Vito, Bogdanowicz, & Reznikoff, 1982). 8.01.3.4.3 Commentary Reviews of research with the HLOC construct have concluded that HLOC is a weak predictor of health behavior, even when

14

Health Behavior diet, etc.). As a result, the development of behavior-specific HLOC scales has been advocated. Georgiou and Bradley's (1992) smokingspecific locus of control scale is a good example of such a scale. This scale was found to have stronger correlations with smokers' behaviors and intentions than the more generalized MHLC scale. Other scales have been developed in relation to exercise (Burk & Kimiecik, 1994), AIDS risk behavior (Kelley et al., 1990), weight loss (Saltzer, 1982), alcohol use (Donovan & O'Leary, 1978) and a range of specific conditions including diabetes (Bradley et al., 1990), arthritis (Nicassio, Wallston, Callahan, Herbert, & Pincus, 1985), cancer (Prwun et al., 1988), hypertension (Stanton, 1987) and heart and lung disease (Allison, 1987). Generally, these scales have been found to be more predictive of health behavior than more generalized measures (Lefcourt, 1991). In conclusion, the amount of variance in health behavior explained by the HLOC construct is low, even when considered in conjunction with health value (Norman & Bennett, 1996; K. A. Wallston, 1991, 1992). Nevertheless, the pattern of results obtained are generally in line with predictions suggesting that HLOC beliefs may have a distal influence on health behavior. In line with this position, K. A. Wallston (1992) has proposed a modified social learning theory in which health behavior is a function of HLOC beliefs, health value, and self-efficacy. Importantly, internal HLOC beliefs are seen to be a necessary, but not sufficient, condition for performing a health behavior. To date, there have been no formal tests of this modified theory, although the attempt to embed HLOC beliefs into a broader theoretical perspective is an encouraging development. 8.01.3.5 Protection Motivation Theory 8.01.3.5.1 Model description Protection motivation theory (Rogers, 1983) was originally developed as a framework for understanding the effectiveness of health-related persuasive communications, although more recently it has also been used to predict health protective behavior. It has its origins in early work on the persuasive impact of fear appeals, which was concerned with the conditions under which fear appeals may influence attitudes and behavior. In an extension of the fear-drive model, Janis (1967) proposed that if a persuasive communication successfully arouses fear, usually through emphasizing the severity of a threat and the likelihood of its occurrence, individuals will be motivated to reduce this unpleasant emotional state. If the message also

considered in conjunction with health value (Norman & Bennett, 1996; K. A. Wallston, 1991, 1992). There are a number of reasons for this poor performance which are outlined below. K. A. Wallston and Wallston (1981, 1982) have argued that certain combinations of HLOC beliefs may be important in predicting health behavior. For example, when being advised to quit smoking by a health professional, it may be advantageous to have a combination of strong powerful others and internal HLOC beliefs. K. A. Wallston and Wallston (1981) have therefore proposed a 2 6 2 6 2 typology based on median splits on the three HLOC dimensions (see also Waller and Bates, 1992). Using this typology, believers in control (i.e., high internal high powerful others, low chance HLOC beliefs) have been found to show better adjustment in relation to rheumatoid arthritis (Roskam, 1986) and diabetes (Bradley, Lewis, Jennings, & Ward, 1990). To date, this approach has not been applied to the prediction of health behavior. However, despite the promise of this approach, its utility may be limited given that large sample sizes are required in order to compare all eight types. In addition, this approach may lead to results which are difficult to interpret, given the large number of comparisons. A number of researchers have questioned the way in which health value is measured. Typically, health value has been measured as an absolute value using Lau et al.'s (1986) fouritem scale. However, when deciding whether or not to perform a specific health behavior, individuals are often faced with more appealing alternatives. As a result, values other than health may be important in determining behavior. In such cases it may be more appropriate to use relative measures of health value. Kristiansen (1986) followed this approach and found that a measure of the value placed on health relative to value of an exciting life was more predictive of health behavior among young people than an absolute measure of health value. However, Wurtele et al. (1985) reported the opposite pattern of results in a sample of female undergraduates. Clearly, more work is needed comparing different approaches to the measurement of health value. The need to consider behavior-specific control beliefs has been highlighted by a number of researchers. In particular it is possible to argue that one reason for the relatively poor performance of the HLOC construct is that it measures generalized expectancy beliefs with respect to health. In other words, while HLOC is specific to a given goal (i.e., health), it cuts across many situations (e.g., smoking, exercise,

Understanding the Basis of Health Behavior contains recommendations for action, then one way in which individuals can reduce this state of arousal is to follow the communicator's advice. If the message does not contain effective behavioral advice, then maladaptive coping reactions may follow such as denial or avoidance. Janis proposed that fear appeals may be most effective when a medium level of fear is evoked. Under such conditions the cognitive responses that promote adaptive reactions (e.g., following behavioral advice) outweigh those that promote maladaptive reactions (e.g., denial). However, later work has failed to confirm this hypothesis (Sutton, 1982). Leventhal (1970) made a similar distinction between adaptive and maladaptive reactions in the parallel response model which differentiates between two independent control processes that are initiated by a fear appeal. The first, fear control, focuses on attempts to reduce the emotional threat (e.g., avoidance) while the second, danger control, focuses on attempts to reduce the threatened danger (e.g., following behavioral advice). The parallel response model is important in proposing that protection motivation results from danger control processes (i.e., cognitive responses) rather than from fear control processes (i.e., emotional responses). Roger's (1983) (PMT) outlines the cognitive responses resulting from fear appeals in more detail (see Figure 3). It is argued that various environmental (e.g., fear appeals) and intrapersonal (e.g., personality variables) sources of information can initiate two appraisal processes: threat appraisal and coping appraisal. Threat appraisal, which is similar to Lazarus and Launier's (1978) primary appraisal, focuses on the source of the threat and the factors that may increase or decrease the probability of the maladaptive response. Both the perceived severity of the threat and the individual's perceived vulnerability to the threat are seen to inhibit maladaptive responses. However, there may be a number of intrinsic (e.g., pleasure) and extrinsic (e.g., social approval) rewards which may serve to increase the likelihood of maladaptive responses. Coping appraisal, which is similar to Lazarus and Launier's (1978) secondary appraisal, focuses on one's ability to cope with the threat and the factors that may increase or decrease the probability of an adaptive response. Both the belief that the recommended action will be effective in reducing the danger (i.e., response efficacy) and the belief that one is capable of performing the recommended action (i.e., selfefficacy) are likely to increase the probability of an adaptive response, although various response costs (e.g., financial cost) associated with

15

performing an adaptive response will serve to inhibit such a response. Protection motivation results from the two appraisal processes and is a positive function of beliefs about severity, vulnerability, response efficacy, and self-efficacy, and a negative function of beliefs about the rewards associated with the maladaptive response and the response costs of the protective behavior. Moreover, for protection motivation to be elicited, it is necessary for the rewards associated with the maladaptive response to be outweighed by perceptions of severity and vulnerability, and the response costs of the protective behavior to be outweighed by perceptions of response efficacy and self-efficacy. Protection motivation, which is usually measured by behavioral intentions, is seen to arouse, direct, and sustain protective behavior. 8.01.3.5.2 Review of research Protection motivation theory provides a framework for understanding both the effects of fear appeals and the social cognitive variables underlying health behavior. As a result, tests of PMT have taken two forms. In the first, the key components of PMT are manipulated in persuasive communications and their effects on protection motivation tested (see Section 8.01.4.1). In the second, PMT is considered as a general attitudebehavior model and its components used as predictors of health behavior. This work is reviewed below. One area in which PMT has been used as a general attitudebehavior model is in relation to cancer-related preventive behavior. In one of the few longitudinal tests of PMT, Hodgkins and Orbell (1998) examined the social cognitive predictors of breast self-examination (BSE) in a sample of young women (1740 year olds) over a one-month period. Each of the main components of PMT was measured (i.e., severity, vulnerability, response efficacy, self-efficacy) as well as the response costs of performing BSE (e.g., I would feel awkward examining my breasts). In a path analysis, only self-efficacy was related to intentions to perform BSE. Time one behavioral intention was in turn found to be the most important predictor of performance of BSE at one month follow-up. Similar results have been reported by Seydel, Taal, and Wiegman (1990) who found response efficacy and self-efficacy to be predictive of intentions to engage in, and concurrent performance of, a number of cancer-related preventive behaviors. Boer and Seydel (1996) also found response efficacy and self-efficacy to be predictive of intentions to participate in mammography screening.

16

Health Behavior

SOURCES OF INFORMATION
Environmental Fear appeals Observation

Threat appraisal Perceived vulnerability

Perceived severity

Intrinsic rewards Intrapersonal Personality Prior experience Coping appraisal Response costs Other external variables Age, sex, occupation, socioeconomic status, religion, education Response efficacy Extrinsic rewards Protection Motivation Behavior

Self-efficacy
Figure 3 Protection motivation theory.

A second area of application of PMT has been in relation to AIDS risk-reducing behaviors such as condom use. Aspinwall, Kemeny, Taylor, Schneider, and Dudley (1991) examined the ability of the PMT components to predict reductions in a number of AIDS risk-reducing behaviors in a sample of gay men over a sixmonth period. The results showed that strong levels of self-efficacy and a high level of perceived vulnerability at time one were predictive of a reduction in the number of sexual partners over the six-month follow-up period. In addition, self-efficacy also emerged as the most important predictor of reductions in the number of anonymous sexual partners. Van der Velde and van der Pligt (1991) used PMT as a framework for assessing the coping responses of heterosexual men and women and homosexual men with multiple partners. Considering the heterosexual sample first, it was found that vulnerability, response efficacy, and self-efficacy all had a direct positive effect on behavioral intentions to use condoms. In addition, severity had an indirect effect on

behavioral intentions through a measure of fear. Similar results were found with the homosexual sample with response efficacy, self-efficacy, and severity having a positive effect on behavioral intentions to engage in safe sex. However, contrary to expectations, a negative relationship was found between vulnerability and behavioral intentions. In a sample of male and female adolescents, Abraham, Sheeran, Abrams, and Spears (1994) found that self-efficacy had a positive influence and response costs (i.e., concern about reputation) had a negative influence on behavioral intentions to use a condom. In addition, vulnerability had a negative effect on behavioral intentions to limit the number of sexual partners. In a cross-sectional study of male and female heterosexuals, Bengel, Beltz-Merk, and Farin (1996) found that self-efficacy was related to a greater use of condoms and fewer sexual partners. However, perceptions of vulnerability were related to greater use of condoms, but also higher number of sexual partners.

Understanding the Basis of Health Behavior Taken together, the above results suggest that PMT is a useful framework for understanding HIV-related protective behavior. Self-efficacy emerges as the most important predictor of such behavior, with response efficacy and severity also emerging as significant predictors in some studies. A conflicting pattern of results has been found with the vulnerability component. Similar conclusions have been reached by Farin (1994) in a meta-analysis of PMT and HIVprotective behavior, in which self-efficacy and response efficacy emerged as the best predictors of protective behavior. However, these two components were only able to explain 2.2% and 1.8% of the variance in such behavior. Severity was seen to be less important, and vulnerability had a conflicting pattern of results. 8.01.3.5.3 Commentary Despite relatively few studies in the area, the above review highlights the potential utility of PMT as a framework for considering the social cognitive predictors of health protective behavior. However, there are a number of issues which future work needs to address. First, some studies have reported a positive relationship between perceptions of vulnerability and protection motivation, while others have reported a negative relationship. Seydel et al. (1990) suggest that the negative relationship between vulnerability and intentions to engage in cancer-related preventive behavior in their study may be due to a defensive avoidance style of coping, in which perceptions of vulnerability to cancer may lead to feelings of anxiety which may inhibit adaptive responses and promote avoidance. However, it is more likely that the mixed pattern of results is due to measurement issues. As Weinstein and Nicolich (1993) argue, a negative correlation may be expected between perceptions of vulnerability and concurrent protective behavior given that one's current behavior may be used to make vulnerability judgments. In contrast, a positive correlation may be expected between perceptions of vulnerability and future protective behavior to the extent that perceptions of vulnerability motivates protective behavior. Considering behavioral intentions, individuals may feel vulnerable and therefore intend to engage in a protective behavior (i.e., positive correlation), or may feel vulnerable because they do not intend to engage in a protective behavior (i.e., negative correlation). Clearly, more consideration needs to be given to the measurement of perceptions of vulnerability and the potential use of conditional measures of risk. Second, while PMT provides a framework for considering the social cognitive predictors of

17

protective behavior, a number of researchers have commented on its lack of specification in terms of the nature of its components and the relationships between them (Bengel et al., 1996). For example, Rogers (1983) states that the response costs of the protective behavior need to be outweighed by perceptions of response efficacy and self-efficacy for protection motivation to be elicited. However, no guidance is given as to how these variables are to be measured and combined in order to predict protection motivation. Overall, PMT identifies many of the social cognitive variables which are important predictors of health behavior. It shares a number of similarities with the HBM (i.e., measures of perceived susceptibility, severity, benefits, and barriers) although it also includes self-efficacy, which has been found to be one of the most powerful explanatory constructs in relation to health behavior (Schwarzer & Fuchs, 1996), and a measure of behavioral intention, which is seen to mediate the influence of threat appraisal and coping appraisal. It is also important for providing a synthesis between social cognitive approaches and coping models as outlined by Lazarus and Launier (1978). However, to date there have been relatively few longitudinal tests of PMT in relation to health behavior. Such studies should help clarify some of the issues raised above. 8.01.3.6 Self-efficacy 8.01.3.6.1 Model description Self-efficacy is one of the most powerful predictors of health behavior (K. A. Wallston, 1992). It has its origins in Bandura's (1977) social cognitive theory which states that behavior is a function of both incentives (i.e., reinforcements) and expectancies. Three kinds of expectancies can be identified, these being situation-outcome expectancies which refer to beliefs about how events are connected, outcome expectancies which refer to beliefs about the consequences of performing a behavior, and self-efficacy expectancies which refer to beliefs about one's ability to perform the behavior. Thus in order to perform a health behavior, individuals must value their health (i.e., incentive), believe that their current lifestyle poses a threat to their health (i.e., situation-outcome expectancy), believe that adopting the new behavior will reduce the threat to their health (i.e., outcome expectancy) and believe that they are capable of performing the behavior (i.e., self-efficacy expectancy). While all these beliefs are seen to be important in the initiation and maintenance of health behavior, self-efficacy

18

Health Behavior Wandersman, 1991; Kok, de Vries, Mudde, & Strecher, 1991; Morrison, Gillmore, & Baker, 1995). While the majority of studies examining the relationship between self-efficacy and AIDS risk-reducing behaviors have reported significant results, a few studies have failed to do so (Boyd & Wandersman, 1991; Morrison et al., 1995). Considering exercise behavior, selfefficacy has been found to be an important predictor of both intentions to engage in regular exercise and actual exercise behavior (Desharnais, Bouillon, & Godin, 1986; McAuley, 1993; Sallis, Howell, Hofsteffer, & Barrington, 1992). Finally, self-efficacy has also been related to dieting and weight control (Bernier & Avard, 1986; Hofstetter, Sallis, & Howell, 1990; Jeffrey et al., 1984). For example, Jeffrey et al. examined the relationship between self-efficacy and weight loss following participation in a behavioral treatment program over a two-year period. Making a distinction between emotional self-efficacy (i.e., confidence in one's ability to refrain from eating during various emotional states) and situational self-efficacy (i.e., confidence in one's ability to refrain from eating in various situations), they found pretreatment measures of both types of self-efficacy to be predictive of weight loss at one and two years. However, post-treatment measures of self-efficacy were not so powerful with only situational self-efficacy predictive of weight loss at one year. 8.01.3.6.3 Commentary The self-efficacy construct has been found to be one of the most important predictors of health behavior (K. A. Wallston, 1992). However, there are a number of issues which future work needs to address. First, the relationship between outcome and self-efficacy expectancies has been a source of some debate (Corcorcan, 1991; Kirsch, 1986, Maddux, 1993). As Schwarzer (1992) argues, there may be a temporal and causal order among the two types of expectancy beliefs inasmuch as individuals are unlikely to consider their ability to perform a behavior before first considering the efficacy of the behavior. Bandura (1991) argues that the effect of outcome expectancies on intentions and behavior are partly governed by self-efficacy expectancies; even if outcome expectancy is high, performance of a behavior is unlikely if self-efficacy is low. As a result, when selfefficacy is partialed out, any relationship between outcome expectancy and behavior should disappear. However, some research has indicated that outcome and self-efficacy expectancies can be independent predictors of intentions and behavior (Maddux, 1993).

expectancies are seen to be the most important. Individuals with strong self-efficacy beliefs are believed to develop stronger intentions to act, to expend more effort to achieve their goals, and to persist longer in the face of barriers and impediments (Bandura, 1991). Self-efficacy beliefs are therefore believed to play a crucial role in the determination of health behavior. According to Bandura (1977, 1982), such beliefs can be conceptualized and measured in terms of three parameters; magnitude, strength, and generality. The first parameter refers to the level of difficulty of the behavior. Individuals with low-level expectations feel capable of performing only very simple behaviors, whereas individuals with high-level expectations feel capable of performing even the most difficult of behaviors. In this way it is possible to assess individuals' expectations about their level, or magnitude, of performance. The second parameter refers to individuals' confidence that they could perform a specific behavior, while the third parameter refers to the generality of expectations across situations or domains. The measurement of self-efficacy usually focuses on the strength of the selfefficacy expectation (e.g., I am confident that I can refrain from smoking), although it will often incorporate the magnitude of expectation (e.g., I am confident that I can refrain from smoking, even if someone offers me a cigarette). 8.01.3.6.2 Review of research The self-efficacy construct has been successfully applied to the prediction of a range of health behaviors (see Bandura, 1991; O'Leary, 1985; Schwarzer & Fuchs, 1996). One of the main areas of application has been in relation to smoking cessation (Condiotte & Lichtenstein, 1981; Colletti, Supnick, & Payne, 1985; Kavanagh, Piere, Lo, & Shelley, 1993). These studies have found self-efficacy to be a consistent predictor of smoking cessation. For example, Condiotte and Lichtenstein (1981) found that post-treatment self-efficacy beliefs were predictive of both the probability of relapse and the amount of time before relapse. In addition, a close correspondence was noted between selfefficacy beliefs for a range of specific tempting situations (e.g., after a meal, when drinking coffee) and the actual situation in which the relapse occurred. The results therefore indicate that self-efficacy is important not only in predicting the likely success of smokers who are trying to quit, but also the situations in which they are most likely to relapse. Self-efficacy has also been related to AIDS risk-reducing behaviors, such as condom use (Basen-Engquist & Parcel, 1992; Boyd &

Encouraging the Adoption of Health Behaviors Second, the role of incentives, or outcome values, has tended to be overlooked in research with the self-efficacy construct. This is despite the fact that a large body of research in expectancy value theory has found outcome value to be an important predictor (Kirsch, 1986; McCelland, 1985). Unfortunately, those studies which have considered the role of outcome and self-efficacy expectancies in conjunction with incentives or outcome values have produced mixed results (Maddux, Norton, & Stoltenberg, 1986; Manning & Wright; 1983). Third, there is some disagreement over whether it is appropriate to consider generalized self-efficacy expectancies. According to Bandura's (1977) original conceptualization, selfefficacy beliefs should focus on specific behaviors in specific situations. Self-efficacy is not seen to be a personality trait as self-efficacy beliefs can be seen to vary across behaviors and situations, although self-efficacy beliefs may generalize to other behaviors and situations to the extent that the new behaviors require similar skills and the new situations have similar features (Bandura, 1986). Nevertheless, Schwarzer and Fuchs (1996) have argued that self-efficacy can be viewed as a generalized trait reflecting a personal resource factor, pointing to studies which have successfully employed generalized measures of self-efficacy to predict behavior (Mittag & Schwarzer, 1993; Snyder et al., 1991; K. A. Wallston, 1992). Generalized self-efficacy may be closely related to dispositional optimism (Scheier & Carver, 1992), although on an empirical level Schwarzer (1994) found a correlation of only 0.60 between dispositional optimism and generalized selfefficacy. In conclusion, self-efficacy has been found to be one of the most powerful and consistent predictors of health behavior. For this reason alone, it is not surprising that it has been incorporated into a number of the main social cognition models of health behavior; for example, Rogers' (1983) PMT and Ajzen's (1991) TPB. In addition, Rosenstock, Strecher, and Becker (1988) have called for the inclusion of self-efficacy in the HBM. It is likely that selfefficacy will continue to attract considerable interest and continue to be a key predictor of health behavior. 8.01.4 ENCOURAGING THE ADOPTION OF HEALTH BEHAVIORS 8.01.4.1 Intervention Studies The models presented earlier in this chapter outline some of the key social cognitive determinants of health behavior. In this section

19

we consider the use of these models in the design of interventions to encourage new health behaviors. As Fishbein (1993) has argued, the ultimate test of the utility of these models lies in their ability to inform the design of effective interventions. To the extent that these models outline the key social cognitive determinants of health behavior, interventions which target these variables should lead to associated changes in behavior. However, to date there have been relatively few theoretically driven interventions. As a result there is still a need for more tests of these models in action (Fishbein, 1993). Brawley (1993) argues that we need to take account of the practicality of employing SCMs when designing interventions. In short, a model which offers a high level of practicality must be shown to have predictive utility, to describe the relationships between key constructs, to offer guidelines for the assessment of these constructs, to allow the translation of these constructs into operational manipulations, and to provide the basis for detecting the reasons why an intervention succeeds or fails. The SCMs considered here can be been seen to have a high level of practicality and therefore should provide a good framework for the design of effective interventions. One model which has been used widely to design interventions is the PMT. A good example of the use of PMT in this respect is provided by Wurtele and Maddux (1987) in their study on exercise intentions and behavior. In this study essays recommending beginning a regular exercise program were presented to a sample of nonexercising female undergraduates. The essays were designed so that each of the PMT's main components (i.e., severity, vulnerability, response-efficacy, and self-efficacy) were independently manipulated, resulting in a 2626262 between-subjects factorial design with two levels (present vs. absent) of each factor. For example, the severity message emphasized the seriousness of the threat of a sedentary lifestyle by vividly describing the negative effects of a heart attack (e.g., Nauseated, the victim vomits; pink foam comes out of the mouth. The face turns ashen grey, sweat rolls down the face . . .). After reading an essay, subjects completed a questionnaire containing measures of the main components of PMT and were followed up two weeks later to chart any changes in exercise behavior. Manipulation checks revealed that each of the messages successfully manipulated their corresponding component from PMT, as is the case in most PMT intervention studies. Of more interest was the effect of the messages on protection motivation as measured by

20

Health Behavior that high levels of arousal or anxiety may indicate to the individual that he or she is not capable of performing a given action. As a result, relaxation techniques may be employed to help maintain feelings of self-efficacy. Each of these techniques have been used in intervention studies to try to enhance feelings of self-efficacy. Maibach, Flora, and Nass (1991) report the results of a year-long community health campaign to encourage the adoption of health behaviors. The campaign materials were all designed to reflect the main principles of Bandura's (1986) social cognitive theory and used a number of strategies for enhancing feelings of self-efficacy and encouraging behavior change attempts. These included encouraging participants to set behavior change goals, using community members who had successfully changed their behavior as role models, using health experts to give advice about behavior change and focusing on the skills needed to support behavior change. The campaign was found to successfully increase feelings of self-efficacy which, in turn, were related to the adoption of new health behaviors. A number of interventions have focused on more situation-specific feelings of self-efficacy. Stevens and Hollis (1989) designed an intervention to help smokers quit smoking which built on the results of earlier research which had shown that situation-specific ratings of selfefficacy were predictive of the circumstances in which relapses occurred (Condiotte & Lichtenstein, 1981). Abstinent smokers have identified potential relapse situations in which they perceived low levels of self-efficacy and then developed and rehearsed appropriate coping strategies over three weekly meetings. This intervention led to a greater abstinence rate at one year than both a discussion-only intervention and a no-treatment control. Other studies which have attempted to improve behavioral skills to enhance feelings of self-efficacy have produced positive results in relation to alcohol use (Baer et al., 1992) and dental hygiene (McCaul, Glasgow, & O'Neil, 1992). Few studies have attempted to use the TRA/ TPB as a framework for developing interventions, despite quite clear guidelines outlined by Ajzen and Fishbein (1980). In fact, van den Putte (1993) reports a mere five studies which have followed such an approach, with only limited evidence for success compared to approaches not based upon the model. For example, Brubaker and Fowler (1990) examined the effect of persuasive messages upon men's intentions to perform testicular self-examination. A persuasive message based on the theory of reasoned action was found to increase intentions to perform testicular self-examination compared

behavioral intentions to start a regular exercise program. The results showed that only the vulnerability and self-efficacy messages had a significant effect on behavioral intentions, although a three-way interaction between vulnerability, response efficacy, and self-efficacy was also found. In relation to changes in exercise behavior over the two-week follow-up period, a significant interaction was found between severity and self-efficacy, such that the self-efficacy message only had a significant effect when the severity message was absent. Overall, studies which have used PMT to design interventions have shown that selfefficacy is the most powerful component that can be manipulated in persuasive messages (Boer & Seydel, 1996). In addition, manipulating response efficacy has been found to have a significant effect on intentions in a majority of studies. These two variables have been shown to influence intentions to engage in a range of health behaviors including exercise (Stanley & Maddux, 1986), quit smoking (Maddux & Rogers, 1983), dietary intake (Wurtele, 1988), and breast self-examination (Rippetoe & Rogers, 1987). Manipulating perceptions of vulnerability has been found to influence intentions in other studies (Maddux & Rogers, 1983; Wurtele & Maddux, 1987), while manipulating perceptions of severity has not been found to have an influence on behavioral intentions. However, many PMT intervention studies have combined the vulnerability and severity components so that the potential threat of a maladaptive behavior is emphasized. This has been found to have a significant effect on intentions in relation to alcohol use (Stainback & Roger, 1983), dental flossing (K. H. Beck & Lund, 1981), dietary behavior (Wurtele, 1988), information seeking (Brouwers & Sorrentino, 1993), and breast self-examination (Rippetoe & Rogers, 1987). Some studies have focused more specifically on enhancing feelings of self-efficacy as a means for encouraging health behavior change. As Bandura (1986) outlines, there are four main sources of self-efficacy, each of which could be addressed in interventions. First, individuals can develop feelings of self-efficacy from personal mastery experience. For example, it may be possible to split a behavior into various subgoals, such that the easiest subgoals are achieved before more difficult tasks are attempted. Second, individuals may develop feelings of self-efficacy through observing other people succeed on a task (i.e., vicarious experience). Third, it is possible to use standard persuasive techniques to try to instil feelings of self-efficacy. Finally, one's physiological state may be used as a source of information, such

Limitations and Extensions of Existing Models to a no-message control, but was no more effective than a knowledge-only message. More recently, Parker, Manstead, and Stradling (1996) have developed intervention videos based on the TPB to discourage speeding in residential areas by car drivers. The results showed that the normative belief video had a significant effect on a postintervention measure of normative beliefs, while the behavioral belief video had no effect on behavioral beliefs, and the perceived behavioral control video had a negative effect on perceptions of control. Overall, the videos had no effect on subjects' expectations of speeding in residential areas over the next year. Finally, there is some evidence that tailoring interventions to fit in with individuals' existing belief orientations may lead to more effective interventions. Chambliss and Murray (1979a) devised a weight control program in which participants were given placebo medication to help control their metabolism. After two weeks participants in one group were debriefed about the placebo medication and encouraged to attribute any weight loss to their own efforts over the previous two weeks. Participants in a second group were given further information about the efficacy of the medication and encouraged to attribute any weight loss to the medication. At two-week follow-up, a significant interaction was found between the giving of information and participants' preprogram locus of control orientation, such that internals lost more weight than externals in the self-efficacy information group, while the opposite pattern of results was found for the drug information group. Similar results have been reported by Chambliss and Murray (1979b) in relation to smoking cessation. These results led to further work exploring the match between the control orientation of the intervention and participants' existing health locus of control beliefs. For example, Quadrel and Lau (1989) found an interaction between health locus of control beliefs and the control orientation of a message to encourage breast self-examination among female students. In particular, those females with strong internal health locus of control beliefs who received a message in a control frame were more likely to perform breast selfexamination at follow-up, although this effect was reversed if a neutral reminder was sent. Further evidence for a matching hypothesis has been provided in relation to weight reduction (B. S. Wallston, Wallston, Kaplan, & Maides, 1976) and smoking cessation (Best, 1975). Despite offering a high level of practicality, there have been relatively few studies testing SCMs in action. As the above review shows,

21

practical applications of these models have met with some success and suggest that health behavior change interventions may have a lot to gain from using these models as a guiding framework. As Lewin (1951) concludes, there is nothing so practical as a good theory (p. 169). 8.01.5 LIMITATIONS AND EXTENSIONS OF EXISTING MODELS 8.01.5.1 Limitations of Current Models The SCMs described above represent one widely used approach to understanding health behaviors. Here we provide a critique of this approach to understanding health behavior and outline ways in which research might develop through consideration of additional theoretical constructs, the processes by which cognitions influence behavior, stage models of behavior change, potential integrations, and future directions for work in this area. There are several advantages of using social cognition models in health psychology (Conner, 1993; Conner & Norman, 1996b). First, they provide a clear theoretical background to research, guiding the selection of variables to measure, the procedure for developing reliable and valid measures, and how these variables are combined in order to predict health behaviors and outcomes. Second, to the extent that the models identify the important variables in predicting health outcomes and behaviors, they enable us to develop effective behavioral interventions. Third, the models provide us with a description of the cognitive processes determining individuals' motivation to perform different behaviors. There are parallel disadvantages in too exclusive a focus upon social cognition models as the way to understand health behaviors. First, in providing such an explicit general theoretical framework, these models may lead us to neglect variables (cognitive and noncognitive) potentially important in understanding a particular health behavior or outcome. For example, the decision to use a condom may be a function of cognitions, emotional reactions, and also a complex interaction between the individuals involved. Social cognition models on their own are unlikely to provide considerable predictive power in these situations. In addition, SCMs are open to extension when empirically and theoretically justified (Fishbein, 1993). Second, while such models provide us with targets for interventions to produce behavior change, they do not specify how such cognitions are best changed. Effective interventions need to consider both the targets (e.g., cognitions) and the persuasion process itself. This process of

22

Health Behavior social comparison processes and how the social image or prototype of the person who performs a particular behavior influences the performance of various health behaviors (particularly among young people). Other researchers have suggested other forms of normative influence we might consider. These include descriptive norms and personal or moral norms. Descriptive norms are perceptions of the behavior of salient others. For example, Jane's eating behavior may be influenced not only by her perceptions of what others think she should eat, but also perceptions of what they actually do eat. Several studies have reported that perceptions of others' behavior contributed to the prediction of intentions independently of perceived injunctive norms (e.g., Conner et al., 1996; De Vries, Backbier, Kok, & Dijkstra, 1995; Grube, Morgan, & McGee, 1986). Personal, or moral, norms are the individual's perception of the moral correctness or incorrectness of performing a behavior (Ajzen, 1991; Sparks, 1994) and take account of, . . . personal feelings of . . . responsibility to perform, or refuse to perform, a certain behavior (Ajzen, 1991, p. 199). Moral norms should have an important influence on the performance of those behaviors with a moral or ethical dimension (L. Beck & Ajzen, 1991; Gorsuch & Ortberg, 1983; Kurland, 1995). A number of studies have found measures of moral norms to be predictive of blood donating behavior (Pomazal & Jaccard, 1976; Zuckerman & Reiss, 1978) as well as intentions to donate organs (Schwartz & Tessler, 1972), eat genetically produced food (Sparks, Shepherd, & Frewer, 1995), buy milk (Raats, 1992), use condoms (Nucifora et al., 1993), and commit driving violations (Parker, Manstead, & Stradling, 1995). Anticipated affective reactions to the performance or nonperformance of a behavior may also be an important determinant of behavior (Triandis, 1977; Van der Pligt & de Vries, 1998), especially in situations where the consequences of the behavior are unpleasant or negatively affectively laden. In the 1990s research has focused on the influence of anticipated regret (Parker et al., 1995; R. Richard & van der Pligt, 1991; Richard, Van der Pligt, & de Vries, 1995, 1996a, 1996b). It is argued that if individuals anticipate feeling regret after performing a behavior then they will be unlikely to perform the behavior. Richard et al. (1995, 1996a) investigated the role of anticipated regret in relation to condom use among adolescents and found such feelings to be an important predictor of intentions. Richard et al. (1996b) examined the influence of anticipated regret on subsequent behavior. Participants in their study were asked to either focus on their anticipated

persuasion is described by other models of social cognitive processes (e.g., the elaboration likelihood model: Petty & Cacioppo, 1986; the systematic-heuristic model: Chaiken, Lieberman, & Eagly, 1989). In addition, applications of SCMs should not lead to neglect of alternatives to persuasion in producing behavior change such as extrinsic changes to the rewards and costs of a given behavior. For example, increased taxation and legal restrictions can be effective in producing change in health behaviors either in isolation or in tandem with persuasion. Third, although SCMs have furthered our understanding of motivational processes and their influence upon behavior, they have neglected other aspects of behavior change. For example, few of the models consider volitional processes beyond attempting to explain intentions (Bagozzi, 1993; Gollwitzer, 1990). However, many individuals who intend to change fail to do so. Hence, we need to consider the other important volitional processes associated with attempts to change and maintain behavior change (see Norman & Conner, 1996a; Schwarzer & Fuchs, 1996). 8.01.5.2 Additional Theoretical Constructs While the social cognition models outlined here provide an important framework for considering the social psychological determinants of health behavior, it is clear that in some instances they only account for a modest amount of the variance in health behavior. For example, Sheppard et al. (1988) noted that about 10% of studies they reviewed reported correlations between behavioral intentions and behavior below 0.2. This suggests that key variables have failed to be included in these models. We review here the most promising of such variables appearing in the literature: measures of norms, anticipated affective reactions, self-identity, and past behavior. A number of researchers have argued that further attention needs to be paid to the concept of normative influences (e.g., Cialdini, Reno & Kallgren, 1990; Conner, Martin, Silverdale, & Grogan, 1996). Of the major SCMs, only the TPB incorporates perceived social pressures to perform a behavior as a predictor of intentions. Based upon social identification theory, Terry and Hogg (1996) suggest that such (injunctive) normative measures might be more predictive of behavior if they employed a measure of group identification rather than motivation to comply (e.g., I identify with my friends with regard to smoking). Similarly, Gibbons and Gerrard (1995, 1997) have noted the need to consider

Limitations and Extensions of Existing Models feelings following safe and unsafe sexual behavior or on their present feelings about these behaviors. At follow-up, participants in the anticipated feelings condition were more likely to have used condoms in casual sexual encounters in the intervening five months. The effects of anticipated affective reactions have been confirmed in studies of driving (Parker, Manstead, Stradling, Reason, & Baxter, 1992) and consumer behavior (Simonson, 1992). However, in terms of developing social cognition models of health behavior, it is possible to argue that anticipated affective reactions may be incorporated into constructs that focus on the consequences of behavior (e.g., behavioral beliefs in the TPB; see Van der Pligt & de Vries, 1998). The concept of self-identity has also been suggested as a predictor of behavior (Biddle, Bank, & Slavings, 1987; Charng, Piliavin, & Callero, 1988). For example, the extent to which individuals think of themselves as healthy eaters should predict their dietary intentions and behavior. In support, Sparks and Shepherd (1992) found that respondents who thought of themselves as green consumers had stronger intentions to consume organic vegetables. Sparks (1994) noted that self-identity may simply be a proxy for past behavior, although Sparks and Shepherd (1992) found that the relationship between self-identity and future intentions remained when past consumption of organic vegetables was controlled for. Selfidentity as someone who is concerned about the health consequences of one's diet has also been related to intentions to reduce fat consumption (Sparks, Shepherd, Wieringa, & Zimmermanns, 1995), although in an earlier study Sparks, Shepherd, Wieringa, and Zimmermanns (1994) failed to find an independent effect for selfidentity. Role identity (regarded as synonymous with self-identity) was measured by Theodorakis (1994) and found to be a significant predictor of exercise behavior. Future work needs to assess the influence of self-identity across of range of behaviors as it may be the case that self-identity is only important in a restricted range of situations. The influence of past on current behavior in SCMs has attracted much attention. It is argued that many health behaviors are determined by one's previous behavior rather than cognitions. The argument is based on the results of a number of studies showing past behavior to be the best predictor of future behavior. For example, Mullen, Hersey, and Iverson (1987) found initial behavior to be the strongest predictor of the consumption of sweet and fried foods, smoking, and exercise over an eightmonth period. Similar results have been

23

reported in relation to drug use (Bentler & Speckart, 1979; Huba, Wingard, & Bentler, 1981), exercise (Godin, Valois, & Lepage, 1993; Norman & Smith, 1995; Valois, Desharnais, & Godin, 1988), breast self-examination (Hodgkins & Orbell, 1998), attendance at health checks (Norman & Conner, 1993, 1996b) and seat belt use (Sutton & Hallett, 1989). Such results have led to calls for past behavior to be considered as an independent predictor of future behavior (Bentler & Speckart, 1979; Fredricks & Dossett, 1983). However, there are problems with this view. Ajzen (1988) argues that the effects of past behavior on future behavior should be mediated by the variables included in social cognition models; past behavior shapes individuals' beliefs about the behavior in question, and it is these cognitions that determine subsequent behavior. When past behavior is found to have a direct effect on future behavior it is because key social cognitive variables have not been considered (Ajzen, 1991). There has also been focus on the concept of habit (i.e., behaving in a way you have acted before without thinking about it). Eagly and Chaiken (1993) review numerous studies where the addition of habit has added to the prediction of future behavior over and above the influence of variables such as intention, attitude, and subjective norm. Despite this evidence, few SCMs incorporate a measure of habit. A notable exception is Triandis (1977) who argues that it is possible to make a distinction between habitual and intentional behaviors. He argues that novel behaviors will be primarily determined by intention, while repeated behaviors will be primarily determined by habit. Ronis, Yates, and Kirscht (1989) make a similar distinction between habits and decisions, arguing that the performance of repeated behavior is determined by habit rather than social cognitive variables. For example, Dishman (1982) distinguished between the initiation and maintenance of behavior in relation to clinical exercise programs and found that only the initiation of exercise behavior was predicted by social cognitive variables. Sutton (1994) has proposed a further distinction between habits and routines. He argues that many health behaviors commonly considered habitual may be more appropriately considered routines. Sutton (1994) describes a routine as a sequence of behaviors which is repeated on a regular basis. However, what distinguishes them from habits is their need to be supported by selfreminders. It may be possible therefore to make the distinction between occasions when the influence of past behavior is mediated by social cognitive variables and those occasions when it

24

Health Behavior the accessibility of relevant attitudes influences the strength of the relationship between attitudes and behavior (Fazio & Williams, 1986) and that highly accessible attitudes can lead to selective perception (Houston & Fazio, 1989; Fazio & Williams, 1986). Both findings are consistent with the spontaneous processing model. This above work has important implications for SCMs and health behaviors. The SCMs outlined here are deliberative processing models inasmuch as they focus on the conscious processing of information and fail to consider spontaneous or automatic influences on behavior. Hence, current SCMs may provide only a partial account of the social cognitive determinants of behavior, that is, they may only be applicable in situations where the individual has the ability and motivation to engage in deliberative processing of information (Conner, 1993). For many behavioral decisions, simplified or spontaneous decision-making rules may be employed instead (Norman & Conner, 1993). Fazio's (1990) spontaneous model has considerable potential in helping to provide a full account of the cognitive influences on behavior. However, it is clear that most of the empirical work to date has focused on issues surrounding the activation of attitudes and their influence on perception; later components of the model have received less attention. 8.01.5.4 Stage Models of Health Behavior A number of researchers have suggested that there may be qualitatively different stages in the initiation and maintenance of health behavior, and that to obtain a full understanding of the determinants of health behavior it is necessary to conduct a detailed analysis of the nature of these stages. From a social cognitive perspective, an important implication of this position is that different cognitions may be important at different stages in promoting health behavior. One of the first stage models was put forward by Prochaska and DiClemente (1984) in their transtheoretical model of change (TTM). Their model has been widely applied to analyze the process of change in alcoholism treatment (DiClemente & Hughes, 1990), smoking cessation (DiClemente et al., 1991), head injury rehabilitation (Lam, McMahon, Priddy, & Gehred-Schutlz, 1988), and psychotherapy (McConnaughly, DiClemente, Prochaska, & Velicer, 1989). In its most recent form, DiClemente et al. (1991) identify five stages of change: precontemplation, contemplation, preparation, action, and maintenance. Individuals are seen to progress through each stage to achieve successful maintenance of a new behavior. Taking the

is seen to have a direct influence via habitual responses. In particular, future work should develop measures of habit and routine that are discriminable from frequency of past behavior and outline the processes through which habit and routine determine behavior. 8.01.5.3 Processes by which Cognitions Influence Behavior One important implication of Ronis et al.'s (1989) distinction between habits and decisions is the suggestion that social cognition models may only predict health behavior under certain conditions. This issue has been addressed by Fazio (1990) in the development of the MODE model of attitudebehavior relationships. He suggests that attitudes (and presumably other cognitions) influence behavior via two distinct processes: a deliberative (or controlled) process and a spontaneous process. Most social cognition models can be labeled as deliberative processing models as they assume that behavior results from a controlled process of conscious deliberation. However, Fazio (1990) argues that individuals may only make a behavioral decision in such a manner when they have the opportunity and motivation to do so. Under other conditions, attitudes which are highly accessible in memory may determine behavior in a spontaneous fashion (Fazio, Powell, & Williams, 1989). When the spontaneous process is operating, an attitude may be automatically activated from memory following the presentation of relevant cues, with the likelihood of activation determined by the accessibility of the attitude. Once activated, the attitude shapes the perception of the attitude object in an automatic, attitude-congruent, fashion. For example, if a positive attitude is activated then this will lead the individual to attend to and notice the positive qualities of the attitude object. This automatic process of selective perception will therefore shape the individual's definition of the event, and thus determine behavior. If the event is defined on the basis of positive perceptions of the attitude object, for example, then approach behaviors will follow. In addition, it is argued that normative guidelines (e.g., social norms or rules) may also influence the definition of the event and thus may help to determine behavior in some situations. One important feature of the spontaneous processing model is that it outlines one way in which social cognitive variables (i.e., highly accessible attitudes) may determine behavior without systematic deliberation. To date, there has been little research on Fazio's model, although it has been successfully shown that

Limitations and Extensions of Existing Models example of smoking cessation, it is argued that in the precontemplation stage smokers are unaware that their behavior constitutes a problem and have no intention to quit. In the contemplation stage, smokers start to think about changing their behavior, but are not committed to try to quit. In the preparation stage, the smoker has an intention to quit and starts to make plans about how to quit. The action stage is characterized by active attempts to quit, and after six months of successful abstinence the individual moves into the maintenance stage characterized by attempts to prevent relapse and to consolidate the newly acquired nonsmoking status. Whilst relative widely applied, the evidence in support of the model and the different stages is at present relatively weak (see Weinstein, Rothman, & Sutton, in press). Heckhausen (1991) has similarly identified phases in the initiation and maintenance of behavior change; these being the predecisional, postdecisional, actional, and evaluative phases, which follow a similar progressive sequence as that outlined by Prochaska and DiClemente (1984). It is further suggested that different types of cognitions are important in each of these phases. So in the predecisional phase, cognitions about the desirability and feasibility of the behavior are believed to be important determinants of a desire to perform the behavior in question. This phase ends with the formation of an intention to change. In contrast, the decisional phase focuses on the development of plans and ends with the successful initiation of the behavior. In the actional phase the individual focuses on effectively achieving performance of the behavior and ends with the conlusion of the behavior. In the final, evaluative phase the individual compares achieved outcomes with initial goals in order to regulate and maintain behavior. While this four phase model of behavior was not developed for the prediction of health behavior, the potential for its application is clear (see Gollwitzer, 1993). Other stage models have been developed including the health action process approach (Schwarzer, 1992; Schwarzer & Fuchs, 1996), the precaution-adoption process (Weinstein, 1988, Weinstein & Sandman, 1992), and goal setting theory (Bagozzi, 1992, 1993; Bagozzi & Edwards, 1998). There are two important themes in each of the stage models outlined above. First, they emphasize a temporal perspective with different stages of behavior change. While the models postulate different numbers of stages, they all follow the same pattern from a precontemplation stage through a motivation stage to the initiation and maintenance of behavior. The

25

important point is that these models are dynamic in nature; people move from one stage to another over time. Second, these stage models imply that different cognitions are important at different stages (Sandman & Weinstein, 1993). For example, in the earlier stages information may be processed about the costs and benefits of performing a behavior, while in the later stages cognitions become more focused on the development of plans of action to initiate and support the maintenance of a behavior. This earlier motivational phase is assumed to end with the formation of an intention and only when the level of motivation or intention reaches a particular level is the individual assumed to be likely to move on to later stages. The main SCMs of health behavior are primarily concerned with people's motivations to perform a health behavior and, as such, provide strong predictions of behavioral intentions (i.e., the end of a motivational stage). Ajzen (1991), for example, reports an average multiple correlation of 0.71 between variables in the TPB and behavioral intention. However, intentions do not always lead to corresponding actions. Studies examining the intention behavior relationship have reported a wide range of correlations. For example, Sheppard et al. (1988) reported intentionbehavior correlations ranging from 0.10 to 0.94. Clearly, many people who intend to perform a behavior fail to do so. However, the SCMs considered do not address the issue of translating intentions into action. They can be conceptualized as static models that stop at the formation of an intention without distinguishing between intenders who become actors and those who do not. As Bagozzi (1993) argues, the variables outlined in the main social cognition models are necessary but not sufficient determinants of behavior. Clearly, a detailed analysis of the social cognitive variables important in translating intentions into action is required to provide a full account of the determinants of health behavior. Relatively little detailed attention has focused on the cognitive processes underlying the successful implementation of intentions. The main social cognition models contain few measures that account for the intention behavior gap (Abraham & Sheeran, 1993). The TPB attempts to do this by proposing a direct link between perceived behavioral control and behavior. Thus, people's perceptions about the amount of control they have over a behavior influence the likely performance of behavior independently of their intentions, although an analysis of the volitional processes underlying performance of a health behavior is required,

26

Health Behavior any other environmental cue. This suggests that the making of an implementation intention can significantly increase the performance of a behavior. In a study on exercise behavior, Kendzierski (1990) found that respondents were more likely to implement their intentions to exercise when they had engaged in some prior planning. Further work needs to establish the utility of implementation intentions in predicting health behavior. However, initial findings are encouraging and suggest that those who make such plans of action are more likely to initiate and maintain behavior. The work of Gollwitzer (1993) is important in that it identifies one way in which goal intentions may be translated into behavior. A similar but more comprehensive approach has been put forward by Bagozzi (Bagozzi, 1992, 1993; Bagozzi & Edwards, 1998; Bagozzi & Warshaw, 1990) in his model of goal achievement. He focuses on goal-directed behavior and argues that to initiate behavior individuals need to form an intention to try to achieve their desired goal. Once an intention to try has been formed, the individual focuses on the means, or instrumental acts, by which to attempt to achieve the desired goal. Considering the example of weight loss, a number of instrumental acts can be identified, including restricting between-meal foods, reducing overall calorie consumption, avoidance of high calorie foods, exercise, and so on. Bagozzi (1993) argues that for each of these instrumental acts, three appraisal tasks are performed. First, the individual considers the extent to which they are confident that they could perform the instrumental act (i.e., specific self-efficacies). Second, the likelihood that the instrumental act will help in achieving the desired goal is assessed (i.e., instrumental beliefs). Third, the individual considers an affective response towards the instrumental act (i.e., affect towards means). Once an individual initiates efforts to achieve a goal, there are a number of cognitive activities that support the successful initiation and maintenance of goal-directed behavior. First, the individual can develop plans in order to ensure that instrumental acts are performed. This involves identifying the situation or triggering conditions under which the instrumental act is performed (Bagozzi & Warshaw, 1990). This idea that certain environmental conditions may trigger behavior has a clear overlap with Gollwitzer's (1993) work on implementation intentions and Weinstein's (1988) messy desk analogy. One way in which plans are more likely to be acted upon is through the development of scripts or cognitive rehearsal, whereby the individuals imagine themselves performing the instrumental act (Anderson,

and a number of researchers have focused attention on this issue (e.g., Kuhl, 1985; Schwarzer, 1992; Weinstein, 1988). Here we focus on Gollwitzer's (1993) work on implementation intentions and Bagozzi's (1992) model of goal achievement to highlight the social cognitive variables important in the initiation and maintenance of behavior. Gollwitzer (1993) made the distinction between goal intentions and implementation intentions. While the former are concerned with intentions to perform a behavior or achieve a goal (i.e., I intend to achieve x), the latter are concerned with plans as to when, where, and how the goal intention is to be translated into behavior (i.e., I intend to initiate the goaldirected behavior x when situation y is encountered). Goal intentions are most like the intention construct in the TPB, although in the TPB such intentions usually refer to actions or behaviors rather than goals. The important point about implementation intentions is that they commit the individual to a specific course of action when certain environmental conditions are met; in so doing they help translate goal intentions into action. Gollwitzer (1993) argues that by making implementation intentions individuals pass over control to the environment. The environment acts as a cue to action, such that when certain conditions are met, the performance of the intended behavior follows. These ideas have similarities with Weinstein's (1988) messy desk analogy, whereby people may have intentions to achieve a number of goals (i.e., projects) which get lost on the messy desk. Which project is actually worked upon is determined by environmental factors in a similar way as outlined by Gollwitzer (1993). Gollwitzer (1993, 1996) has compiled a range of experimental evidence to support the view that the making of implementation intentions can aid the performance of intended behavior. To date, the only application of implementation intentions to the prediction of health behavior is a study by Orbell, Hodgkins, and Sheeran (1997) on breast self-examination. At the end of a questionnaire about breast self-examination, half the women were asked to indicate when and where in the next month they intended to perform breast self-examination. A one month follow-up found that 64% of these women had performed breast self-examination that month compared with only 16% of women who had not made an implementation intention, despite no difference in goal intentions. In addition, everyone in the implementation condition who actually performed the behavior reported doing so in response to the environmental cue in the implementation intention and not in response to

Limitations and Extensions of Existing Models 1983). Another is through the use of precommitting devices whereby the behavior is made more likely by precommitting oneself to it (e.g., avoiding eating butter at home by not having butter in the house). Bagozzi (1993) also proposes that ongoing behavior has to be monitored to ensure, for example, that the instrumental acts achieve their objectives. If any unforeseen impediments are encountered then these need to be taken into consideration and any future plans modified accordingly. These ideas overlap with Kuhl's (1985) theory of action control which identifies a number of processes by which individuals attempt to control their actions and achieve their goals. These processes may be particularly important in allowing individuals to overcome temptations to break their new behavior (see Loewenstein, 1996). Finally, goal-directed behavior is likely to be stronger and more persistent if the individual has a strong sense of commitment to the decision to try to achieve the goal and the means to achieve it. 8.01.5.5 Theoretical Integrations and Future Directions The above research demonstrates some of the ways in which research into health behaviors based upon social cognitive approaches is developing. Here we outline possible directions for the future development of an integrative social cognition model of health behavior, outline some of the basic requirements for such a model, and indicate some fruitful avenues for future research. It is clear that to fully explain health behavior it is necessary to develop a more dynamic model that examines different stages or phases in the contemplation, initiation, and maintenance of behavior. What is being proposed is an integration of current SCMs (such as the TPB) with stage models of health behavior (such as the TTM). Several authors have recommended such an integration (e.g., Courneya, Nigg, & Estabrooks, 1998; Godin, Desharnais, Valois, & Bradet, 1995; Marcus, Eaton, Rossi, & Harlow, 1994). Though the stage models considered in this chapter have suggested differing number of stages, it is likely that an integrative model should address at least four or five main stages: precontemplation, contemplation, planning, action, and maintenance (Norman & Conner, 1996a). One implication of the identification of different stages is that different cognitive variables may be important in ensuring movement from one stage to the next. In the first stage, the individuals are not thinking about

27

making a change to their behavior. However, this stage may be brought to an end by a range of cues to action, as outlined in the HBM, which may motivate the individual to start thinking about performing a health-related behavior. One such cue to action may be perceived threat (i.e., perceived susceptibility and perceived severity). While perceived susceptibility and perceived severity are seen to be important determinants of behavior in the HBM, research with these dimensions has tended to show that they are relatively weak predictors of behavior. However, as Schwarzer (1992) has argued, it may be more appropriate to consider these variables to have an indirect or more distal influence on behavior. Thus they may act as a cue to action, motivating the individual to start deliberating over performing a health-related behavior, and thus ensuring movement from the first to the second stage. In the second stage, a decision-making or motivation stage, the individual is thinking about adopting a new behavior, and the stage ends when the individual forms an intention to perform the behavior. To date, most social cognition models have been primarily concerned with this stage. These models distinguish between three distinct determinants of individuals' intentions to perform a health behavior. First, are outcome expectancies, which focus on the perceived consequences of performing a behavior. These expectancies may also cover the notion of behavioral beliefs as considered in the TPB and include anticipated affective reactions. Second, are normative influences which are primarily tapped by the subjective norm and normative belief components of the TPB. This group of variables could also include descriptive norms, moral norms, and perceived social support. The third influence on individuals' intentions to perform a behavior is control beliefs (or self-efficacy expectancies) and may be based on a consideration of perceived barriers (HBM) and control beliefs (TPB). In the motivation stage, it is likely that other variables may have a more distal influence on behavioral intention via the variables outlined above. For example, health locus of control beliefs may help shape self-efficacy expectancies, self-identity and health values may influence the interpretation of the potential consequences of a behavior, and past behavior or experience may provide information which is used to determine the ease or difficulty of performing a behavior (i.e., self-efficacy). Once a behavioral intention has been formed, it has to be translated into behavior. In the third stage, the individual is therefore concerned with planning; focusing on the specific actions or instrumental acts that need to be performed and

28

Health Behavior the contention that important determinants of health behaviors are identified in these models. Further refinement and development of these models along the lines we have suggested and reviewed may lead to even better predictions of behavior and greater understanding of health behavior and how individuals may be encouraged to change. However, persuasive messages targeted at relevant cognitions identified by SCMs may not be sufficient to produce the major behavior change necessary for health benefits to accrue. It may be that strategies which employ multiple level interventions which take account not only of the psychosocial factors influencing performance of the behavior (derived from SCMs) but also models of the process of persuasion of how people change and the context in which changes are made will be important (Glanz et al., 1990; Hockbaum & Lorig, 1992; Winett, 1985). 8.01.7 REFERENCES
Abraham, C., & Sheeran, P. (1993). Inferring cognitions, predicting behaviour: Two challenges for social cognition models. Health Psychology Update, 14, 1823. Abraham, C. S., Sheeran, P., Abrams, D., & Spears, R. (1994). Exploring teenagers' adaptive and maladaptive thinking in relation to the threat of HIV infection. Psychology and Health, 9, 253272. Adler, N., & Matthews, K. (1994). Health psychology: Why do some people get sick and some stay well? Annual Review of Psychology, 45, 229259. Aguirre-Molina, M., & Gorman, D. M. (1996). Community-based approaches for the prevention of alcohol, tobacco, and other drug use. Annual Review of Public Health, 17, 337358. Aho, W. R. (1979). Smoking, dieting and exercise: Age differences in attitudes and behavior relevant to selected health belief model variables. Rhode Island Medical Journal, 62, 95102. Aiken, L. S., West, S. G., Woodward, C. K., Reno, R. R., & Reynolds, K. D. (1994). Increasing screening mammography in asymptomatic women: Evaluation of a second-generation, theory-based program. Health Psychology, 13, 526538. Ajzen, I. (1988). Attitudes, personality and behavior. Milton Keynes, UK: Open University Press. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179211. Ajzen, I. (1996). The directive influence of attitudes on behavior. In P. Gollwitzer & J. A. Bargh (Eds.), Psychology of action (pp. 385403). New York: Guilford Press. Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood-Cliffs, NJ: Prentice-Hall. Allied Dunbar Fitness Survey (1992). London: Health Education Authority. Allison, K. R. (1987). Perceived control as a determinant of preventive health behaviour for heart disease and lung cancer. Unpublished doctoral dissertation, University of Toronto. Amler, R. W., & Dull, H. B. (1987). Closing the gap. New York: Oxford University Press. Anderson, C. A. (1983). Imagination and expectation: The effect of imagining behavioral scripts on personal

the resources required to support them. Thus, a number of authors have highlighted the importance of action plans in this stage (Bagozzi & Warshaw, 1990; Schwarzer, 1992). Similarly, Gollwitzer (1993) focuses on implementation intentions which help ensure performance of the target behavior. Despite slight differences between definitions of these two concepts, both emphasize the need to construct fairly detailed plans of action in order to bridge the intention behavior gap. As Schwarzer (1992) argues, selfefficacy may have an important role to play in the development and implementation of such plans, as might self-identity (Sparks & Shepherd, 1992) and a sense of commitment (Bagozzi, 1993). The planning stage is brought to an end when the individual initiates behavior. In the fourth stage, the individual has to ensure that the behavior is successfully enacted. Various cognitive processes which are concerned with the monitoring and controlling of behavior may be important in this stage. Schwarzer (1992), for example, highlights the need for action control in which the behavior is re-evaluated against initial goals in order to regulate and maintain behavior (see also Bagozzi & Warshaw, 1990; Heckhausen, 1991). As with the development of action plans, self-efficacy, self-identity, and commitment may all be important variables in ensuring that behavior is maintained. In addition, Kuhl (1985) has argued that there may be individual differences in people's propensity to engage in monitoring activities that may account for why behavior is not always maintained. These monitoring activities help to ensure that the behavior is successfully completed. The same activities will be important in the fifth, maintenance stage for ongoing behaviors, where the individual is concerned with ensuring that the behavior is successfully repeated. 8.01.6 CONCLUSIONS Health behaviors have been demonstrated to have important consequences for both the quality and length of life through influencing various disease outcomes. We have attempted to justify the interest in understanding health behaviors as a basis for attempting to change their occurrence in order to increase both length and quality of life. SCMs provide one approach to understanding health behavior in describing the important social cognitive variables underlying such behaviors. We believe that these models provide an important way of achieving these aims by providing a means for identifying useful targets for persuasion. It would seem that there is already sufficient literature to support

References
intentions. Journal of Personality and Social Psychology, 45, 293305. Armitage, C. J., & Conner, M. (in press). Distinguishing perceptions of control from self-efficacy: Predicting consumption of a low fat diet using the theory of planned behavior. Journal of Applied Social Psychology. Ashworth, H. W. (1963). An experiment in presymptomatic diagnosis. Journal of the Royal College of General Practitioners, 6, 71. Aspinwall, L. G., Kemeny, M. E., Taylor, S. E., Schneider, S. G., & Dudley, J. P. (1991). Psychosocial predictors of gay men's AIDS risk-reduction behavior. Health Psychology, 10, 432444. Babrow, A. S., Black, D. R., & Tiffany, S. T. (1990). Beliefs, attitudes, intentions, and a smoking-cessation program: a planned behavior analysis of communication campaign development. Health Communication, 2, 145163. Baer, J. S., Marlatt, G. A., Kivlahan, D. R., Fromme, K., Larimer, M. E., & Williams, E. (1992). An experimental test of three methods of alcohol risk reduction with young adults. Journal of Consulting and Clinical Psychology, 60, 974979. Bagozzi, R. P. (1992). The self-regulation of attitudes, intentions and behavior. Social Psychology Quarterly, 55, 178204. Bagozzi, R. P. (1993). On the neglect of volition in consumer research: A critique and proposal. Psychology and Marketing, 10, 215237. Bagozzi, R. P., & Edwards, E. A. (1998). Goal setting and goal pursuit in the regulation of body weight. Psychology and Health, 13. Bagozzi, R. P., & Warshaw, P. R. (1990). Trying to consume. Journal of Consumer Research, 17, 127140. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191215. Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122147. Bandura, A. (1986). Social foundations of thought and action: A cognitive social theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1991). Self-efficacy mechanism in physiological activation and health-promoting behavior. In J. Madden (Ed.), Neurobiology of learning, emotion and affect (pp. 229270). New York: Raven Press. Basen-Engquist, K., & Parcel, G. S. (1992). Attitudes, norms, and self-efficacy: A model of adolescents' HIVrelated sexual risk behavior. Health Education Quarterly, 19, 263277. Beale, D. A., & Manstead, A. S. R. (1991). Predicting mothers' intentions to limit frequency of infants' sugar intake: Testing the theory of planned behavior. Journal of Applied Social Psychology, 21, 409431. Beck, K. H. & Lund, A. K. (1981). The effects of health threat seriousness and personal efficacy upon intentions and behavior. Journal of Applied Social Psychology, 11, 401415. Beck, K. H. (1981). Driving while under the influence of alcohol: Relationship to attitudes and beliefs in a college population. American Journal of Drug and Alcohol Abuse, 8, 377388. Beck, L., & Ajzen, I. (1991). Predicting dishonest actions using the theory of planned behavior. Journal of Research in Personality, 25, 285301. Becker, M. H. (1974). The health belief model and sick role behavior. Health Education Monographs, 2, 409419. Becker, M. H., Kaback, M. M., Rosenstock, I. M., & Ruth, M. (1975). Some influences on public participation in a genetic screening program. Journal of Community Health, 1, 314. Becker, M. H., & Maiman, L. A. (1983). Models of healthrelated behavior. In D. Mechanic (Ed.), Handbook of

29

health, health care and the health professions (pp. 539 568). New York: Free Press. Becker, M. H., Radius, S. M., & Rosenstock, I. M. (1978). Compliance with a medical regimen for asthma: A test of the health belief model. Public Health Reports, 93, 268277. Becker, M. H., & Rosenstock, I. M. (1987). Comparing social learning theory and the health belief model. In W. B. Ward (Ed.), Advances in health education and promotion (Vol. 2, pp. 245249). Greenwich, CT: JAI Press. Belloc, N. B. (1973). Relationship of health practices to mortality. Preventive Medicine, 2, 6781. Belloc, N. B., & Breslow, L. (1972). Relationship of physical health status and health practices. Preventive Medicine, 9, 409421. Bengel, J., Belz-Merk, M., & Farin, E. (1996). The role of risk perception and efficacy cognitions in the prediction of HIV-related preventive behavior and condom use. Psychology and Health, 11, 505525. Bentler, P. M., & Speckart, G. (1979). Models of attitudebehavior relations. Psychological Review, 86, 452464. Bernier, M. & Avard, J. (1986). Self-efficacy, outcome and attrition in a weight reduction program. Cognitive Therapy and Research, 10, 319338. Best, J. A. (1975). Tailoring smoking withdrawal procedures to personality and motivational differences. Journal of Consulting and Clinical Psychology, 43, 18. Biddle, B. J. Bank, B. J., & Slavings, R. L. (1987). Norms, preferences, identities and retention decisions. Social Psychology Quarterly, 50, 322337. Black Report (1988). Inequalities in health: the Black report and the health divide. London: DHSS. Blair, S. N., Goodyear, N. N., Gibbons, L. W., & Cooper, K. H. (1984). Physical fitness and incidence of hypertension in healthy normotensive men and women. Journal of the American Medical Association, 252, 487490. Blane, D., Smith, G. D., & Bartley, M. (1990). Social class differences in years of potential life lost: Size trends, and principal causes. British Medical Journal, 301, 429432. Blaxter, M. (1990). Health and lifestyles. London: Tavistock. Boer, I., & Seydel, E. R. (1996). Protection motivation theory. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 95120). Buckingham, UK: Open University Press. Boldero, J., Moore, S., & Rosenthal, D. (1992). Intention, context, and safe sex: Australian adolescents' responses to AIDS. Journal of Applied Social Psychology, 22, 13741396. Borgida, E., Conner, C., & Manteufel, L. (1992). Understanding living kidney donation: a behavioral decision perspective. In S. Spacapan & S. Oskamp (Eds.), Helping and being helped: Naturalistic studies (pp. 183212). Newbury Park, CA: Sage. Boulton, M., Schramm Evans, Z., Fitzpatrick, R., & Hart, G. (1991). Bisexual men: Women, safer sex, and HIV infection. In P. Aggleton, P. M. Davies, & G. Hart (Eds.), AIDS: Responses, policy and care (pp. 5567). London: Falmer Press. Boyd, B., & Wandersman, A. (1991). Predicting undergraduate condom use with the Fishbein and Ajzen and the Triandis attitude-behavior models: Implications for public health interventions. Journal of Applied Social Psychology, 21, 18101830. Bradley, C., Lewis, K. S., Jennings, A. M., & Ward, J. D. (1990). Scales to measure perceived control developed specifically for people with tablet-treated diabetes. Diabetic Medicine, 7, 685694. Brawley, L. R. (1993). The practicality of using social psychological theories for exercise and health research and intervention. Journal of Applied Sport Psychology, 5, 99115.

30

Health Behavior
Colliver, J. D., & Mallin, H. (1986). State an national trends in alcohol-related mortality, 19751982. Alcohol Health and Research World, 10, 6064. Committee on Medical Aspects of Food Policy (1991). Dietary reference values for food energy and nutrients for the United Kingdom. Report on Health and Social Subjects, 41. London: HMSO. Condiotte, M. M., & Lichtenstein, E. (1981). Self-efficacy and relapse in smoking cessation programs. Journal of Consulting and Clinical Psychology, 49, 648658. Connelly, C. E. (1984). Compliance with outpatient lithium therapy. Perspectives in Psychiatric Care, 22, 4450. Conner, M. (1993). Pros and cons of social cognition models in health behaviour. Health Psychology Update, 14, 2431. Conner, M., Martin, E., Silverdale, N., & Grogan, S. (1996). Dieting in adolescence: An application of the theory of planned behavior. British Journal of Health Psychology, 1, 315325. Conner, M., & Norman, P. (1994). Comparing the health belief model and the theory of planned behaviour in health screening. In D. R. Rutter & L. Quine (Eds.), Social psychology and health: European perspectives (pp. 124). Aldershot, UK: Avebury. Conner, M., & Norman, P. (Eds.) (1996a). Predicting health behaviour. Buckingham, UK: Open University Press. Conner, M., & Norman, P. (1996b). The role of social cognition in health behaviours. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 122). Buckingham, UK: Open University Press. Conner, M., & Sparks, P. (1996). The theory of planned behaviour and health behaviours. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 121162). Buckingham, UK: Open University Press. Consensus Development Conference on Lowering Blood Cholesterol to Prevent Heart Disease (1985). Lowering blood cholesterol to prevent heart disease. Journal of the Australian Medical Association, 253, 20802086. Corcoran, K. J. (1991). Efficacy, skills, reinforcement, and choice behavior. American Psychologist, 46, 155157. Costello, R. M., & Manders, K. R. (1974). Locus of control and alcoholism. British Journal of Addiction, 69, 1117. Courneya, K. S., Nigg, C. R., & Estabrooks, P. A. (1998). Relationships among the theory of planned behavior, stages of change, and exercise behavior in older persons over a three year period. Psychology & Health, 13. Cox, B. D, Huppert, F. A., & Whichelow, M. J. (1993). The health and lifestyles survey: Seven years on. Aldershot, UK: Dartmouth. Dean, K. (1991). Relationships between knowledge and belief variables and health maintenance behaviours in a Danish population over 45 years of age. Journal of Ageing and Health, 3, 386406. Desharnais, R., Bouillon, J., & Godin, G. (1986). Selfefficacy and outcome expectations as determinants of exercise adherence. Psychological Reports, 59, 1155 1159. DeVellis, B. M., Blalock, S. J., & Sandler, R. S. (1990). Predicting participation in cancer screening: the role of perceived behavioral control. Journal of Applied Social Psychology, 20, 639660. DeVito, A. J., Bogdanowicz, J., & Reznikoff, M. (1982). Actual and intended health-related information seeking and health locus of control. Journal of Personality Assessment, 46, 6369. DeVries, H., Backbier, E., Kok, G., & Dijkstra, M. (1995). The impact of social influences in the context of attitude, self-efficacy, intention, and previous behavior as predictors of smoking onset. Journal of Applied Social Psychology, 25, 237257.

Breslow, L., & Enstrom, J. E. (1980). Persistence of health habits and their relationship to mortality. Preventive Medicine, 9, 469483. British Medical Journal (1982). Alcohol problems. London: BMJ Publications. Brock, B. M., Haefner D. P., & Noble, D. S. (1988). Alameda County Redux: Replication in Michigan. Preventive Medicine, 17, 483495. Brouwers, M. C., & Sorrentino, R. M. (1993). Uncertainty orientation and protection motivation theory: The role of individual differences in health compliance. Journal of Personality and Social Psychology, 65, 102112. Brown, J. D. (1991). Staying fit and staying well: Physical fitness as a moderator of life stress. Journal of Personality and Social Psychology, 60, 555561. Brown, N., Muhlenkamp, A., Fox, L., & Osborn, M. (1983). The relationship among health beliefs, health values, and health promotion activity. Western Journal of Nursing Research, 5, 155163. Brubaker, R. G., & Fowler, C. (1990). Encouraging College males to perform testicular self-examination: Evaluation of a persuasive message based on the revised theory of reasoned action. Journal of Applied Social Psychology, 17, 14111422. Bundek, N. I., Marks, G., & Richardson, J. L. (1993). Role of health locus of control beliefs in cancer screening of elderly Hispanic women. Health Psychology, 12, 193199. Burk, C., & Kimiecik, J. (1994). Examining the relationship among locus of control, health value and exercise. Health Values, 18, 1423. Butts, S. V., & Chotlas, J. A. (1973). A comparison of alcoholics and nonalcoholics on perceived locus of control. Quarterly Journal of Studies on Alcohol, 34, 13271332. Calnan, M. (1989). Control over health and patterns of health-related behaviour. Social Science and Medicine, 29, 131136. Carlson, B., & Petti, K. (1989). Health locus of control and participation in physical activity. American Journal of Health Promotion, 3, 3237. Chaiken, S., Lieberman, A., & Eagly, A. H. (1989). Heuristic and systematic information processing within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp. 212252). New York: Guilford Press. Chambliss, C. A., & Murray, E. J. (1979a). Efficacy attribution, locus of control, and weight loss. Cognitive Therapy and Research, 3, 349353. Chambliss, C. A., & Murray, E. J. (1979b). Cognitive procedures for smoking reduction: symptom attribution versus efficacy attribution. Cognitive Therapy and Research, 3, 9195. Champion, V. L. (1984). Instrument development of the health belief model constructs. Advances in Nursing Science, 6, 7385. Charng, H. W., Piliavin, J. A., & Callero, P. L. (1988). Role identity and reasoned action in the prediction of repeated behavior. Social Psychology Quarterly, 51, 303317. Chen, M. M., & Land, K. C. (1986). Testing the health belief model: LISREL analysis of alternative models of the causal relationships between health beliefs and preventive dental behavior. Social Psychology Quarterly. 49, 4560. Cialdini, R. B., Reno, R. R., & Kallgren, C. A. (1990). A focus theory of normative conduct: Recycling the concept of norms to reduce littering in public places. Journal of Personality and Social Psychology, 58, 10151026. Colletti, G., Supnick, J. A., & Payne, T. J. (1985). The smoking self-efficacy questionnaire (SSEQ): Preliminary scale development and validation. Behavioral Assessment, 7, 249260.

References
DeVries, H., Dijkstra, M., & Kuhlman, P. (1988). Selfefficacy: The third factor besides attitude and subjective norm as a predictor of behavioural intentions. Health Education Research, 3, 273282. DiClemente, C. C., & Hughes, S. O. (1990). Stages of change profiles in outpatient alcoholism treatment. Journal of Substance Abuse, 2, 217235. DiClemente, C. C., Fairhurst, S. K., Velasquez, M. M., Prochaska, J. O., Velicer, W. F., & Rossi, J. S. (1991). The process of smoking cessation: An analysis of precontemplation, contemplation, and preparation stages of change. Journal of Consulting and Clinical Psychology, 59, 295304. Dishman, R. K. (1982) Compliance/adherence in healthrelated exercise. Health Psychology, 1, 237267. Doll, R., Peto, R., Wheatley, K., Gray, R., & Sutherland, I. (1994). Mortality in relation to smoking: 40 years' observations on male British doctors. British Medical Journal, 309, 901911. Donovan, D. M., & O'Leary, M. R. (1978). The drinkingrelated locus of control scale. Journal of Studies on Alcohol, 39, 759784. Duffy, M. E. (1988). Determinants of health promotion in midlife women. Nursing Research, 37, 358362. Dzewaltowski, D. A., Noble, J. M., & Shaw, J. M. (1990). Physical activity participationsocial cognitive theory versus the theories of reasoned action and planned behavior. Journal of Sport and Exercise Psychology, 12, 388405. Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX: Harcourt Brace Jovanovich. Edwards, W. (1954). The theory of decision making. Psychological Bulletin, 51, 380417. Edwards, W. (1992). Utility theories: measurements and applications. Boston: Kluwer. Expert Panel on Detection, Evaluation, and Treatment of High Cholesterol in Adults (1993). Summary of the second report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high cholesterol in adults. Journal of the American Medical Association, 269, 30153023. Farin, E. (1994). Eine Metaanalyse Empirischer Studien zum Pra diktiven Wert Kognitiver Variablen der HIVBezogenen Risikowahrnehmung und-Verabeitung fur das HIV-Risikoverhalten. Dissertation. Universitat Frei burg. Fazio, R. H. (1990). Multiple processes by which attitudes guide behavior: The MODE model as an integrative framework. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp. 75109). San Diego, CA: Academic Press. Fazio, R. H., & Williams, C. J. (1986). Attitude accessibility as a moderator of the attitude-perception and attitude-behavior relations: An investigation of the 1984 presidential election. Journal of Personality and Social Psychology, 51, 505514. Fazio, R. H., Powell, M. C., & Williams, C. J. (1989). The role of attitude accessibility in the attitude-to-behavior process. Journal of Consumer Research, 16, 280288. Feather, N. T. (1982). Expectations and actions: Expectancy-value models in psychology. Hillsdale, NJ: Erlbaum. Fife-Schaw, C.R., & Breakwell, G. M. (1992). Estimating sexual behaviour parameters in the light of AIDS: A review of recent UK studies of young people. AIDS Care, 4, 187201. Fishbein, M. (1993). Foreword. In D. J. Terry, C. Gallois, & M. McCamish (Eds.). The theory of reasoned action: Its application to AIDS-related behavior (pp. xvxxv). Oxford, UK: Pergamon. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior. New York: Wiley. Folkins, C. H., & Sime, W. E. (1981). Physical fitness

31

training and mental health. American Psychologist, 36, 373389. Forrest, A.P.M. (1986). Breast cancer screening. London: HMSO. Fredricks, A. J., & Dossett, D. L. (1983). Attitude behavior relations: A comparison of the FishbeinAjzen and BentlerSpeckart models. Journal of Personality and Social Psychology, 45, 501512. Freeling, P. (1965). Candidates for exfoliative cytology of the cervix. Journal of the Royal College of General Practitioners, 10, 261. Friedman, G. D., Dales, L. G., & Ury, H. K. (1979). Mortality in middle-aged smokers and non-smokers. New England Journal of Medicine, 300, 213217. Fries, J. F., Green, L. W., & Levine, S. (1989). Health promotion and the compression of mortality. Lancet, i, 481483. Frisch, D., & Clemen, R. T. (1994). Beyond expected utility: rethinking behavioral decision making. Psychological Bulletin, 116, 4654. General Household Survey (1989). London: OPCS. General Household Survey (1992). London: OPCS. General Household Survey (1993). London: OPCS. General Household Survey (1994). London: OPCS. Georgiou, A., & Bradley, C. (1992). The development of a smoking specific locus of control scale. Psychology and Health, 6, 227240. Gerrard, M., Gibbons, F. X., & Bushman, B. J. (1996). Relation between perceived vulnerability to HIV and precautionary sexual behavior. Psychological Bulletin, 119, 390409. Gianetti, V. J., Reynolds, J., & Rihen, T. (1985). Factors which differentiate smokers from ex-smokers among cardiovascular patients: A discriminant analysis. Social Science and Medicine, 20, 241245. Gibbons, F. X., & Gerrard, M. (1995). Predicting young adults' health risk behavior. Journal of Personality and Social Psychology, 69, 505517. Gibbons, F. X., & Gerrard, M. (1997). Health images and their effects on health behavior. In B. P. Buuuk & F. X. Gibbons (Eds.), Health, coping, and social comparisons (pp. 6394). Hillsdale, NJ: Erlbaum. Gibbons, F. X., Gerrard, M., Ouelette, J., & Burzette, R. (1998). Cognitive antecedents to adolescent health risk: Discriminating between behavioral intention and behavioral willingness. Psychology and Health, 13. Glanz, K., Lewis, F. M., & Rimmer, B. K. (Eds.) (1990). Health behavior and health education: Theory, research and practice. San Francisco: Josey-Bass. Godin, G., Desharnais, R., Valois, P., & Bradet, R. (1995). Combining behavioral and motivational dimensions to identify and characterize the stages in the process of adherence to exercise. Psychology and Health, 10, 333344. Godin, G., & Shepherd, R. J. (1987). Psychosocial factors influencing intentions to exercise in a group of individuals ranging from 45 to 74 years of age. In M. E. Berridge & G. R. Ward (Eds.), International perspectives on adapted physical activity (pp. 243249). Champaign, IL: Human Kinetics Publishers. Godin, G., Valois, P., & Lepage, L. (1993). The pattern of influence of perceived behavioral control upon exercise behavior: An application of Ajzen's theory of planned behavior. Journal of Behavioral Medicine, 16, 81102. Godin, G., Valois, P., Lepage, L., & Desharnais, R. (1992). Predictors of smoking behaviourAn application of Ajzen's theory of planned behaviour. British Journal of Addiction, 87, 13351343. Gollwitzer, P. M. (1990). Action phases and mind-sets. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition: foundations of social behaviour (Vol. 2, pp. 5392). New York: Guilford Press.

32

Health Behavior
theoretical specificity. Social Science and Medicine, 31, 10511056. Honkanen, R., Ertama, L., Kuosmanen, P., Linnoila, M., Ahla, A., & Visuri, T. (1983). The role of alcohol in accidental falls. Journal of Studies on Alcohol, 44, 231245. Horwitz, M. B., Hindi-Alexander, M., & Wagner, T. J. (1985). Psychosocial mediators of abstinence, relapse, and continued smoking: A one-year follow-up of a minimal intervention. Addictive Behaviors, 10, 2939. Hounsa, A. M., Godin, G., Alihonou, E., Valois, P., & Girard, J. (1993). An application of Ajzen's theory of planned behaviour to predict mothers' intention to use oral rehydration therapy in a rural area of Benin. Social Science and Medicine, 37, 253261. Houston, D. A., & Fazio, R. H. (1989). Biased processing as a function of attitude accessibility: Making objective judgements subjectively. Social Cognition, 7, 5166. Huba, G. J., Wingard, J. A., & Bentler, P. M. (1981). Intentions to use drugs among adolescents: A longitudinal analysis. Journal of the Additions, 16, 331339. Janis, I. L. (1967). Effects of fear arousal on attitude change: recent developments in theory and experimental research. In L. Berkowitz (Ed.) Advances in experimental social psychology (Vol. 3, pp. 166224). New York: Academic Press. Janz, N. K., & Becker, M. H. (1984). The health belief model: A decade later. Health Education Quarterly, 11, 147. Jeffrey, R. W., Bjornson-Benson, W. M., Rosenthal, B. S., Lindquist, R. A., Kurth, C. L., & Johnson, S. L. (1984). Correlates of weight loss and its maintenance over two years of follow-up among middle-aged men. Preventive Medicine, 13, 155168. Jonas, K. (1993). Expectancy-value models of health behaviour: An analysis by conjoint measurement. European Journal of Social Psychology, 23, 167183. Kaplan, G. D., & Cowles, A. (1978). Health locus of control and health value in the prediction of smoking behavior. Health Education Monographs, 6, 129137. Kavanagh, D. J., Pierce, J., Lo, S. K., & Shelley, J. (1993). Self-efficacy and social support as predictors of smoking after a quit attempt. Psychology and Health, 8, 231242. Kelley, J. A., St. Lawrence, J. S., Brasfield, T. L., Lemke, A., Amidei, T., Roffman, R. E., Hood, H. V., Smith, J. E., Kilgore, H., & McNeill, C. (1990). Psychological factors that predict AIDS high-risk versus AIDS precautionary behavior. Journal of Consulting and Clinical Psychology, 58, 117120. Kendzierski, D. (1990). Decision making versus decision implementation: An action control approach to exercise adoption and adherence. Journal of Applied Social Psychology, 20, 2745. Keys, A. (1970). Coronary heart disease in seven countries. Circulation, 41(Suppl. 1), entire issue. King, A. C., Blair, S. N., Bild, D. E., Dishman, R. K., Dubbert, P. M., Marcus, B. H. Oldridge, N. B., Paffenbarger, R. S., Powell, K. E., & Yeager, K. K. (1992). Determinants of physical activity and interventions in adults. Medicine and Science in Sports and Exercise, 24, 10791092. King, J. (1982). The impact of patients' perceptions of high blood pressure on attendance at screening. Social Science and Medicine, 16, 10791091. Kinket, B., Paans, L., & Verplanken, B. (1992). Patient education by general practitioners: Application of the theory of planned behavior. Gedrag and Gezondheid Tijdschrift voor Psychologie and Gezondheid, 20, 187195. Kirsch, I. (1986). Early research on self-efficacy: What we already know without knowing we knew. Journal of Social and Clinical Psychology, 4, 339358. Kirscht, J. P., Becker, M. H., & Eveland, P. (1976). Psychological and social factors as predictors of medical

Gollwitzer, P. M. (1993). Goal achievement: The role of intentions. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 4, pp. 141 185). Chichester, UK: Wiley. Gollwitzer, P. M. (1996). The volitional benefits of planning. In P. Gollwitzer & J. A. Bargh (Eds.), Psychology of action (pp. 287312). New York: Guilford Press. Gorsuch, R. L., & Ortberg, J. (1983). Moral obligation and attitudes: Their relation to behavioral intentions. Journal of Personality and Social Psychology, 44, 10251028. Gottlieb, N. H., & Baker, J. A. (1986). The relative influence of health beliefs, parental and peer behaviours and exercise program participation on smoking, alcohol use and physical activity. Social Science and Medicine, 22, 915927. Gottlieb, N. H., & Green, L. W. (1984). Life events, social network, life-style and health: An analysis of the 1979 National survey of personal health practices and consequences. Health Education Quarterly, 11, 91105. Graham, H. (1987). Women's smoking and family health. Social Science and Family Health, 25, 4756. Gregg, I. (1966). The recognition of early chronic bronchitis. Journal of the American Medical Association, 176, 11141116. Grube, J. W., Morgan, M., & McGee, S. T. (1986). Attitudes and normative beliefs as predictors of smoking intentions and behaviours: A test of three models. British Journal of Social Psychology, 25, 8193. Hallal, J. (1982). The relationship of health beliefs, health locus of control, and self-concept to the practice of breast self-examination in adult women. Nursing Research, 31, 137142. Harris, R., & Lynn, M. W. (1985). Health beliefs, compliance and control of diabetes mellitus. Southern Medical Journal, 2, 162166. Harrison, J. A., Mullen, P. D., & Green, L. W. (1992). A meta-analysis of studies of the Health Belief Model with adults. Health Education Research, 7, 107116. Hartman, P. E., & Becker, M. H. (1978). Non-compliance with prescribed regimen among chronic haemodialysis patients. Journal of Dialysis and Transplantation, 7, 978985. Hayes, D., & Ross, C. E. (1987). Concern with appearance, health beliefs and eating habits. Journal of Health and Social Behaviour, 28, 120130. Health Promotion Authority for Wales (1990) Health in Wales. Cardiff, UK: Author. Heckhausen, H. (1991). Motivation and action. Berlin, Germany: Springer-Verlag. Hennekens, C. H. (1983). Alcohol. In N. M. Kaplan & J. Stamler (Eds.) Prevention of coronary heart disease: practical management of the risk factors (pp. 130138). Philadelphia: Saunders. Hester, N. R., & Macrina, D. M. (1985). The health belief model and the contraceptive behavior of college women: Implications for health education. Journal of American College Health, 33, 245252. Hochbaum, G. M. (1958). Public participation in medical screening programs: A socio-psychological study. Public Health Service Publication No. 572. Washington, DC: US Government Printing Office. Hockbaum, G. M., & Lorig, K. (1992). Roles and uses of theory in health education practice. Health Education Quarterly, 19, 289412. Hodgkins, S. & Orbell, S. (1998). Can protection motivation theory predict behaviour? A longitudinal study exploring the role of previous behaviour. Psychology and Health, 13. Hofstetter, C. R., Sallis, J. F., & Howell, M. F. (1990). Some health dimensions of self-efficacy: analysis of

References
behaviour. Journal of Medical Care, 14, 422431. Kok, G., de Vries, H., Mudde, A. N., & Strecher, V. J. (1991). Planned health education and the role of selfefficacy: Dutch research. Health Education Research, 6, 231238. Kristiansen, C. M. (1986). A two-value model of preventive health behaviour. Basic and Applied Social Psychology, 7, 173184. Kuhl, J. (1985). Volitional mediators of cognition-behavior consistency: Self-regulatory processes and action versus state orientation. In J. Kuhl & J. Beckman (Eds.), Action control: From cognition to behavior (pp. 101128). New York: Springer. Kurland, N. B. (1995). Ethical intentions and the theories of reasoned action and planned behavior. Journal of Applied Social Psychology, 25, 297313. Lam, C. S., McMahon, B. T., Priddy, D. A., & GehredSchutlz, A. (1988). Deficit awareness and treatment performance among traumatic head injury adults. Brain Injury, 2, 235242. Langlie, J. K. (1977). Social networks, health beliefs and preventive health behaviours. Journal of Health and Social Behaviour, 18, 244260. Lau, R. R. (1982). Origins of health locus of control beliefs. Journal of Personality and Social Psychology, 42, 322334. Lau, R. R., Hartman, K. A., & Ware, J. E. (1986). Health as value: methodological and theoretical considerations. Health Psychology, 5, 2543. Lazarus, R. S. & Launier, R. (1978). Stress related transactions between person and environment. In L. A. Pervin & M. Lewis (Eds.), Perspectives in interactional psychology (pp. 287327). New York: Plenum. Lefcourt, H. M. (1991). Locus of control. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes (pp. 661753). New York: Academic Press. Leitner, M., Shapland, J., & Wiles, P. (1993). Drug usage and drugs prevention: The views and habits of the general public. London: HMSO. Lennon, D. L. F., Stratman, F. W., Shrago, E., Nagle, F. J., Hanson, P. G., Maddon, M., & Spennetta, T. (1983). Total cholesterol and DHL cholesterol changes during acute, moderate intensity exercise in men and women. Metabolism, 32, 244249. Levenson, H. (1974). Multidimensional locus of control in psychiatric patients. Journal of Consulting and Clinical Psychology, 41, 397404. Leventhal, H. (1970). Findings and theory in the study of fear communications. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 5, pp. 119186). New York: Academic Press. Leventhal, H., Nerenz, D. R., & Steele, D. F. (1984). Illness representations and coping with health threats. In A. Baum & J. Singer (Eds.), A handbook of psychology and health (pp. 219252). Hillsdale, NJ: Erlbaum. Lewin, K. (1951). Field theory in social science. New York: Harper. Liao, K. L., Hunter, M., & Weinman, J. (1995). Healthrelated behaviours and their correlates in a general population sample of 45-year old women. Psychology and Health, 10, 171184. Loewenstein, G. (1996). Out of control: Visceral influences on behavior. Organizational Behavior and Human Decision Processes, 65, 272292. Madden, T. J., Ellen, P. S., & Ajzen, I. (1992). A comparison of the theory of planned behavior and the theory of reasoned action. Personality and Social Psychology Bulletin, 18, 39. Maddux, J. E. (1993). (Ed.) Self-efficacy, adaptation and adjustment: theory, research and application. New York: Plenum.

33

Maddux, J. E., Norton, L. W., & Stoltenberg, C. D. (1986). Self-efficacy expectancy, outcome expectancy, and outcome value: relative effects on behavioral intentions. Journal of Personality and Social Psychology, 51, 783789. Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self-efficacy: A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology, 19, 469479. Maibach, E. W., Flora, J. A., & Nass, C. (1991). Changes in self-efficacy and health behavior in response to a minimal community health campaign. Health Communication, 3, 115. Manning, M. M. & Wright, T. L. (1983). Self-efficacy expectancies, outcome expectancies, and the persistence of pain control in childbirth. Journal of Personality and Social Psychology, 54, 421431. Marcus, B. H., Eaton, C. A., Rossi, J. S., & Harlow, L. L. (1994). Self-efficacy, decision-making, and stages of changeAn integrative model of physical exercise. Journal of Applied Social Psychology, 24, 489508. Marteau, T. M. (1989). Health beliefs and attributions. In A. K. Broome (Ed.), Health psychology: processes and applications (pp. 123). London: Chapman Hall. McAuley, E. (1993). Self-efficacy and the maintenance of exercise participation in older adults. Journal of Behavioral Medicine, 16, 103113. McCaul, K. D., Glasgow, R. E., & O'Neill, H. K. (1992). The problem of creating habits: establishing healthprotective dental behaviors. Health Psychology, 11, 101110. McCaul, K. D., Sandgren, A. K., O'Neill, H. K., & Hinsz, V. B. (1993). The value of the theory of planned behavior, perceived control, and self-efficacy expectations for predicting health-protective behaviors. Basic and Applied Social Psychology, 14, 231252. McCelland, D. C. (1985). How motives, skills, and values determine what people do. American Psychologist, 40, 812825. McConnaughly, E. A., DiClemente, C. C., Prochaska, J. O., & Velicer, W. F. (1989). Stages of change in psychotherapy: A follow-up report. Psychotherapy, 26, 494503. McCusker, J., & Morrow, G. (1979). The relationship of health locus of control to preventive health behaviors and health beliefs. Patient Counselling and Health Education, 1, 146150. McNeil, A. D., Jarvis, M. J., Stapleton, J. A., Russell, M. A. H., Eiser, J. R., Gammage, P., & Gray, E. M. (1988). Prospective study of factors predicting uptake of smoking in adolescents. Journal of Epidemiology and Community Health, 43, 7278. Mechanic, D., & Cleary, P. D. (1980). Factors associated with the maintenance of positive health behavior. Preventive Medicine, 9, 805814. Metzner, H. L., Carman W. J., & House, J. (1983). Health practices, risk factors and chronic disease in Tecumeseh. Preventive Medicine, 12, 491507. Mittag, W. & Schwarzer, R. (1993). Interaction of employment status and self-efficacy on alcohol consumption: A two-wave study on stressful life transitions. Psychology and Health, 8, 7787. Moore, L., Smith, C., & Catford, J. (1994). Binge drinking: prevalence, patterns and policy. Health Education Research, 9, 497505. Morrison, D. M., Gillmore, M. R., & Baker, S. A. (1995). Determinants of condom use among high-risk heterosexual adults: A test of the theory of reasoned action. Journal of Applied Social Psychology, 25, 651676. Muhlenkamp, A. F., Brown, N. J., & Sands, D. (1985). Determinants of health promotion activities in nursing clinic clients. Nursing Research, 34, 327333. Mullen, P. D., Hersey, J. C., & Iverson, D. C. (1987).

34

Health Behavior
cervical screening uptake. British Journal of Health Psychology, 1, 3550. Orbell, S., Hodgkins, S., & Sheeran, P. (1997). Translating intentions into action. Personality and Social Psychology Bulletin, 23, 953962. Paisley, C. M., & Sparks, P. (1998). Expectations of reducing fat intake: The role of perceived need with the theory of planned behaviour. Psychology and Health, 13. Parker, D., Manstead, A. S. R., & Stradling, S. G. (1995). Extending the theory of planned behaviour: The role of personal norm. British Journal of Social Psychology, 34, 127137. Parker, D., Manstead, A. S. R., & Stradling, S. G. (1996). Modifying beliefs and attitudes to exceed the speed limit: An intervention study based on the theory of planned behavior. Journal of Applied Social Psychology, 26, 119. Parker, D., Manstead, A. S. R., Stradling, S. G., Reason, J. T., & Baxter, J. S. (1992). Intention to commit driving violationsAn application of the theory of planned behavior. Journal of Applied Psychology, 77, 94101. Peak, H. (1955). Attitude and motivation. In M. R. Jones (Ed.), Nebraska symposium on motivation (Vol. 3, pp. 149188). Lincoln, NE: University of Nebraska Press. Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes of attitude change. New York: Springer-Verlag. Pomazal, R. P., & Jaccard, J. J. (1976). An informational approach to altruistic behavior. Journal of Personality and Social Psychology, 33, 317326. Prochaska, J. O., & DiClemente, C. C. (1984). The transtheoretical approach: Crossing traditional boundaries of therapy. Homewood, IL: Dow Jones Irwin. Prwun, J., van der Borne, H., de Reuver, R., de Boer, M., Ter Pelkwijk, M., & de Jong, P. (1988). The locus of control scale for cancer patients. Tijdscrift vour Sociale Gezondherdszong, 66, 404408. Quadrel, M. J. & Lau, R. R. (1989). Health promotion, health locus of control, and health behavior: Two field experiments. Journal of Applied Social Psychology, 18, 14971521. Quine, L., Rutter, D. R., & Arnold, L. (1998). Predicting safety helmet use among schoolboy cyclists: A comparison of the theory of planned behavior and the health belief model. Psychology and Health, 13. Raats, M. M. (1992). The role of beliefs and sensory to milk in determining the selection of milks of different fat content. Unpublished doctoral thesis, University of Reading. Redeker, N. S. (1989). Health beliefs, health locus of control, and the frequency of practice of breast selfexamination in women. Journal of Obstetrics, Gynaecology and Neonatal Nursing, 18, 4551. Redhead, I. H. (1960). The incidence of glycosuria and diabetes mellitus in general practice. British Medical Journal, 1, 695. Reiss, I. L., & Leik, R. K. (1989). Evaluating strategies to avoid AIDS: Numbers of partners versus use of condoms. Journal of Sex Research, 26, 411433. Richard, R., & van der Pligt, J. (1991). Factors affecting condom use among adolescents. Journal of Community and Applied Social Psychology, 1, 105116. Richard, R., van der Pligt, J., & de Vries, N. (1995). Anticipated affective reactions and prevention of AIDS. British Journal of Social Psychology, 34, 921. Richard, R., van der Pligt, J., & de Vries, N. (1996a). Anticipated affect and behavioral choice. Basic and Applied Social Psychology, 18, 111129. Richard, R., Van der Pligt, J., & de Vries, N. (1996b). Anticipated regret and time perspective: Changing sexual risk-taking behavior. Journal of Behavioral Decision Making, 9, 185199. Rigotti, N. A. (1989). Cigarette smoking and body weight.

Health behavior models compared. Social Science and Medicine, 24, 973981. Nemeck, M. A. (1990). Health beliefs and breast selfexamination among black women. Health Values, 14, 4152. Netemeyer, R. G., Burton, S., & Johnston, M. (1991). A comparison of two models for the prediction of volitional and goal-directed behaviors: A confirmatory analysis approach. Social Psychology Quarterly, 54, 87100. Nicassio, P., Wallston, K., Callahan, L., Herbert, M., & Pincus, T. (1985). The measurement of helplessness in rheumatoid arthritis: The development of the arthritis helplessness scale. Journal of Rheumatology, 12, 462467. Norman, P. (1990). Health locus of control beliefs and preventive health behaviours. Journal of the Institute of Health Education, 28, 113117. Norman, P. (1991). Social learning theory and the prediction of attendance at screening. Psychology and Health, 5, 231239. Norman, P. (1995). Heath locus of control and health behaviour: an investigation into the role of health value and behaviour-specific efficacy beliefs. Personality and Individual Differences, 18, 213218. Norman, P., & Bennett, P. (1996). Health locus of control. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 6294). Buckingham, UK: Open University Press. Norman, P., & Conner, M. (1993). The role of social cognition models in predicting attendance at health checks. Psychology and Health, 8, 447462. Norman, P., & Conner, M. (1996a). The role of social cognition models in predicting health behaviours: future directions. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 197225). Buckingham, UK: Open University Press. Norman, P., & Conner, M. (1996b). Predicting healthcheck attendance among prior attenders and nonattenders: The role of prior behavior in the theory of planned behavior. Journal of Applied Social Psychology, 26, 10101026. Norman, P., & Smith, L. (1995). The theory of planned behaviour and exercise: An investigation into the role of prior behaviour, behavioural intentions and attitude variability. European Journal of Social Psychology, 25, 403415. Nucifora, J., Gallois, C., & Kashima, Y. (1993). Influences on condom use among undergraduates: Testing the theories of reasoned action and planned behaviour. In D. J. Terry, C. Gallois & M. McCamish (Eds.), The theory of reasoned action: Its application to AIDS-preventive behaviour (pp. 4764). Oxford, UK: Pergamon. Nutbeam, D., & Catford, J. (1990). Modifiable risks for cardiovascular disease among general practitioners in Wales. Public Health, 104, 353361. Oberman, A. (1985). Exercise and the primary prevention of cardiovascular disease. American Journal of Cradiology, 55, 10D20D. O'Connell, J. K., & Price, J. H. (1982). Health locus of control of physical fitness program participants. Perceptual and Motor Skills, 55, 925926. O'Leary, A. (1985). Self-efficacy and health. Behavioral Research and Therapy, 23, 437451. O'Leary, A., & Raffaelli, M. (1996). Preventing the sexual transmission of HIV: Current status and future directions. Psychology and Health, 11, 7594. Office of Population Census and Surveys (1968). Drug use in the UK. London: Government Statistical Services. Oliver, R. L. & Berger, P. K. (1979). A path analysis of preventive health care decision models. Journal of Consumer Research, 6, 113122. Orbell, S., Crombie, I., & Johnson, G. (1996). Social cognition and social structure in the prediction of

References
New England Journal of Medicine, 320, 931933. Rippetoe, P. A., & Rogers, R. W. (1987). Effects of components of protection motivation theory on adaptive and maladpative coping with a health threat. Journal of Personality and Social Psychology, 52, 596604. Rodin, J., & Salovey, P. (1989). Health psychology. Annual Review of Psychology, 40, 533579. Rogers, R. W. (1983). Cognitive and physiological processes in fear appeals and attitude change: A revised theory of protection motivation. In J. T. Cacioppo & R. E. Petty (Eds.), Social psychophysiology: a source book (pp. 153176). New York: Guilford Press. Ronis, D. L., & Harel, Y. (1989). Health beliefs and breast self-examination behaviours: Analysis of linear structural relations. Psychology and Health, 3, 259285. Ronis, D. L., Yates, J. F., & Kirscht, J. P. (1989). Attitudes, decisions and habits as determinants of repeated behavior. In A. R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure and function (pp. 213239). Hillsdale, NJ: Erlbaum. Rosen, T. J., & Shipley, R. H. (1983). A stage analysis of self-initiated smoking reductions. Addictive Behaviors, 8, 263272. Rosenstock, I. M. (1966). Why people use health services. Millbank Memorial Fund Quarterly, 44, 94124. Rosenstock, I. M. (1974). Historical origins of the health belief model. Health Education Monographs, 2, 18. Rosenstock, I. M., Strecher, V. J., & Becker, H. M. (1988). Social learning theory and the health belief model. Health Education Quarterly, 15, 175183. Rosenthal, M., Haskell, W. L., Solomon, R., Widstrom, A., & Reavan, G. M. (1983). Demonstration of a relationship between level of physical training and insulin stimulated glucose utilization in normal humans. Diabetes, 32, 408411. Roskam, S. (1986). Application of health locus of control typology toward predicting depression and medical adherence in rheumatoid arthritis. Unpublished doctoral dissertation, Vanderbilt University, Nashville, TN. Rotter, J. B. (1954). Social learning and clinical psychology. Englewood Cliffs, NJ: Prentice-Hall. Rotter, J. B. (1966). Generalized expectancies for internal and external control of reinforcement. Psychological Monographs: General and Applied, 80 (whole no. 609), 128. Royal College of Physicians (1983). Health or smoking: Follow-up report. London: Pitman. Rutter, D. R., & Quine, L. (1996). Social psychological mediators of the relationship between demographic factors and health outcomes: A theoretical model and some preliminary data. Psychology and Health, 11, 522. Sachs, M. L. (1984). Psychological well-being and vigorous activity. In J. M. Silva & R. S. Weinberg (Eds.), Psychological foundations of sport (pp. 435444). Champaign, IL: Human Kinetics. Sallis, J. F., Howell, M. F., Hofstetter, C. R., & Barrington, E. (1992). Explanation of vigorous physical activity during two years using social learning variables. Social Science and Medicine, 34, 2532. Saltzer, E. B. (1982). The weight locus of control (WLOC) scale: A specific measure for obesity research. Journal of Personality Assessment, 46, 620628. Sampos, C., Fulwood, R., Haines, C., Carroll, M., Anda, R., Williamson, D. F., Remington, P., & Cleeman, J. (1989). The prevalence of high blood cholesterol levels among adults in the United States. Journal of the American Medical Association, 262, 4552. Sandman, P. M., & Weinstein, N. D. (1993). Predictors of home radon testing and implications for testing promotion programs. Health Education Quarterly, 20, 471487. Scheier, M. F. & Carver, C. S. (1992). Effects of optimism on psychological and physical well-being: Theoretical

35

overview and empirical update. Cognitive Therapy and Research, 16, 201228. Schifter, D. E., & Ajzen, I. (1985). Intention, perceived control and weight loss: An application of the theory of planned behaviour. Journal of Personality and Social Psychology, 49, 843851. Schlegel, R. P., D'Avernas, J. R., Zanna, M. P., DeCourville, N. H., & Mansue, S. R. (1992). Problem drinking: A problem for the theory of reasoned action? Journal of Applied Social Psychology, 22, 358385. Schwartz, S. H., & Tessler, R. C. (1972). A test of a model for reducing measured attitude-behavior discrepancies. Journal of Personality and Social Psychology, 24, 225236. Schwarzer, R. (1992). Self-efficacy in the adoption and maintenance of health behaviors: Theoretical approaches and a new model. In R. Schwarzer (Ed.), Self-efficacy: thought control of action (pp. 217243). London: Hemisphere. Schwarzer, R. (1994). Optimism, vulnerability, and selfbeliefs as health-related cognitions: a systematic overview. Psychology and Health, 9, 161180. Schwarzer, R., & Fuchs, R. (1996). Self-efficacy and health behaviour. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 163196). Buckingham, UK: Open University Press. Seeman, M., & Seeman, T. E. (1983). Health behaviour and personal autonomy: A longitudinal study of the sense of control in illness. Journal of Health and Social Behaviour, 24, 144160. Segall, M. E., & Wynd, C. A. (1990). Health conception, health locus of control, and power as predictors of smoking behavior change. American Journal of Health Promotion, 4, 338344. Seydel, E., Taal, E., & Wiegman, O. (1990). Risk-appraisal, outcome and self-efficacy expectancies: Cognitive factors in preventive behaviour related to cancer. Psychology and Health, 4, 99109. Shaper, A. G., Pocock, S. J., Walker, M., Cohen, N. M., Wale, C. J., & Thomson, A. G. (1981). British Regional Heart Survey: Cardiovascular risk factors in middle-aged men in 24 towns. British Medical Journal, 283, 179186. Sheeran, P., & Abraham, C. (1996). The health belief model. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 2361). Buckingham, UK: Open University Press. Sheppard, B. H., Hartwick, J., & Warshaw, P. R. (1988). The theory of reasoned action: A meta-analysis of past research with recommendations for modifications and future research. Journal of Consumer Research, 15, 325343. Sherlock, S. (1982). (Ed.) Alcohol and Disease. British Medical Bulletin, 38 (whole issue). Simonson, I. (1992). The influence of anticipating regret and responsibility on purchase decisions. Journal of Consumer Research, 19, 105118. Singer, R. N. (1992). Physical activity and psychological benefits: A position statement. The Sport Psychologist, 6, 199203. Slenker, S. E., Price, J. H., & O'Connell, J. K. (1985). Health locus of control of joggers and nonexercisers. Perceptual and Motor Skills, 61, 323328. Snyder, C. R., Harris, C., Anderson, J. R., Holleran, S. A., Irving, L. M., Sigmon, S. T., Yoshinobu, L., Gibb., J., Langelle, C., & Harney, P. (1991). The will and the ways: Development and validation of an individual-differences measure of hope. Journal of Personality and Social Psychology, 60, 570585. Sparks, P. (1994). Food choice and health: Applying, assessing, and extending the theory of planned behaviour. In D. R. Rutter & L. Quine (Eds.), Social psychology and health: European perspectives (pp. 25 46). Aldershot, UK: Avebury.

36

Health Behavior
Taylor, S. E. (1991). Health psychology. New York: McGraw-Hill. Terry, D. J., & O'Leary, J. E. (1995). The theory of planned behaviour: The effects of perceived behavioural control and self-efficacy. British Journal of Social Psychology, 34, 199220. Terry, D. J., & Hogg, M. A. (1996). Group norms and the attitude-behavior relationship: A role for group identification. Personality and Social Psychology Bulletin, 22, 776793. Theodorakis, Y. (1994). Planned behavior, attitude strength, role identity, and the prediction of exercise behavior. The Sport Psychologist, 8, 149165. Triandis, H. C. (1977). Interpersonal behavior. Monterey, CA: Brooks-Cole. US Department of Health and Human Services (1988). The Surgeon General's report on nutrition and health. Washington, DC: US Government Printing Office. US Department of Health and Human Services (1990). Health people 2000: National health promotion and disease prevention objectives. Washington, DC: US Government Printing Office. Valois, P., Desharnais, R., & Godin, G. (1988). A comparison of the Fishbein and the Triandis attitudinal models for the prediction of exercise intention and behavior. Journal of Behavioral Medicine, 11, 459472. Van den Putte, H. (1993). On the theory of reasoned action. Unpublished doctoral dissertation, University of Amsterdam. Van der Pligt, J., & de Vries, N. (1998). Expectancy-value models of health behaviours: The role of salience and anticipated affect. Psychology and Health, 13. Van der Pligt, J., & Eiser, J. R. (1984) Dimensional salience, judgement and attitudes. In J. R. Eiser (Ed.), Attitudinal judgement (pp. 4363). New York: SpringerVerlag. Van der Velde, W., & van der Pligt, J. (1991). AIDS-related behavior: Coping, protection motivation, and previous behavior. Journal of Behavioral Medicine, 14, 429451. Vernon, S. W., Laville, E. A., & Jackson, G. L. (1990). Participation in breast screening programs: a review. Social Science and Medicine, 30, 11071118. Waldron, I. (1988). Why do women live longer than men? Journal of Human Stress, 2, 213. Waller, K. V., & Bates, R. C. (1992). Health locus of control and self-efficacy beliefs in a healthy elderly sample. American Journal of Health Promotion, 6, 302309. Wallston, B. S., Wallston, K. A., Kaplan, G. D., & Maides, S. A. (1976). Development and validation of the health locus of control (HLC) scale. Journal of Consulting and Clinical Psychology, 44, 580585. Wallston, K. A. (1991). The importance of placing measures of health locus of control beliefs in a theoretical context. Health Education Research: Theory and Practice, 6, 251252. Wallston, K. A. (1992). Hocus-pocus, the focus isn't strictly on locus: Rotter's social learning theory modified for health. Cognitive Therapy and Research, 16, 183199. Wallston, K. A., Maiders, S., & Wallston, B. S. (1976). Health-related information seeking as a function of health-related locus of control and health value. Journal of Research in Personality, 10, 215222. Wallston, K. A., & Wallston, B. S. (1980). Health locus of control scales. In H. Lefcourt (Ed.), Advances and innovations in locus of control research (pp. 198234). New York: Academic Press. Wallston, K. A., & Wallston, B. S. (1981). Health locus of control scales. In H. Lefcourt (Ed.), Research with the locus of control construct (Vol. 1, pp. 189243). New York: Academic Press. Wallston, K. A., & Wallston, B. S. (1982). Who is responsible for your health? The construct of health

Sparks, P., & Shepherd, R. (1992). Self-identity and the theory of planned behaviour: Assessing the role of identification with green consumerism. Social Psychology Quarterly, 55, 388399. Sparks, P., Shepherd, R., & Frewer, L. J. (1995). Assessing and structuring attitudes toward the use of gene technology in food productionThe role of perceived ethical obligation. Basic and Applied Social Psychology, 16, 267285. Sparks, P., Shepherd, R., Wieringa, N., & Zimmermanns, N. (1994). Barriers to healthy eating: An examination of perceived behavioural control and unrealistic optimism. Paper presented at the BPS Social Psychology Section Annual Conference, University of Hertfordshire, UK, September 1994. Sparks, P., Shepherd, R., Wieringa, N., & Zimmermanns, N. (1995). Perceived behavioural control, unrealistic optimism and dietary change: An exploratory study. Appetite, 24, 243355. Speake, D. L., Cowart, M. E., & Stephens, R. (1991). Healthy lifestyle practices of rural and urban elderly. Health Values, 15, 4551. Stacy, R. D., & Lloyd, B. H. (1990). An investigation of beliefs about smoking among diabetes patients: Information for improving cessation efforts. Journal of Patient Education and Counselling, 15, 181189. Stainback, R. D., & Rogers, R. W. (1983). Identifying effective components of alcohol abuse prevention programs: effects of fear appeals, message style and source expertise. International Journal of Addictions, 18, 393405. Stanley, M. A. & Maddux, J. E. (1986). Cognitive processes in health enhancement: Investigation of a combined protection motivation and self-efficacy model. Basic and Applied Social Psychology, 7, 101113. Stanton, A. (1987). Determinance of adherence to medical regimens by hypertensive patients. Journal of Behavioural Medicine, 10, 377394. Steptoe, A., & Wardle, J. (1992). Cognitive predictors of health behaviour in contrasting regions of Europe. British Journal of Clinical Psychology, 31, 485502. Steptoe, A., & Wardle, J. (1996). The European health and behaviour survey: The development of an international study in health psychology. Psychology and Health, 11, 4973. Steptoe, A., Wardle, J., Vinck, J., Tuomisto, M., Holte, A., & Wickstrom, L. (1994). Personality and attitudinal correlates of healthy and unhealthy lifestyles in young adults. Psychology and Health, 9, 331343. Stevens, V. J. & Hollis, J. F. (1989). Preventing smoking relapse, using an individually tailored skills-training technique. Journal of Consulting and Clinical Psychology, 57, 420424. Strax, P. (1984). Mass screening for control of breast cancer. Cancer, 53, 665670. Stroebe, W., & Stroebe, M. S. (1995). Social psychology and health. Buckingham, UK: Open University Press. Sutton, S. (1982). Fear-arousing communications: A critical examination of theory and research. In J. R. Eiser (Ed.), Social psychology and behavioural medicine (pp. 303337). London: Wiley. Sutton, S. (1994). The past predicts the future: Interpreting behaviourbehaviour relationships in social psychological models of health behaviours. In D. R. Rutter & L. Quine (Eds.), Social psychology and health: European perspectives (pp. 7188). Aldershot, UK: Avebury. Sutton, S., & Hallett, R. (1989). Understanding seat-belt intentions and behavior: A decision-making approach. Journal of Applied Social Psychology, 19, 13101325. Taylor, D. W. (1979). A test of the health belief model in hypertension. In R. B. Haynes, D. W. Taylor, & K. L. Sackett (Eds.), Compliance in health care (pp. 103109). Baltimore: Johns Hopkins University Press.

References
locus of control. In G. Sanders & J. Suls (Eds.), Social psychology of health and illness (pp. 6595). Hillsdale, NJ: Erlbaum. Wallston, K. A., Wallston, B. S., & DeVellis, R. (1978). Development of multidimensional health locus of control (MHLC) scales. Health Education Monographs, 6, 160170. Wardle, J. (1995). Cholesterol and psychological wellbeing. Journal of Psychosomatic Research, 39, 549562. Warshaw, P. R., & Davis, F. D. (1985). Disentangling behavioral intention and behavioral expectation. Journal of Experimental Social Psychology, 21, 213228. Weatherburn, P., Hunt, A. J., Davies, P. M., Coxon, A. P. M., & McManus, T. J. (1991). Condom use in a large cohort of homosexually active men in England and Wales. AIDS Care, 3, 3141. Weinstein, N. D. (1988). The precaution adoption process. Health Psychology, 7, 355386. Weinstein, N. D. & Nicolich, M. (1983). Correct and incorrect interpretations of correlations between risk perceptions and risk behaviors. Health Psychology, 12, 235245. Weinstein, N. D., Rothman, A. J., & Sutton, S. R. (in press). Stage theories of health behavior. Health Psychology. Weinstein, N. D., & Sandman, P. M. (1992). A model of the precaution adoption process: evidence from home radon testing. Health Psychology, 11, 170180. Weiss, G. L., & Larsen, D. L. (1990). Health value, health locus of control, and the prediction of health protective

37

behaviours. Social Behavior and Personality, 18, 121135. Wilson, D., Zenda, A., McMaster, J., & Lavelle, S. (1992). Factors predicting Zimbabwean students' intentions to use condoms. Psychology and Health, 7, 99114. Winett, R. A. (1985). Ecobehavioral assessment in health life-styles: Concepts and methods. In P. Karoly (Ed.), Measurement strategies in health psychology (pp. 147 181). Chichester, UK: Wiley. Wojcik, J. V. (1988). Social learning predictors of the avoidance of smoking relapse. Addictive Behaviors, 13, 177180. Wurtele, S. K. (1988). Increasing women's calcium intake: the role of health beliefs, intentions and health value. Journal of Applied Social Psychology, 18, 627639. Wurtele, S. K., Britcher, J. C., & Saslawsky, D. A. (1985). Relationships between locus of control, health value and preventive health behaviors among women. Journal of Research in Personality, 19, 271278. Wurtele, S. K., & Maddux, J. E. (1987). Relative contributions of protection motivation components in predicting exercise intentions and behavior. Health Psychology, 6, 453466. Young, H. M., Lierman, L., Powell-Cope, G., Kasprzyk, D., & Bonoliel, J. Q. (1991). Operationalizing the theory of planned behaviour. Research in Nursing and Health, 14, 137144. Zuckerman, M., & Reiss, H. T. (1978). Comparison of three models for predicting altruistic behavior. Journal of Personality and Social Psychology, 36, 498510.