Psychological Reports, 2003,92, 809-819.

O Psychological Reports 2003

DIMENSIONALITY OF MILITARY JOB SATISFACTION ITEMS: AN

EXPLORATORY FACTOR ANALYSIS OF DATA FROM THE SPRING
1996 SAMPLE SURVEY OF MILITARY PERSONNEL '

WALTER R. SCHUMM

Kanansas State University

PAUL A. GADE, AND D. BRUCE BELL

U.S.A m y Research Institule

Summary.-The Sample Survey of Military Personnel in the spring of 1996 asked

detailed questions of Army personnel about job sadsfacdon. Maximum likehood fac-
tor analysis yielded four factors: satisfaction with supervision, with job environment,
with the duty specifically, and with opportunity for development, factors similar to
those used in Smith's Job Descriptive Index, 1992. Scales developed from these fac-
tors showed acceptable internal consistency reliabihty and correlated as expected with
measures of satisfaction with Army Me, retention, morale, combat preparedness, and
overall job sadsfacdon. Some divergent validity was established with measures of per-
sonaUfamily stress and with previous job stress. Demographic data suggested that sol-
diers with higher education found more satisfying job opportunities while those with
higher rank, more years of service, and higher age reported greater satisfaction with
job h l f i e n t , perhaps a consequence of selection effects.

Job satisfaction has been an important concept in civilian and mhtary

research. It has been related to better health and higher Me satisfaction
(Wdiarnson, 1996) as well as reduced job turnover (Brief, 1998) although,
unfortunately, job satisfaction has been declining in both civilian (Brief,
1998, p. 177) and rnhtary (U.S. Army Research Institute, 2001) sectors.
However, in much c i d a n and mihtary research it has often been conceptu-
alized as a unidimensional construct (Wood & LeBold, 1970), even though
recent research has demonstrated that the concept is multidimensional, at
least for engineers (Wood - & LeBold, 1970), secondary school teachers
(Knoop, 1994), elernentary/junior high teachers (Lester, 19871, social senrice
workers (Wdharnson, 1996), and special education teachers (Abelson, 1986).
Smith (1992) reported that job satisfaction includes five facets-satisfaction
with work, promotion, pay, supervision, and coworkers, as measured by the
Job Descriptive Index (Smith, Kendall, & Hulin, 1969). James and James

'The opinions or assertions contained herein are the private views of the authors and are not to
be construed as official or as reflecting the views of the Deparunent of the Army or of [he
Deparunent of Defense of che United States of America. Our gratitude is ex ressed to Dr.
Morris Peterson for permitting use of the data. Requests for re rints should ge sent to Dr.
Paul A. Gade, U.S. Army Research Institute for the Behavioral a n j ~ o c i a Sciences,
l 5001 Eisen-
hower Avenue, Alexandria, VA 22333-5600.
810 W. R. SCHUMM, ET AL.

(1992) found four factors in job satisfaction-work satisfaction, satisfaction

with work group (coworkers), satisfaction with organization (prestige, train-
ing received, and opportunity for promotion and growth), and satisfaction
with supervisors. However, global job satisfaction, although related to facets
of job satisfaction, is not equivalent to the sum of scales measuring those
facets (Brief, 1998, p. 15). Further, some of the more recent large sample
surveys conducted by the U.S. Army have not assessed job satisfaction either
drectly, e.g., the 1989 Army Family Research Program survey of 11,035 sol-
diers, or if directly, not in terms of multiple dimensions (the 1992 Depart-
ment of Defense Survey of Officers and Enlisted Personnel). However, job
satisfaction has been linked to retention within the military system (Kocher
& Thomas, 1994; Kim, Price, Mueller. & Watson, 1996), as well as civhan
occupations (Fuller, Hester, Dickson. Ahson, & Birdseye, 1996). Rohall, Se-
gal, and Segal (1999, p. 56) h k e d rerenrlon, morale, preparedness, and job
satisfaction so closely they felt comfortable combining those types of ques-
tions within one scale for solder morale. Although we interpreted morale, re-
tention, readiness, and job satisfaction as distinct, separate concepts, we also
anticipated that job satisfaction would be correlated with indvidual morale
and with satisfaction with overall quality of life in the Army as well as with
retention. The goal of this study was to use an extant data set that had a va-
riety of measures of job satisfaction to assess the dirnensionahty of those
measures and to assess how any dimensions might be related to key mhtary
outcome variables such as retention, morale, satisfaction with the quality of
Army life, overall job satisfaction, and wartime preparedness.

Fourteen job satisfaction items were selected for data-reduction analysis.

The section of the questionnaire contains a total of 54 items. The other
items were not assessed because they assessed employment benefits (e.g., sat-
isfaction with medical care), situations that were not relevant to all soldiers
(e.g., living condtions overseas, because not all soldiers have lived overseas),
or contained a very high amount of missing data (e.g., retirement benefits,
amount of time separated from family). One item was not used because it
was a key outcome measure for several aspects of military life in addition to
job satisfaction (satisfaction with overall quahty of Army kfe). All items fea-
tured 4-point response categories, 4: very satisfied, 3: satisfied, 2: dissatis-
fied, and 1: very dissatisfied (and a not applicable category, converted into
missing data). For example, observed satisfaction with amount of enjoyment
from my job were 19.3% very satisfied, 51.5% satisfied, 20.6% dissatisfied,
and 8.7% very dissatisfied.
Several single items available in the survey were expected to be corre-
 MILITARY JOB SATISFACTION ITEMS 811

lated with at least some of the possible dimensions of these professional val-
ues. Those items included the following variables, with their response cate-
gories and numerical codes.
Satisfaction with overall quality of Army life
Very dissatisfied (1)
Dissatisfied (2)
Satisfied (3)
Very Satisfied (4)
Retention Intentions
Definitely leave upon completion of present obhgdr~on(1)
Probably leave upon completion of present obl~gat~on (2)
Probably stay in beyond my present obligation, but not necessarily to retirement (3)
Definitely stay in beyond my present obligation, but not necessarily to retirement (4)
Probably stay in until retirement (5)
Definitely stay in until reurement (6)
Perceived Preparedness for Performing Wartime Duties/Mission (for soldier and or his per-
ception of his unit)
Not at all prepared ( I )
Not well prepared (2)
Moderately prepared (3)
Well prepared (4)
Very well prepared ( 5 )
Current Individual Morale
Very low (1)
Low (2)
Moderate (3)
High (4)
Very high (5)
Current Job Satisfaction
Very dissatisfied (1)
Dissatisfied (2)
Neither satisfied nor dissatisfied (3)
Satisfied (4)
Very satisfied ( 5 )
Unit Morale (same responses as for individual morale)
Personal Readiness or Deployment (same responses as for preparedness for wartime du-
ties/rnission)
Current Military Job Stress
None (1)
Slight (2)
Moderate (3)
High (4)
Very high (5)
Extremely high (6)
Current Family/l?ersonal Life Stress (same responses as for military job stress)
Military Job Stress a year ago (same responses as for current military job stress)
Family/Personal Life Stress a year ago (same responses as for current family/personal life
stress)
812 W. R. SCHUMM. ET AL.

A variety of demographic variables concerning sex, age, education, m h -

tary rank, racial heritage, and years of mditary service were included in the
survey.
Data Collection
The Army Personnel Survey Office of the U.S. Army Research Lnstitute
for the Behavioral and Social Sciences under the U.S. Total Army Personnel
Command conducts the Sample Survey of Mhtary Personnel nvice a year.
The survey is administered to a selected sample of permanent party Army
personnel in Grades E-2 (private) to 0 - 6 (colonel), who are avadable for
duty.
The samples for the survey are drawn randomly from the Standard In-
stdation/Division Personnel System using the final one or two digits of
soldiers' social security numbers, including d permanent party, Active Com-
ponent Army personnel (excluding all PVl/El soldiers and PV2/E2 solders
in Europe and Korea). Soldiers in Europe and Korea are sent surveys by di-
rect mail. Those elsewhere are administered surveys by their installation
Personnel Survey Control Officer (PSCO). For the Spring 1996 survey, ap-
proximately 6% of Army active duty personnel in eligible paygrades were
sampled; of those, 7,860 responded, estimated to be approximately half of
those el~g~ble who received correctly addressed surveys. Approximately 10%
of off~cersand 2-3% of enlisted personnel are sampled. Sampling errors are
estimated to be f 2-3% for most ranks except for warrant officers (+_4%).
The Army Personnel Survey Office and the Personnel Service Commands are
not resourced to track the number of unavailable soldiers from each com-
mand and have not established the actual response rate, but it is thought to
be approximately 50%.
Sample Characteristics
Normally, survey data are weighted to permit generalization to the en-
tire U.S. Army, to a total of 465,801 cases for the Spring 1996 SSMP. How-
ever, for purposes of our analysis, we used unweighted data from 7,860 sol-
diers. Most of these respondents were males (87.3%), Euro-Americans
(75.1%), African Americans (19.2%),or were commissioned officers (39.1%)
or enlisted personnel (53.7%). Most of the soldiers were married (54.3%),
remarried (13.9%), or single (23.3%) with a few widowed (0.2%), dvorced
(5.6%), or separated (2.6%). Among those not married, 44.4% were en-
gaged or involved in a significant relationship with a boyfriend or girlfriend.
Some had one (18.1%), two (23.8%), or three or more (15.4%) children. A
few were single parents (2.9%) with custody. Some soldiers had more than
20 yr. of mibtary service (18.7%), while others had less than 5 yr. service
(30.8%) or between 5 and 9 yr. (17.0%) service. In terms of education, the
subjects had the following attainments: high school diploma/GED (18.2%),
 MILITARY JOB SATISFACnON ITEMS 813

some college (36.1%), a college degree (24.4%), some graduate study

(4.5%), or a graduate degree (17.7%). In terms of age, the soldiers reported
the following categories: under 20 years (3.9%), 20-24 years (20.7%), 25-29
years (18.0%), 30-34 years (15.7%), 35-39 years (16.8%), 40-44 years
(13.5%), 45-49 years (9.0%), and 50 years or older (2.4%). As noted else-
where (Schumm, Gade, & Bell, 2003), weighted and unweighted data ~ i e l d
relatively similar percentage breakdowns by race, sex, and rank.
Analysis
The 14 items pertaining to job satisfaction were subjected to a maxi-
mum k e l h o o d factor analysis, using varirnax rotation. Varirnax rotation was
preferred over oblique rotations given its simplicity and our working as-
sumption that the factors be uncorrelated (McDonald, 1985, pp. 85-87).
Furthermore, according to Child (1990, pp. 57-58) oblique rotations are s d
quite controversial. Maximum likehhood methods were used to extract fac-
tors, rather than principal components or other common factor approaches,
because goodness of fit tests for comparing different number of factor mod-
els are available in SPSS (Norugis, 1994) only for the maxirnwn hkelihood
approach to factor extraction. Initially, the items were subjected to a three-
factor extraction. However, a four-factor solution greatly reduced residual
errors (>.05) from 12% to 2%. Table 1 below presents the results of the
four-factor solution. Factor items were then subjected to item and rehabhty
analyses. Subsequently, scales were computed from the factored items and cor-
related with satisfaction with overd quality of Army life, retention inten-
tions, perceived preparedness for wartime duties/mission, current individual
morale, current stress in mhtary job, current stress in family/personal Me,
and current overall job satisfaction. We expected that job stress and wartime
preparation would be less strongly correlated with job satisfaction measures
because (1) a soldier can have a very enjoyable job yet not be prepared for
war because the job is not a combat position and (2) because a solher could
have a high stress job that was more enjoyable than some low stress but bor-
ing jobs. Furthermore, we thought that personallfamily stress would be less
related to job satisfaction measures than job stress and that previous job
stress would be less related to current job satisfaction than current job stress.
We expected job satisfaction to be correlated with rank and also, probably,
the related variables of age and years of mhtary service, based on previous
research in which rank was correlated significantly with mhtary job satisfac-
tion (Blount, LeClair, Miser, Manness, Schlrner, Weightman, &Jones, 1995).
Cronbach (1951) alpha was used to estimate mternal consistency reliabhty
for all scales. All analyses were performed using SPSS for Windows (Noru-
Sis, 1994).
 W. R. SCHUMM, ET AL.

Factor Analysis
The factor structure, as shown in Table 1, was relatively clear. None of
the items loaded more than .35 on a secondary factor and but three items
loaded at least .54 on their primary factor. The KMO sampling adequacy
was very good, and the Bartlett Test of Sphericity was significant, as ex-
pected (Dzuiban & Shirkey, 1974). The chi-square goodness of fit improved
considerably by using four (x2=551.22)instead of three (x2=2,346.32)fac-
tors in the extraction. The percentage of residuals in the reproduced correla-
tion matrix was very low. Even though the goodness of fit indicated less than
perfect fit, it is virtually impossible to obtain a perfect fit in such a large
sample, because even very small deviations from fit will be statistically signif-
icant (in this sample, correlations as small as r = .02 were significant).

Satisfaction Item Rotated Factor Loading

1 2 3 4

Level of job hlfillmendchallenge

Amount of enjoyment from job
Use of my skills and training on the job
Level of competence of supervisors
Quality of leadership at your place of duty
Amount of respect from superiors .21 .66 .16 .31
Opportunity to selecr a job, training, or station of my choice .15 .11 .80 .15
Control over my job assignments .24 .15 .60 .15
Assignments to jobs offering technicaVproessional
development
Amount of pay (basic)
Length of working hours
Level of recognition for my accomplishments
Level of fairness in how my performance is evaluated
Promotion/advancement opportunities
Number of Cases
Eigenvalues (unrotated factors)
KMO Sampling Adequacy
Matrix Determinant
Bartlett Test of Sphericicy
Iterations to Convergence 6
Chi-square rest of goodness of fir (df=41) 551.22, p < ,0001
% Residuals in Reproduced Correlation Matrix > .05 2.0

The first factor was labeled Satisfaction with Job F u l f h e n t because

the item content seemed to focus on fulfillment A d enjoyment associated
with the soldier's job. The second factor was labeled Satisfaction with Su-
 MILITARY JOB SATISFACTION ITEMS 815

pervision because items all pertained to the quality of supervision or leader-

ship experienced by the soldier. The third factor was labeled Satisfaction
with Job Opportunities because items concerned opportunity to select or
control one's job assignments in a desirable manner. The fourth factor was
labeled Satisfaction with Recognition because items all referred to recogni-
tion, performance evaluations, and promotion opportunities. The factors ob-
tained here resemble those found in Smith's Job Descriptive Index (19921,
concerning satisfaction with work, supervision, pay, and promotion, respec-
tively.

The three highest loading items for each factor were used to create
summed scales of three items each, associated with each factor. The items
for pay and working hours were not included in the scale for satisfaction
with job opportunities because the loadmgs on their best factor did not
reach even .35. Cronbach (1951) alpha was used to estimate the internal con-
sistency reliabhty of each scale, as shown in Table 2. Item analysis indicated
that no item could be deleted from any of the scales without lowering the
overall alpha level, indicating that each of the three items belonged in each
scale. The internal consistency reliabfity of the scales for job f u l f h e n t and
for supervision were higher than for the other two scales and the standard
deviations were also larger. The mean scores for the scales for job opportu-
nities and for recognition were lower than for the other two scales, suggest-
ing that the Army may be having more success in terms of providing a
chance for job fulfillment and high qu&ty supervision than with providing
choice of job opportunities and appropriate recognition. However, all the
scales showed evidence of adequate to good internal consistency reliability.
TABLE 2
CHARACTEIU~TIC~
OF FOURJOBSATISFACI~ON
SCALES

Scale N M SD a
satisfaction with
Job FulEillment 7,736 8.52 2.23 .86
Supervision 7,697 8.22 2.25 .86
Job Opportunities 7,639 7.20 2.05 .74
Recognition 7,664 7.83 2.04 .75
Note.-Each scale contained three items, each item coded as 1: ve dissatisfied, 2: dissatis-
fied, 3: satisfied, and 4: very satisfied; therefore, scores for each s c z e ranged benveen 3 and
12 points.

Validity
As shown in Table 3 , the four job satisfaction scales are intercorrelated
with each other, as might be expected. The closest item to pure job satisfac-
tion was, indeed, a single-item measure of job satisfaction; it correlated more
816 W. R. SCHUMM. ET AL.

TABLE 3
PEARSONZERO-ORDER
CORRELATIONS SWES WITH~ L ~ T JOB
OF JOB SAT~SFACT~ON E D
OUTCOMEVARIABLES
(EVALUATING
CONVERGEN~ AND DWERGENTVALIDITY)
Outcome Variable Satisfaction With
Fulfillment Suvenlision Ovvortunitv Recognition

Rerenr~onintentions
Satisfaction with overall of
Army Me
Wartime preparedness (soldier)
Wartime preparedness (unit)
Individual morale
Unit morale
Current job stress
Current job satisfaction
Satisfaction with supervision
Satisfaction with job opportunities
Satisfaction with recognition
Divergent Validity
Current personaVfamily srress
Personallfamily stress a year ago
Military job stress a year ago
Demographics
Sex (1=male, 2 =female)
Education
Age of solher
Years of military senlice
Overall rank
Enhsted rank
Officer rank
Nole.-AU correlations are si niEicant ( p < ,001) unless otherwise indicated. Warrant Officer
rank was not correlated sigmficantly with any of the job satisfaction variables. +Nor signifi-
cant. *p<.05.

strongly with job f u l f h e n t than with the other job satisfaction scales, sug-
gesting that satisfaction with job fulfillment was more closely related to job
satisfaction overall than the other three job satisfaction scales. Likewise, re-
cencion intentions, wartime preparedness, individual morale, and sacisfaction
with overall qual~tyof Army life were, with one exception, more strongly cor-
related with job f u l f h e n t satisfaction than with any of the other job satis-
faction scales. As we expected, wartime preparedness and job stress were not
as strongly correlated with the job satisfaction measures, compared to the
other independent variables. It was interesting that the measures of unit mo-
rale and wartime preparedness correlated a bit more strongly with supervi-
sion satisfaction than they did with job f u I f h e n t satisfaction, the reverse of
what was found for the individual measures of those same variables. That dif-
ference may reflect the important role of supervisors in influencing the en-
 MILITARY JOB SATISFACTION ITEMS 817

tire organizational system more consistently than the job satisfaction of any
particular member of an organization.
As expected, personaVfamily stress measures were less related to job
satisfaction than job stress; likewise, prior job stress was less strongly related
to job satisfaction than current stress. Although most of the correlations were
still significant, the much lower correlations for the dvergent vahdity vari-
ables suggest that the convergent validity correlations were not primarily ar-
tifacts of methods variance, though it is possible that methods variance might
have accounted for less than three percent (.I8 x .18) of the variance over-
lap in the worst case.
The demographic variables featured fairly low correlations with job sat-
isfaction. Female soldiers reported slightly lower job satisfaction in terms of
job f u l f h e n t and supervision but there were no significant ( p < .05) sex
differences for the other two aspects of job satisfaction. Those soldiers with
higher education reported slightly higher satisfaction with job opportunities,
probably because their educational attainment opened more doors for them
in terms of career options. Older soldiers, those with higher rank, and those
with more years of service reported higher job fulfillment satisfaction, prob-
ably due in part to a selection effect-those who enjoy their jobs stay in the
military and are promoted while those who do not enjoy their jobs leave the
mhtary and are not promoted.
Further Analyses
We also re-analyzed the correlations in Table 3 , based on rank and sex
(12 analyses) to guard against the lunitations of the analysis in terms of the
structure of the survey data, which is stratified by sex and rank (as well as
location, Europe versus other). However, there were only two discernable
patterns, both of whlch were consistent across sex. First, retention and job
satisfaction variables were not strongly related to each other for senior ranks,
probably because most senior personnel have already decided to stay in until
retirement if possible. For example, the correlations for men and women of
field grade rank between job f u l f h e n t satisfaction and retention were -.I2
and -.14, respectively; for senior noncommissioned officers the correlations
were -.I4 and -.26, respectively.
Secondly, the correlations between current job stress and job f u l f h e n t
satisfaction varied considerably by rank, lowest most notably for junior offic-
ers (lieutenants and captains) where the correlations were only -.18 and
-.17, for male and female officers (for senior officers, the corresponding cor-
relations were -.24 and -.25). Ln contrast, the correlations for enlisted
personnel ranged between -.27 and -.36 for men and between -.27 and -.40
for women. It seems possible that job stress may affect job satisfaction less
for officers, especially junior officers, and more for enlisted personnel. Other
 W. R. SCHUMM, ET AL.

than for the situations just mentioned, it appears that the data in Table 3
can be generalized across the entire sample in spite of the problems of strat-
ification (and our using unweighted data). The data can be generaLzed to
the entire Army except for privates and general officers.
Conclusions
The 14 job satisfaction items appear to be reducible to four three-item
scales that appear to measure distinct, although related, aspects of satisfac-
tion with the soldier's military job. The scales created from the factors all
appear to have acceptable to good estimates of internal consistency reliabil-
ity. While all aspects of job satisfaction appear related to outcome variables
of interest to the U.S. Army, satisfaction with job fulfillment appears to be
most important for individual morale, current overall job satisfaction, satis-
faction with the overall quality of Army Me, wartime preparedness, and re-
tention intentions. Some evidence is provided for convergent vahdity inas-
much as all of the four job satisfaction scales are correlated significantly with
the outcome variables. At the same time, the much lower correlations of the
job satisfaction scales with items about family/personal stress and with job
stress in the previous year provide lunited evidence for divergent v&dity.
The correlations of the demographic variables with the four job satisfaction
scales also provide some support for the scales' vali&ty as education appears
to influence satisfaction with job opportunity more than the other scales
while rank, age, and years of mditary service are strongly correlated with job
fulfhent.
All in all, the core component of job satisfaction in the U.S. Army for
soldiers in this sample appears to be job fulfillment. Possibly that is what is
captured by the popular theme song, "Be all that you can be, in the Army."
Good supervision, recognition, and control over one's opportunities are use-
ful, but job fulfillment appears to be the central factor to positive outcomes
- -

from the Army's perspective. However, current job stress may reduce job
fulfillment satisfaction far more for e k s t e d personnel than for officers.
REFERENCES
ELS SON, A. G. (1986) A [actor-analytic study of job satisfaction among special educators. Edu-
catiorral and Psychological Measurement, 46, 37-43.
BLOUNT,B. W., LECLAIR, B. M., MISER.W. F.. MANNESS, D. L.. SCHIRNER,
W. A,, WELGHTMAN,G.
F., &JONES,R. (1995) Army family physician satisfaction. Militaty Medicine, 160, 501-
505.
A. P. (1998) Attit~rde.7in and around organizations. Thousand Oaks, CA: Sage.
BRIEF,
CHILD.D. (1990) The essentials ofjacfor a~ralysi.~.(2nd ed.) London: Cassell Educational Ltd
CRONBACH, L. J., JR. (1951) Coefficient alpha and the internal structure of resrs. Psj'chomelrika,
16, 297-234.
DZUIBAN.
C. D., &SHIRKEY,
E. C. (1974) When is a correlation matrix appropriate for factor
analysis: some decision rules. Psychological Bz~ffetin,81, 358-361.
J . B., HESTER,K., DICKSON.
FULLER, P., ALLISON,B.. &BIRDSNE,M. (1996) A closer look at se-
 MILITARY JOB SATISFACTION ITEMS 819
lect cognitive precursors to organizational turnover: what has been missed and why.
Psychological Reports, 78, 1331-1352.
JAMES, L. R., &JAMES,L. A. (1992) Psychological c h a t e and affect: test of a hierarchical dy-
namlc model. In C. J. Cranny, P. C. Smith, & E. F. Stone (Eds.), Job satisfaction: how
people feel about their jobs and how it affects their performance. New York: Lexington
Books. Chap. 5, pp. 89-117.
KIM, S., PRICE,1. L., MUELLER, C. W., &WATSON, T. W. (1996) The determinants of career in-
tent among physicians at a U.S. Air Force hospital. Human Relations, 49, 947-976.
KNOOP,R. (1994) Work values and job satisfaction. The lotirnal of Psychology, 128, 683-690.
KOCHER,K. M., &THOMAS.G. W. (1994) Retaining Army nurses: a longitudinal model. Re-
search in Nrrrsing and Health, 17, 59-65.
LESTER,P. E. (1987) Development and hctor analysis of the Teacher Job Satisfaction Ques-
uonnaire (TJSQ). Educational and Psychological Measrirement, 47, 223-233.
MCDONALD, R. l? (1985) Factor analysis and related methods. Hillsdale, NJ: Erlbaum.
NORUSIS, M. 1. (1994) SPSS Professional Statistics G.1. Chicago, E.SPSS, Inc.
ROHALL,D. E., SBGAL,M. W., &SECAL,D. R. (1999) Examining the importance of organiza-
tional sup orts on family adjustment to Army life in a period of increasing separation.
Journal ojgolitical and Military Sociology, 27, 49-65.
SCHUMM, W. R., GADE,P. A,, &BELL,D. B. (2003) Dimensionalicy of military professional val-
ues items: an explorato factor analysis of data from the Spring 1996 Sample Survey of
Military Personnel. ~sycXolo~ica1 Reports, 92, 83 1-841.
SMITH,P. C. (1992) I n pursuit of happiness: why study general job satisfaction? In C. J.
Cranny, P. C. Smith, & E. F. Stone (Eds.), Job satisfaction: how eople feel about their
jobs and how it affects !heir performance. New York: Lexington B ~ ~ Chap. R ~ .1, pp. 5-19.
SMITH,l? C., KENDALL, L. M., & HUWN,C. L. (1969) The measurement of sati.facfion in work
and retiremenl. Chicago, IL: Rand McNally.
UNITEDSTATES ARMY SEARCH INS~ITUTEFOR THE BEHAVIORAL A N D SOCLAL
SCIENCES. (2001) Sfi-
report: quality of life (well-being) and job satisfaction. (Report No. 2001-07, February)
x u a n d r i n , VA: Author.
WILLLAMSON, D. A. (1996) Job satisjaction in social services. New York: Garland.
WOOD.D. A., &LEBOLD.W. K. (1970) The multivariate nature oE proEessional job satisfaction.
Per.ronnel Psychologv, 23, 173-189.

Accepted March 13, 2003.