locus of control and well-being at work: how generalizable are western...

15
® Academy of Management Journal 2002, Vol. 45. No. 2, 453-466. LOCUS OF CONTROL AND WELL-BEING AT WORK: HOW GENERALIZABLE ARE WESTERN FINDINGS? PAUL E. SPECTOR University of South Florida CARY L. COOPER University of Manchester Institute of Science and Technology JUAN I. SANCHEZ Florida International University MICHAEL O'DRISCOLL University of Waikato KATE SPARKS University of Manchester Institute of Science and Technology PEGGY BERNIN, National Institute for Psychosocial Factors and Health, Sweden ANDRE BUSSING, Technical University of Miinchen PHILIP DEWE, Massey University PETER HART, social research consultant, Australia LUO LU, Kaohsiung Medical University KAREN MILLER, University of Witwatersrand LUCIO RENAULT DE MORAES, Federal University of Minas Gerais GABRIELLE M. OSTROGNAY, social research consultant, Australia MILAN PAGON, University of Ljuhljana and University of Maribar HOREA D. PITARIU, Babes-Bolyai University STEVEN A.Y. POELMANS, IESE, University of Navarra PHANI RADHAKRISHNAN, University of Texas at El Paso VESSELINA RUSSINOVA, Bulgarian Academy of Sciences VLADIMIR SALAMATOV, Ukrainian Academy of Public Administration JESIJS F. SALGADO, University of Santiago SATORU SHIMA, Tokyo Keizai University OI-LING SIU, Lingnan University JEAN BENJAMIN STORA, Hautes Etudes Commerciales Groupe MARE TEICHMANN, Tallinn Technical University TORES THEORELL, National Institute for Psychosocial Factors and Health, Sweden PETER VLERICK, University of Ghent MINA WESTMAN, Tel Aviv University MARIA WIDERSZAL-BAZYL, Central Institute for Labor Protection, Poland PAUL T. P. WONG, Trinity Western University SHANFA YU, Henan Institute of Occupational Medicine Managers from 24 geopohtical entities provided data on work locus of control, job satisfaction, psychological strain, physical strain, and individualism/collectivism. The hypothesis that the salutary effects of perceived control on well-being are universal was supported hecause relations of work locus of control with well-heing at work were similar in almost all the sampled areas. Furthermore, the individualism/collectivism level of each sample did not moderate the magnitude of correlations of work locus of control with measures of well-being. Findings indicate that control beliefs contribute to well-being universally, hut we suggest that how control is manifested can still differ. 453

Upload: vuhanh

Post on 17-May-2018

214 views

Category:

Documents


0 download

TRANSCRIPT

® Academy of Management Journal2002, Vol. 45. No. 2, 453-466.

LOCUS OF CONTROL AND WELL-BEING AT WORK:HOW GENERALIZABLE ARE WESTERN FINDINGS?

PAUL E. SPECTORUniversity of South Florida

CARY L. COOPERUniversity of Manchester Institute of Science and Technology

JUAN I. SANCHEZFlorida International University

MICHAEL O'DRISCOLLUniversity of Waikato

KATE SPARKSUniversity of Manchester Institute of Science and Technology

PEGGY BERNIN, National Institute for Psychosocial Factors and Health, SwedenANDRE BUSSING, Technical University of Miinchen

PHILIP DEWE, Massey UniversityPETER HART, social research consultant, Australia

LUO LU, Kaohsiung Medical UniversityKAREN MILLER, University of Witwatersrand

LUCIO RENAULT DE MORAES, Federal University of Minas GeraisGABRIELLE M. OSTROGNAY, social research consultant, AustraliaMILAN PAGON, University of Ljuhljana and University of Maribar

HOREA D. PITARIU, Babes-Bolyai UniversitySTEVEN A.Y. POELMANS, IESE, University of Navarra

PHANI RADHAKRISHNAN, University of Texas at El PasoVESSELINA RUSSINOVA, Bulgarian Academy of Sciences

VLADIMIR SALAMATOV, Ukrainian Academy of Public AdministrationJESIJS F. SALGADO, University of SantiagoSATORU SHIMA, Tokyo Keizai University

OI-LING SIU, Lingnan UniversityJEAN BENJAMIN STORA, Hautes Etudes Commerciales Groupe

MARE TEICHMANN, Tallinn Technical UniversityTORES THEORELL, National Institute for Psychosocial Factors and Health, Sweden

PETER VLERICK, University of GhentMINA WESTMAN, Tel Aviv University

MARIA WIDERSZAL-BAZYL, Central Institute for Labor Protection, PolandPAUL T. P. WONG, Trinity Western University

SHANFA YU, Henan Institute of Occupational Medicine

Managers from 24 geopohtical entities provided data on work locus of control, jobsatisfaction, psychological strain, physical strain, and individualism/collectivism. Thehypothesis that the salutary effects of perceived control on well-being are universalwas supported hecause relations of work locus of control with well-heing at work weresimilar in almost all the sampled areas. Furthermore, the individualism/collectivismlevel of each sample did not moderate the magnitude of correlations of work locus ofcontrol with measures of well-being. Findings indicate that control beliefs contribute towell-being universally, hut we suggest that how control is manifested can still differ.

453

454 Academy of Management Journal April

Many authors have noted that cross-nationalmanagement research is now needed more thanever hecause it can no longer he assumed thatAmerican concepts and theories transcend cultureand national boundaries (e.g., Boyacigiller & Adler,1991; Peng, Peterson, & Shyi, 1991; Trompenaars &Hampden-Turner, 1998). For instance, the issue ofemployee control, both as a perception and as ageneralized belief (locus of control), has played aprimary role in theories of organizational behaviorand stress (Ganster & Fusilier, 1989) in the U.S.and other Western management literature. In thepresent study, we examined whether control be-liefs operate in a similar way universally by pro-viding results from 24 geopolitical entities^ regard-ing how locus of control in the workplace relates tojob satisfaction and well-being.

THEORY AND HYPOTHESES

Locus of Control as a Universal Component ofWell-Being

Prominent theories have linked perceptions ofcontrol in various forms to employee well-being(broadly conceptualized to include positive atti-tudes and absence of physical and psychologicalsymptoms) as well as to other variables. For exam-ple, in their job characteristics model, Hackmanand Oldham (1976) considered autonomy to be amajor cause of job satisfaction and positive adjust-ment to work. In Karasek's (1979) demands-controlstress model, the hypothesis is that control at workbuffers the impact of job stressors on well-being.Spector's (1986) meta-analysis showed that per-ceived autonomy and participation at work wererelated to job satisfaction and other measures ofwell-being. In their review of the workplace controlliterature, Ganster and Fusilier (1989) concludedthat control was a vital element of well-being. Inaddition, management approaches that empoweremployees by giving them more control have beenadvocated as both effective and humane (Block,1993; Lawler, Mohrman, & Ledford, 1995).

Research has shown the importance of not onlyperceptions of control in the immediate work envi-ronment, but also of a person's more geheral beliefsabout control. Locus of control, perhaps the moststudied control-related variable, reflects a person'sbelief in personal control in life (internality) ratherthan in control by outside forces or individuals(externality). It has been noted that internal control

^ Most of the entities were nation-states. However,Hong Kong and Taiwan were included, as well as thePeople's Republic of China (PRG).

beliefs are an important component of emotionaladjustment and ability to handle stress in generallife (e.g., Kobasa, Maddi, & Kahn, 1982) and at work(Spector, 1982). Locus of control in the workplacein particular (that is, belief that one has control atwork) has likewise been linked to employee well-being (e.g., Spector, 1988; Spector & O'Connell,1994). In sum, research supports the notion thatinternality is associated with positive well-beingboth on and off the job. Our first hypothesis statesthe universality of such a relation:

Hypothesis 1. Work locus of control will becorrelated with measures of well-being (inter-nality associated with positive well-being) uni-versally across nations and territories.

Locus of Control as an ExclusivelyWestern Phenomenon

There is reason to expect cross-national differ-ences in how work locus of control might relate towell-being (Pervin, 1999). Individualism/collectiv-ism has been studied extensively in relation toculture. As Triandis (1995) explained, individual-ism is a tendency for people to view themselves asindependent entities who are motivated primarilyby their own goals and preferences. It has beendescribed as a reflection of the independent self(Markus & Kitayama, 1998) and an expression ofthe need for self-sufficiency (Kagitgibasi, 1994).Collectivism is a tendency for individuals to viewthemselves as parts of one or more social groupswhose motivation is based mainly on group goalsand norms. It has been described as a reflection ofthe interdependent self (Markus & Kitayama, 1998)and an expression of the need for relatedness(Kagitgibasi, 1994). Nations considered to be indi-vidualistic are to be found primarily in the Anglo-European world, whereas nations considered col-lectivistic come from Asia and Latin America, aswell as other regions of the world.

Gudykunst (1998) noted that members of indi-vidualistic cultures are taught to value indepen-dence and achievement through their own actions.They view themselves and others as having directcontrol over various aspects of life. Members ofcollectivist cultures are taught to value harmonyand solidarity with others (Markus & Kitayama,1991). Because they accept subordination of indi-vidual to group interests, they view the group ashaving legitimate control over their actions. Thesedifferences can be seen clearly in a study by Lund-berg and Peterson (1994), who found that Japaneseconsidered work autonomy to be less importantthan did Americans. Smith, Trompenaars, and Du-

2002 Spector et al. 455

gan (1995) also found that, at the country level,individualism was associated with a belief in indi-vidual autonomy.

Weisz, Rothbaum, and Blackburn (1984) com-pared views of control in the individualist UnitedStates and coUectivist Japan. They noted that in theUnited States, the emphasis is on primary control,in which individuals attempt direct control oversituations through independent action. In Japan,there is more emphasis on secondary control,whereby individuals experience feelings of controlindirectly, either by aligning themselves with pow-erful others or by modifying interpretations of asituation, thereby controlling its effects (that is, byregulating emotional reactions). Our United States-developed locus of control scales mainly reflectbeliefs about primary control, neglecting the sec-ondary control that may be more important in othernations.

There have been several studies of general locusof control in the cross-cultural domain, but find-ings have been somewhat inconsistent (Hui, 1982;Smith, Trompenaars, & Dugan, 1995). Studies haveshown that, in general, Confucian Asians (such asthe Chinese and Japanese) are more external intheir locus of control than are the more individu-alistic Americans and other Western nationals(Hamid, 1994; Hui, 1982). Arguments have beenadvanced that the people of the former Soviet blocknations in Eastern Europe should also be more ex-ternal in their locus of control than the people ofWestern nations owing to the prior state-dominatedsystem that limited individual control over workand other life domains (Frese, Kring, Soose, & Zem-pel, 1996; Tobacyk & Tobacyk, 1992).

From a theoretical perspective, for several rea-sons we would predict that the relation betweenlocus of control and well-being will be smaller incoUectivist than in individualist countries. First,coUectivists are socialized to subordinate their ownpersonal control, and thus a coUectivist will de-velop secondary, rather than primary, control strat-egies to deal with his or her environment. Second,in coUectivist countries, people expect to have lim-ited direct and immediate personal control, and sofailure to have it will be less distressing. In theUnited States and other individualist nations, itmight be argued that having an external locus ofcontrol is actually an indicator of poor adjustment(Phares, 1976).

Third, it has been maintained that the idea ofbehavior being driven by stable dispositions is verymuch rooted in individualism, whereas behavior incoUectivist societies is more context-specific anddriven by the environment (Church & Lonner,

1998; Markus & Kitayama, 1998). Thus, we wouldexpect an individual trait such as locus of controlto be predictive of attitudes and behavior moreconsistently in individualist countries. Finally, inindividualist countries, people are encouraged toautonomously regulate their distress (Friedlmeier &Trommsdorff, 1999) rather than to seek consolationfrom others, thereby making personal control amore important element of well-being. In contrast,coUectivist emotional regulation is more depen-dent on others.

In summary, a belief in personal control shouldbe more efficacious in individualist than in coUec-tivist coimtries. This leads to our second hypothesis;

Hypothesis 2. Individualism/collectivism willmoderate the relation between locus of controland well-being in such a way that this relation-ship will be stronger in individualist countries.

The Current Study

In this report, we describe results of a studyconducted in 24 geopolitical entities spread acrossfive continents. The aim was to choose locations sothat a wide range of cultures and individualism/collectivism would be represented. We chose sev-eral Anglo and Western European nations to repre-sent the individualist category and both Asian andEastern European entities to represent collectivism.Since our purpose was to compare employees' lo-cus of control across geopolitical entities, we at-tempted to hold the nature of the jobs constant asmuch as possible and to make our samples as rep-resentative as possible. We chose the manager roleas one that exists across nations and tried to selectparticipants who were representative of this occu-pation in each country. We collected demographicinformation on age, education, gender, marital sta-tus, and organizational tenure to serve as controls.Furthermore, we included a measure of individu-alism in our survey, to provide an up-to-date mea-sure reflecting the values of the managers them-selves. This measure provided a more accurateindication of individualism for each sampled entitythan archival measures that might not reflect pos-sible change over place and time would have pro-vided. Finally, we decided to assess locus of con-trol specifically in fhe workplace. Beliefs aboutcontrol can vary across domains of life (Phares,1976). However, since our interest was in the rela-tion of work-related factors to well-being, it seemedmost appropriate to assess beliefs about workplacecontrol rather than control in general.

456 Academy of Management fournal April

METHODS

Samples and Participants

The results reported here are part of the Collab-orative International Study of Managerial Stress(CISMS), estahlished in 1996 by the two projectdirectors (C. L. Cooper and P. E. Spector). Follow-ing the approach of Peterson and Smith and theircolleagues (1995), the project directors compiled alist of geopolitical entities that would represent abroad range of cultures. For each entity, researcherswere invited to participate. Each of these inviteeswas either personally acquainted with, or was re-ferred by people known to, the project directors.Data were available for 24 samples (see Table 1,below), but data on individualism/collectivismwere not collected in Australia, making it necessaryto leave out this sample for the test of Hypothesis 2.(As was noted, we use the term "geopolitical enti-ties" because we have three Chinese samples, fromthe PRC, Hong Kong, and Taiwan.) In our 24 sam-ples, five of the eight "country clusters" designatedby Ronen and Shenkar (1985) were represented(Anglo, Far Eastern, Cermanic, Latin, European,and Nordic), as well as all four of the areas thoseauthors labeled "independents" (Brazil, India, Is-rael, and Japan). Additional details on this projectcan be obtained from the authors or at the followingWeb site; http://chuma.cas.usf.edu/~spector.

The original plan was to keep participants' jobtype constant (manager) and to collect data usingmethods that would achieve reasonable represen-tativeness. Although we achieved the first objectivein every case, the second was more difficult. In fivesamples, data were collected in only one or twoorganizations (the PRC, Germany, India, Romania,and the United Kingdom), and in Sweden eightorganizations were represented. In the remainingsamples, procedures were used to cover a widerange of organizations. This was accomplished insome instances by sampling members of manage-ment organizations such as the chamber of com-merce (in Canada, for instance), or an institute ofmanagement (in New Zealand, for one). In othercases, we conducted mail surveys using a manage-ment directory as a sampling frame (Hong Kongwas one such case), or by mailing questionnaires toa random sample of managers in businesses andoffices in a city (in the United States). Multiplemethods and locations were used in some samplesto expand representativeness (in Hong Kong,Spain, and the United States, among others).

Participants were 5,185 managers from 24 geopo-litical entities who completed our questionnairevoluntarily and for no compensation. Sample sizesvaried considerably, from 61 for France to 515 for

Japan. The samples varied considerably on the de-mographic variables of gender, marital status, edu-cation level, age, and organizational tenure (a sum-mary of sample characteristics is available from thesenior author). Although we report the simpler sta-tistics here, we also ran hypothesis tests with thesedemographic variables as controls, to rule out thepossibility that our findings could be explained bydemographic differences between samples.

Measures and Procedures

Our questionnaire included the OccupationalStress Indicator-2 (0SI2), which contains measuresof well-being (Cooper & Williams, 1996); the WorkLocus of Control Scale (WLCS; Spector, 1988); theValues Survey Module 1994 (VSM94; Hofstede,1994; not used in Australia); and additional demo-graphic questions.

The WLCS is a 16-item scale that assesses em-ployee beliefs about their control at work in gen-eral. Half the items are written to indicate an exter-nal locus of control (for example, "Getting the jobyou want is mostly a matter of luck") and half toindicate an internal locus (for example, "Promo-tions are given to employees who perform well onthe job"). For all items, the six response choicesrange from "strongly disagree" to "strongly agree."High scores represent externality and low scores,internality. Spector (1988) reported an internalconsistency (coefficient alpha) of .75 to .85 in sixU.S. samples. The scale has been shown to relate asexpected to job satisfaction, control at work, rolestress, and perceptions of supervisor style, and ithas been shown to correlate from .49 to .57 withRotter's (1966) general locus of control scale (Spec-tor, 1988). It has also been found to be related toboth job performance (Blau, 1993) and counterpro-ductive behavior at work (Fox & Spector, 1999).

The 0SI2 is a 90-item short form of the Occupa-tional Stress Indicator (OSI; Cooper, Sloan, & Wil-liams, 1988). Most available reliability and validitydata concern the longer version (e.g., Robertson,Cooper, & Williams, 1990). Job satisfaction was as-sessed with 12 items concerning a respondent'sorganization and his or her work itself. Items hadsix response choices ranging from "very much dis-satisfaction" to "very much satisfaction." A sampleitem is "Communication and the way informationflows around the organization." Psychologicalwell-being was assessed with the 12-item mentalstrain scale that asks about psychological reactionsat work. All items had six response choices, but thechoices varied across items. For example, the item"Are there times at work when you feel so exasper-ated that you sit back and think to yourself that 'life

cdu

egS03

TAI

vari

an

oU0)uecd•ccd>

fact

ion

•aItC/3

c

CO

cn

1T3

CS

Eo

I

00 ID

d dCD ^

G) cnd d

CO O G) COO5 O 05 CT

d

t ^ CO05 O5

COO5

d d d d d dt ^ CD fv CO » ( ^

O5 O3 G3 Ol O5 O3

o <6 a a o o

CO f^ O5 COOl 01 Ol ^

Ol CDOl CO

CO CD in in in txOl Ol Ol Ol Gl CO

d d d d <6 d

CD CO CM rH CM COCT CT O l CT CT tx

d d d d d din ^ CO CO CO inCT CT CT CT Ol CO

d d d d d d

CDc:i

d

CDCO

ddid

C31

d

COO l

d

CD01

d

in

dCO

d

COCD C D C M C O C O O r Hr H C O ^ C O r H i n O OT H C O r H C M C M C M C V l ^

^ t in 1% ^CT CT CT O l CT CT

d d d d d d

CO0 1

o

COO l

o

01O

mOlc-i

in01o

CD

o

0 1CO

o

mO l

o

CD0 1

o

CDOJ

CM

cno

O l

n

in01o

CDO l

n

oOlo

mOlo

cno>o

•J

CO

o

rH(31

O

a iCO

o

CMO l

o

roO l

o oCO

o

CT i nCM CMr H CO

<N CDO l CO

d d

CT COCO CO

d d

o inO! CO

d d

O CMCM in^ in

3

a.

T}̂ IX CD i n CO COCO CD o ^ i n ^CM i n CM CO CM C^

CM tN. O CO CD t xOJ CO Ol CO IX CO

d d d d d d

CD CD CO CT hs txCO IX CO tx in tx

d d d d d dCT ^ tx CD ^ ^CO CO CO CO tx CO

d d d: <D d o

0 Cl O •* r-l tv01 01 01 01 ^ CO

d d d d d d

CD ^ CO CO fvl COC31 CO I-^ CO O CMCO CO • * CD O l i n

I'llS o

X

T3Ci

458 Academy of Management Journal April

is all really just too much effort?'" had choicesranging from "never" to "often." Physical well-being was assessed with the 6-item physical strainscale that asks ahout somatic symptoms, such asshortness of hreath and muscle tremhling. The sixresponse choices ranged from "never" to "very fre-quently." For all three scales, high scores repre-sented high levels of well-heing—that is, high sat-isfaction and low psychological and physicalsymptoms. Robertson et al. (1990) reported coeffi-cient alpha reliabilities for the original OSI of .85,.88, and .78, respectively. Validation evidence forthe most recent version of the OSI was summarizedin Williams and Cooper (1998).

The Values Survey Module assesses five culturalvalues, including individualism/collectivism. Itcontains four items for which respondents indicateimportance, using five response choices rangingfrom "of very little or no importance" to "of utmostimportance." The items ask about having sufficienttime for a personal life, having good working con-ditions, having job security, and having an elementof variety and adventure in one's job. High scoresrepresent an individualist orientation, and lowscores, a coUectivist orientation. We computedscores by differentially weighting the items andadding a constant, resulting in means for countriesthat typically range from about 0 to the low 100s.Hofstede (1994) noted that the VSM94 is only ap-propriate at the aggregate or country level of anal-ysis rather than at the individual level of analysis.

The CISMS project directors put together the En-glish version of the questionnaire. In 8 samples(Australia, Canada, India, New Zealand, South Af-rica, Sweden, the United Kingdom, and the UnitedStates) the questionnaire were administered in En-glish. In 16 samples, the questionnaires were trans-lated into the native language of the area. Each ofthe non-English versions was translated into theappropriate language and then independentlyback-translated to assure language equivalence.Where translation equivalence was not maintained,portions were retranslated and then retested untilthe meanings of the two versions were as close aspossible.

Measurement Transportability

These scales maintained adequate internal con-sistency reliabilities as assessed with the widelyaccepted .70 coefficient alpha standard (Nunnally,1978) in 91 out of 96 cases (coefficient alphas areavailable from the first author). We compared theU.S. alpha (as a standard) with all others, using anF-test (1 - smaller alpha/1 - larger alpha, with n —1 degrees of freedom associated with each alpha)

provided by van de Vijver and Leung (1997: 60). In48 of 114 comparisons, the U.S. alpha was signifi-cantly higher than that for the other sample, and 38of these differences involved translations of thescales. These results should not be surprising, asoften internal consistency declines with translation(e.g., DeFrank, Ivancevich, & Schweiger, 1988;Iwata, Roberts, & Kawakami, 1995).

We also tested for scale transportability by con-ducting "pairwise" multisample variance-covari-ance equality tests on the items for each individualscale using LISREL 8 (Joreskog & Sorbom, 1992).This comparison of the equality of interitem covari-ances (for each of the four scales separately) acrosssamples is the most stringent test of factor equalitythat is frequently used for comparing scales acrosssamples (e.g., Riordan & Vandenberg, 1994;Schaubroeck & Green, 1989). We limited the anal-yses to only those samples that had at least 200participants with complete data on each scale,since this method is designed for large samples.Since the scales were developed in English-speak-ing Anglo nations (the United Kingdom and UnitedStates), we chose our largest sample that fit thatcategory—namely. New Zealand—as a standard,and we then compared it to the samples for Aus-tralia, China, Hong Kong, Japan, Poland, Slovenia,Taiwan, and Ukraine. Our view was that this widerange of samples should provide a clear indicationof transportability. Table 1 shows that for six fitindexes, the three well-being measures met thewidely accepted .90 or higher standard in 87 per-cent of cases across the included countries. Fit wasalmost as good for the WLCS, with a third of thestatistics .90 or higher, and almost three-fourths at.85 or higher, across all the fit indexes combined.The only clearly poor fit was for Taiwan. The tablealso shows the chi-square values for each analysis,all but four of which were significant. However, itis generally recognized that chi-square is overlysensitive to sample size, so we placed more empha-sis on the fit indexes in interpretation.

RESULTS

Test of Hjrpothesis 1

Correlations were computed between locus ofcontrol and the three well-being indicators for eachof our 24 samples as a test of Hypothesis 1. Table 2shows these correlations. In addition, using a z-testfor independent correlations, we took the correla-tion from the U.S. sample as a standard and com-pared it with the corresponding correlation for eachother geopolitical entity to see if there were signif-

.s

oUl

II

I«i Ul

W "

ea:0)

oU

3 I

III

IIIIO B-

III

5o •a

G —o m

<*- ftp fl

I

« « «I r r r r r r r r r r r r r

r r r r I I I I I I I I I I I I I I I

r r r r r r r r r r r r r r r r r r

U

X>a

^ "m 2< CQ CQ

>-, o

o

.2 .2 <

I i g :̂2 1 § 5 I s > I I "S .1 .̂

qVn.

•3

•scdu

a

inqVn.

460 Academy of Management Journal April

icant differences (see Table 2]. Since there werelarge differences in sample sizes, care should hetaken in interpreting comparative significance lev-els. Finally, we ran multiple regression analyses ofeach well-heing variable on work locus of controland the five demographic variables to see if resultswere affected. In no case was significance lost forwork locus of control, showing that demographicdifferences among samples hardly account for ourresults.

The most consistent correlations were betweenlocus of control and job satisfaction. All the sam-ples had a significant, negative correlation on thesevariables, except France, which had a small samplesize but a correlation close in magnitude to mostother samples'. Furthermore, no sample's correla-tion was significantly different from that for theUnited States. Relations between locus of controland psychological well-being had a little more dis-persion. Three samples had nonsignificant correla-tions, and four had correlations significantlysmaller than that for the United States. For physicalstrain, seven correlations were nonsignificant, andten were significantly lower than that for theUnited States.

Test of Hypothesis 2: Individualism/Collectivismas a Moderator of the Locus of Control-Well-Being Relationship

To test Hypothesis 2, stating that individualism/collectivism would serve as a moderator, wechecked to see if the individualism/collectivismvalue found for each sample related to the magni-tude of the correlation between work locus of con-trol and each well-being measure for that sample.We did this by computing the correlations betweenthe mean individualism/collectivism scores andthe corresponding magnitude of correlation be-tween work locus of control and each measure ofwell-being. That is, we separately matched eachsample's mean for individualism/collectivism withthe correlation values in columns 2, 3, and 4 ofTable 2. The two statistics are both summaries ofcountry samples, one a correlation coefficient andthe other a mean. Correlating such summary statis-tics is a practice often seen in meta-analysis; forexample, Carsten and Spector (1997) computedcorrelations between correlation coefficients fromstudies and the unemployment rates at the timesthe studies were conducted.

The correlations were -.07, - .25, and .12, for jobsatisfaction, psychological well-being, and physi-cal well-being, respectively, and each was nonsig-nificant, thus failing to support the moderator hy-pothesis. An inspection of the results showed that

the United Kingdom correlation of work locus ofcontrol with psychological strain was an outlierand far lower than would be expected (r = -.06)from values for the other five Anglo countries,which ranged from -.31 to -.40. When the UnitedKingdom was removed, the resulting correlation of-.40 just missed statistical significance (p < .07),and was in the expected direction, with the stron-ger relations between work locus of control andpsychological strain being associated with higherindividualism.

DISCUSSION

Is the Relationship between Work Locus ofControl and Weil-Being Universal?

As noted, we tested Hypothesis 1 by comparingthe correlations of work locus of control and well-being for our 24 samples. We examined whether ornot these correlations were statistically significantand whether they differed significantly from theU.S. correlations used as the standard. For job sat-isfaction, our results support the first hypothesis.Correlations ranged from -.50 to - .23 , with all butthe one for France (which had a very small sample)being statistically significant. Relations with psy-chological well-being were a little less consistent,but in most cases were supportive of Hypothesis 1.For three samples, correlations were both nonsig-nificant and significantly lower than the U.S. one(Bulgaria, Romania, and the United Kingdom).However, it should be noted that, owing to culturalsimilarities, the United Kingdom would be ex-pected not to differ from the United States.

Relations with physical well-being, however,failed to show consistent support for universality.Seven samples had nonsignificant values (Belgium,the PRC, Estonia, France, Hong Kong, South Africa,and the United Kingdom), and ten had values sig-nificantly lower than the U.S. value. AlthoughFrance was not significantly lower than the UnitedStates, its correlation was only -.12 and, as noted,its sample was very small. Thus, in half the cases,the relatively strong relation seen in the UnitedStates (-.39) was not replicated. Therefore, we canconclude that, unlike job satisfaction, physicalwell-being is not universally associated with be-liefs in low control at work.

Individualism/Collectivism as a Moderator ofWork Locus of Control and Well-Being

Our second hypothesis was that individualism/collectivism would moderate relations betweenwork locus of control and well-being in such a way

2002 Spector et al. 461

that they would be stronger in individualist na-tions. We tested this hypothesis by combining theintracultural analysis (Leung & Bond, 1989), in-volving computing correlations separately for eachgeopolitical entity sample, with an interculturalanalysis involving relating the magnitude of thosecorrelations with each sample's mean individual-ism/collectivism score. We failed to find supportfor this prediction, as individualism/collectivismdid not significantly moderate relations betweenwork locus of control and well-being. Even whenwe removed the United Kingdom from our analysisbecause its results were an outlier among Anglonations, at best we found only marginal signifi-cance for one well-being measure. In light of oursupport for Hypothesis 1 with job satisfaction andpsychological well-being, such a relation wouldseem doubtful, but it is surprising that we didn'tfind it for physical well-being, where relationswere not universal.

There are at least three methodological explana-tions, which are not necessarily mutually exclu-sive, for the failure to support our second hypoth-esis. First, at the intercultural level, our number ofsamples (23 for individualism/collectivism) wasquite small, thus reducing the statistical power todetect significant effects at traditional probabilitylevels of less than .05. Second, the Hofstede (1994)VSM94 four-item individualism/collectivism mea-sure is global and unidimensional. It assesses indi-vidualism/collectivism on a single continuum, butit has been recognized that individualism/collec-tivism is complex and multidimensional (Ho &Chiu, 1994; Triandis, 1995). It may be only certainaspects of individualism/collectivism that moder-ate, rather than the global dimension assessed bythe VSM94. Furthermore, using these same data wehave found that the internal consistency of theVSM94 individualism/collectivism scale is poor inmany samples (Spector et al., 2001). Perhaps amore psychometrically sound scale would haveyielded better results. However, it should be notedthat the Hofstede scale has been used extensively,and much of the cross-national literature is basedon its findings. Furthermore, Hofstede (1984) re-ported on 16 of our geopolitical entities in his orig-inal work on cultural values using an earlierversion of the Values Survey Module. A rank-ordercorrelation comparing our ranks with his was .71,showing strong agreement between his order andours (see Spector et al., 2001).

Third, although we attempted to make our sam-ples as representative as possible, in a few casesdata came from just a few organizations. Thus, it ispossible that results in some samples deviatedsomewhat from what might have been found with

more organizations. However, had this been a seri-ous issue, we would have expected our first hy-pothesis to have been unsupported as well, and welikely would not have found such good agreementwith Hofstede on individualism/collectivism.

Study Limitations

Five of our variables were assessed via self-reports, but the crucial variable of geopolitical en-tity was not. It seems unlikely that monomethodbias or method variance accounted for the consis-tent relations observed between locus of controland other variables. First, the idea that all variablesin a questionnaire are related to a degree owing tothe method itself has found very little support(Spector, 1987; Spector & Brannick, 1995). In fact,an inspection of Table 1 shows that quite a few ofour correlations were nonsignificant, a finding thatprovides an argument against pervasive methodvariance. It is still possible that certain variablesshared biases and that these were widespread cross-nationally. The best recognized biases are "re-sponse sets" and "social desirability." Responsesets have been noted as varying considerably acrosscountries (Triandis, 1994; van de Vijver & Leung,1997), even within collectivists, which would de-crease the likelihood of similar findings across oursamples. Social desirability seems a highly un-likely explanation since the WLCS is unrelated to it(Spector, 1988), and little relation has been foundbetween social desirability and job satisfaction(Spector, 1987; Spector & Brannick, 1995).

Conclusions and Management Implications

We found that the relation of work locus of con-trol to job satisfaction was consistent across all 24samples in our study, despite a wide range of cul-tural differences among them. We found consis-tency in almost all samples for the relation betweenwork locus of control and psychological well-being. However, the U.S. findings of internalitybeing associated with reduced physical well-beingdid not hold across our samples. In half of oursamples, the U.S. correlation was significantlyhigher than the comparison. Thus, we can con-clude that the relation of work locus of control maybe universal in terms of job satisfaction and psy-chological well-being, but it does not generalize asbroadly with physical well-being.

What is intriguing about the findings regardingphysical well-being is that they do not follow aneasily interpretable pattern. The geopolitical enti-ties whose samples either differed from the UnitedStates or failed to be significantly different from

462 Academy of Management Journal April

zero included members of country clusters as di-verse as the Anglo cluster (New Zealand, SouthAfrica, and the United Kingdom), Western Europe(Belgium, France, and Sweden), Eastern Europe(Estonia and Slovenia), and Asia (the PRC, HongKong, Japan, and Taiwan). Only the Asian sampleswere consistent, as there were other Anglo, West-ern Europe, and Eastern Europe samples that didnot differ from the United States. Continued re-search is needed to determine why the physicalwell-being results were so unpredictably variable.The explanation of such variability does not seemto lie with individualism/collectivism, at least inthe global manner assessed by the Hofstede scaleemployed here. Work-related legislation that favorshumane working conditions, good diets, availabil-ity of health care, and a plethora of economic andsocial factors may account for differences in phys-ical well-being across nations. This issue is partic-ularly important since it is physical well-being thatlikely relates most closely with physical health andillness.

Although our results hint that stimulating beliefsof control among managers is a healthy organiza-tional strategy across countries, our data do notpinpoint the specific forms of control that mayhave the best adaptive effects in each country. Thatis, despite similarities in relations, it is likely thatthe ways in which people translate their controlbeliefs into action varies across nations. As Trian-dis (1994) noted, etic constructs have emic aspects,as individual cultures find their own ways of ex-pressing them. Thus, the manifestation of controlcan differ across societies, even though individualsmight score the same. In the United States, a personwho is internal expects to have personal, directcontrol over a job, but in a coUectivist society, aperson who is internal may expect to have indirectcontrol via the cultivation of personal relationshipswith others. Certainly individuals in powerful po-sitions in coUectivist nations such as Japan believein their ability to exert control, but they likely seethat control as coming from their networks of rela-tionships with others. Their counterparts in theUnited States more likely see themselves as havingpersonal power less linked to others. Civen a con-trol scale, each might tend to interpret the itemssomewhat differently, but in a way consistent withhis or her own cultural context.

Some aspects of control might well be universalin their effects. It seems plausible that control be-liefs play an adaptive role in a person's interactionwith his or her environment regardless of culture,although the specific nature of those beliefs is col-ored by the cultural context. It should also be keptin mind that although our theories suggest that

control beliefs and perceptions are a causal agent inproducing well-being, the opposite is also possible.It may well be that individuals who are welladapted to their culture have experienced certainsuccesses that enhance beliefs about control. Thesesuccesses contribute to positive attitudes and well-being, so that control beliefs are the effect ratherthan the cause. Again, this phenomenon might oc-cur universally across cultures, although the pre-cise contexts might vary.

As suggested by the popularity of managementapproaches emphasizing autonomy and empower-ment, beliefs of control are a vital element in themanagement of organizations in the West, and ourresults suggest that they can be just as importantelsewhere. Whereas most prior research on controlbeliefs was Umited to reporting mean country dif-ferences in general locus of control, our study wasdesigned to test theoretical hypotheses concerninguniversality of relations, not just means, specifi-cally in the workplace. We have shown that, for jobsatisfaction and psychological well-being, worklocus of control relations generalize. However, itwould be premature to assume that Anglo-Euro-pean management approaches that promote indi-vidual autonomy and empowerment will work uni-versally to enhance well-being. Beliefs aboutcontrol and, presumably, perceptions of control atwork must be culturally appropriate for the contextin which people live, and exploring such contextdifferences should be the next direction for inter-national control research.

REFERENCES

Blau, G. 1993. Testing the relationship of locus of controlto different performance dimensions. Journai of Oc-cupationai and Organizationai Psyciioiogy, 66:125-138.

Block, P. 1993. Stewardsiiip. San Francisco: Berret-Koehler.

Boyacigiller, N. A. & Adler, N. J. 1991. The parochialdinosaur: Organizational science in a global context.Academy of Management Journai, 16: 262-290.

Carsten, J. M., & Spector, P. E. 1987. Unemployment, jobsatisfaction and employee turnovers: A meta-analytic test of the Muchinsky model. Journai ofAppiied Psyciioiogy, 72: 374-381.

Church, A. T., & Lonner, W. J. 1998. The cross-culturalperspective in the study of personality: Rationaleand current research. Journai of Cross-Cuiturai Psy-choiogy, 29: 32-62.

Gooper, C. L., & Williams, S. 1996. Occupational StressIndicator Version 2.0. Windsor, England: NFER-Nelson.

2002 Spector et al. 463

Cooper, G. L., Sloan, S. J., & Williams, S. 1988. Occupa-tionai Stress Indicator management guide. Wind-sor, England: NFER-Nelson.

DeFrank, R. S., Ivancevich, J. M., & Schweiger, D. M.1988. Job stress and psychological well-being: Simi-larities and differences among American, Japanese,and Indian managers. Beiiaviorai Medicine, 14:160-170.

Fox, S., & Spector, P. E. 1999. A model of work frustra-tion-aggression. Journai of Organizationai Beiiav-ior, 20: 915-931.

Frese, M., Kring, W., Soose, A., & Zempel, J. 1996. Per-sonal initiative at work: Differences between Eastand West Germany. Academy of Management Jour-nai, 39: 37-63.

Friedlmeier, W., & Trommsdorff, G. 1999. Emotion regu-lation in early childhood: A cross-cultural compari-son between German and Japanese toddlers. Journaiof Cross-Cuiturai Psychoiogy, 30: 684-711.

Ganster, D. G., & Fusilier, M. R. 1989. Gontrol in theworkplace. In C. L. Gooper & I. T. Robertson (Eds.),Intemationai review of industrial and organiza-tionai psychoiogy 1989: 235-280. Ghichester, En-gland: Wiley.

Gudykunst, W. B. 1998. Individualistic and collectivisticperspectives on communication: An introduction.Intemationai Journai of Intercuiturai Relations,22: 107-134.

Hackman, J. R., & Oldham, G. R. 1976. Motivationthrough the design of work: Test of a theory. Organ-izationai Behavior and Human Performance, 16:250-279.

Hamid, P. N. 1994). Self-monitoring, locus of control,and social encounters of Ghinese and New Zealandstudents. Journai of Cross-Cuiturai Psychoiogy, 25:353-368.

Ho, D. Y., & Ghiu, G. 1994. Gomponent ideas of individ-ualism, collectivism, and social organization: An ap-plication in the study of Chinese culture. In U. Kim,H. C. Triandis, Q. Kagitgibasi, S. Ghoi, & G. Yoon(Eds.), Individuaiism and coiiectivism: Theory,method, and appiications: 137-156. ThousandOaks, CA: Sage.

Hofstede, G. 1984. The cultural relativity of the quality oflife concept. Academy of Management Beview, 9:389-398.

Hofstede, G. 1994. Vaiues Survey Moduie 1994 manualMaastrict, The Netherlands: Institute for Research onIntercultural Gooperation.

Hui, G. H. 1982. Locus of control: A review of cross-cultural research. Intemationai Journai of Intercui-turai Beiations, 6: 301-323.

Iwata, N., Roberts, G. R., & Kawakami, N. 1995. Japan-U.S. comparison of responses to depression scaleitems among adult workers. Psychiatry Besearch,58: 237-245.

Joreskog, K. G., & Sorbom, D. 1992. LISBEL VIII: Aguide to the program and appiications. Moores-ville, IN: Scientific Software.

Kagitgibasi, Q. 1994. A critical appraisal of individualismand collectivism. In U. Kim, H. G. Triandis, Q. Kagit-gibasi, S. Ghoi, & G. Yoon (Eds.), Individuaiism andcoiiectivism: Theory, method, and appiications:52-65. Thousand Oaks, GA: Sage.

Karasek, R. A., Jr. 1979. Job demands, job decision lati-tude, and mental strain: Implications for job rede-sign. Administrative Science Quarterly, 24: 285-307.

Kobasa, S. G., Maddi, S. R., & Kahn, S. 1982. Hardinessand health: A prospective study. Journai of Person-ality and Sociai Psychoiogy, 42: 168-177.

Lawler, E. E., Ill, Mohrman, S. A., & Ledford, G. E., Jr.1995. Creating high performance organizations:Practices and resuits of employee involvement andtotai quaiity management in Fortune 1000 compa-nies. San Francisco: Jossey-Bass.

Leung, K., & Bond, M. H. 1989. On the empirical identi-fication of dimensions for cross-cultural compari-sons. Journai of Cross-Cuiturai Psychoiogy, 20:133-151.

Lundberg, G. D., & Peterson, M. F. 1994. The meaning ofworking in U.S. and Japanese local governments atthree hierarchical levels. Human Beiations, 47:1459-1487.

Markus, H. R., & Kitayama, S. 1991. Gulture and the self:Implications for cognition, emotion, and motivation.Psychoiogicai Beview, 98: 224-253.

Markus, H. R., & Kitayama, S. 1998. The cultural psychol-ogy of personality. Journai of Cross-Cuiturai Psy-choiogy, 29: 63-87.

Nunnally, J. G. 1978. Psychometric theory (2nd ed.).New York: McGraw-Hill.

Peng, T. K., Peterson, M. F., & Shyi, Y. P. 1991. Quanti-tative methods in cross-national management re-search: Trends and equivalence issues. Journai ofOrganizationai Behavior, 12: 87-107.

Pervin, L. A. 1999. The cross-cultural challenge to per-sonality. In Y. T. Lee, G. R. McCauley, & J. G. Draguns(Eds.), Personaiity and person perception acrosscuitures: 23-41. Mahwah, NJ: Erlbaum.

Peterson, M. F., Smith, P. B., Akande, A., Ayestaran, S.,Bochner, S., Callan, V., Cho, N. G., Jesuino, J. G.,D'Amorim, M., Francois, P. H., Hofmann, K., Koop-man, P. L., Leung, K., Lim, T. K., Mortazavi, S.,Munene, J., Radford, M., Ropo, A., Savage, G., Se-tiadi, B., Sinha, T. N., Sorenson, R., & Viedge, G.1995. Role conflict, ambiguity, and overload: A 21-nation study. Academy of Management Journai,38: 429-452.

Phares, E. J. 1976. Locus of controi in personaiity. Mor-ristown, NJ: General Learning Press.

Riordan, G. M., & Vandenberg, R. J. 1994. A central ques-

464 Academy of Management Journal April

tion in cross-cultural research: Do employees of dif-ferent cultures interpret work-related measures in anequivalent manner? Journai of Management, 20:643-671.

Robertson, I. T., Cooper, G. L., & Williams, J. 1990. Thevalidity of the Occupational Stress Indicator. Woric& Stress, 4: 29-39.

Ronen, S., & Shenkar, O. 1985. Clustering countries onattitudinal dimensions: A review and synthesis.Academy of Management Beview, 10: 435-454.

Rotter, J. B. 1966. Generalized expectancies for internalversus external control of reinforcement. Psychoiog-icai Monographs, 80: Whole no. 609.

Schaubroeck, J., & Green, S. G. 1989. Gonfirmatory factoranalytic procedures for assessing change during or-ganizational entry. Journai of Appiied Psychoiogy,74: 892-900.

Smith, P. B., Trompenaars, F., & Dugan, S. 1995. TheRotter locus of control scale in 43 countries: A test ofcultural relativity. Intemationai Journai of Psychoi-ogy, 30: 377-400.

Spector, P. E. 1982. Behavior in organizations as a func-tion of employee's locus of control. PsychoiogicaiBuiietin, 91: 482-497.

Spector, P. E. 1986. Perceived control by employees: Ameta-analysis of studies concerning autonomy andparticipation at work. Human Beiations, 39: 1005-1016.

Spector, P. E. 1987. Method variance as an artifact inself-reported affect and perceptions at work: Myth orsignificant problem? Journai of Appiied Psychoi-ogy, 72: 438-443.

Spector, P. E. 1988. Development of the Work Locus ofGontrol Scale. Journai of Occupationai Psychoiogy,61: 335-340.

Spector, P. E., & Brannick, M. T. 1995. The nature andeffects of method variance in organizational re-search. In G. L. Gooper & I. T. Robertson (Eds.),Intemationai review of industriai and organiza-tionai psychoiogy, 1995: 249-274. West Sussex, En-gland: Wiley.

Spector, P. E., Gooper, G. L., Sparks, K., Bernin, P., Biis-sing. A., Dewe, P., Lu, L., Miller, K., Renault deMoraes, L., O'DriscoU, M., Pagon, M., Pitariu, H.,Poelmans, S., Radhakrishnan, P., Russinova, V.,Salamatov, V., Salgado, J., Sanchez, J. I., Shima, S.,Siu, O. L., Stora, J. B., Teichmann, M., Theorell, T.,Vlerick, P., Westman, M., Widerszal-Bazyl, M.,Wong, P., & Yu, S. 2001. An international study ofthe psychometric properties of the Hofstede ValuesSurvey Module 1994: A comparison of individualand country/province level results. Appiied Psy-choiogy: An Intemationai Beview, 50: 269-281.

Spector, P. E., & O'Gonnell, B. J. 1994. The contributionof personality traits, negative affectivity, locus ofcontrol and Type A to the subsequent reports of job

stressors and job strains. Journai of Occupationaiand Organizationai Psychoiogy, 67: 1-11.

Tobacyk, J. J., & Tobacyk, Z. S. 1992. Comparisons ofbelief-based personality constructs in Polish andAmerican university students. Journai of Cross-Cuiturai Psychoiogy, 23: 311-325.

Triandis, H. G. 1994. Cuiture and sociai behavior. NewYork: McGraw-Hill.

Triandis, H. G. 1995. Individuaiism & coiiectivism.Boulder: Westview.

Trompenaars, F., & Hampden-Turner, G. 1998. Bidingthe waves of cuiture: Understanding cuiturai di-versity in giobai business (2nd ed.). New York:McGraw-Hill.

van de Vijver, F., & Leung, K. 1997. Methods and dataanaiysis for cross-cuiturai research. ThousandOaks, GA: Sage.

Weisz, J. R., Rothbaum, F. M., & Blackburn, T. G. 1984.Standing out and standing in: The psychology ofcontrol in America and Japan. American Psychoio-gist, 39: 955-969.

Williams, S., & Cooper, G. L. 1998. Measuring occupa-tional stress: Development of the pressure manage-ment indicator. Journai of Occupationai HeaithPsychoiogy, 3: 306-321.

Paul E. Spector ([email protected]) receivedhis Ph.D. in industrial/organizational psychology fromthe University of South Florida. He is a professor and I/Oprogram director at the same university. His researchinterests involve the well-being of employees, includingcross-national issues, job satisfaction, job stress, andcounterproductive work behavior.

Cary L. Cooper is the BUPA Professor of OrganizationalPsychology and Health and the deputy vice chancellor ofthe University of Manchester Institute of Science andTechnology in England. He is founding editor of thefournal of Organizational Beiiavior, a Fellow of the Acad-emy of Management, and a coeditor of the 12-volumeBiackwell Encyclopedia of Management.

Juan L Sanchez received his Ph.D. from the Universityof South Florida. He is an associate professor of indus-trial/organizational psychology at Florida Interna-tional University. His research focuses on the identifica-tion of work requirements and work stressors within andacross cultures.

Michael O'Driscoll received his Ph.D. from Flinder Uni-versity of South Australia. He is an associate professor ofpsychology at the University of Waikato, New Zealand.His current research interests include job stress andburnout, the interface between job and off-the-job rolesand experiences, and work attitudes, especially job in-volvement and organizational commitment.

2002 Spector et al. 465

Kate Sparks received her Ph.D. from the University ofManchester Institute of Science and Technology(UMIST). She is a research associate at the ManchesterSchool of Management, UMIST. Her research interestsinclude the impact of long work hours on employeehealth and performance, cross-cultural study of occupa-tional stress, and workplace health promotion.

Peggy Bernin received her MA from Karolinska Institutein Stockholm, Sweden. She is currently a research assis-tant working on several health-related projects, includingGISMS at IPM—The National Institute for PsychosocialFactors and Health, in Stockholm.

Andre Biissing studied mathematics and psychology andreceived a Ph.D. and Dr. rer. nat. habil. He is currently afull professor and chair of psychology at the TechnicalUniversity of Miinchen. His research focuses on tele-working, work satisfaction, occupational stress and burn-out, occupational psychology and health care, and tran-sitions from knowledge to behavior at work.

Philip Dewe received his Ph.D. from the London Schoolof Economics. He is an Anniversary Reader in Organiza-tional Psychology in the Department of OrganizationalPsychology at Birkbeck College, University of London.His research focuses on work stress and coping.

Peter Hart is a social research consultant in Australia.

Luo Lu received her Ph.D. from Oxford University. She isnow an associate professor of psychology at the GraduateInstitute of Behavioral Sciences, Kaohsiung Medical Uni-versity, Taiwan. Her research focuses on work stress andwell-being.

Karen Miller received her MA in industrial/organiza-tional psychology from University of the Witwatersrandin South Africa. She is currently on the faculty of thepsychology department at the same university. Her re-search interests are in the areas of job stress and organi-zational commitment.

Lucio Renault de Moraes is a professor at the FederalUniversity of Minas Gerais in Brazil.

Gahrielle M. Ostrognay is a social research consultant inAustralia.

Milan Pagon received his Sc.D. from University of Mari-bor and his Ph.D. from University of Arkansas. He is anassociate professor of police administration and manage-ment at the GoUege of Police and Security Studies, Uni-versity of Ljubljana, and an associate professor of organ-izational behavior at the Faculty of OrganizationalSciences, University of Maribor, Slovenia. His researchinterests are mainly in the areas of police stress, policeintegrity and deviance, and managerial stress, as well asapplications of organizational bebavior and managementconcepts in police organizations.

Horia D. Pitariu received his Ph.D. from the University ofBabes-Bolyai, Gluj-Napoca, Romania. He is professor ofI/O psychology at the same university. His research fo-cuses on human resource management (personnel psy-

chology, personnel appraisal, and training) and cognitiveergonomics.

Steven A. Y. Poelmans is an assistant professor at IESE(Instituto de Estudios Superiores de Empresa), Univer-sity of Navarra, Barcelona. He works as a research assis-tant at the organizational behavior department. His re-search focuses on work-family conflict, work stress,corporate stress management, and the link between strat-egy, culture, and external communication.

Phanikiran Radhakrishnan received her Ph.D. from theUniversity of Illinois at Urbana-Ghampaign. She is anassistant professor of psychology at the University ofTexas at El Paso. Her research focuses on ethnic harass-ment and discrimination in organizations and on therelation between culture and the self-concept.

Vesselina Russinova is a researcher with the BulgarianAcademy of Sciences.

Vladimir Salamatov is a research scientist with theUkrainian Academy of Public Administration in Kyiv,Ukraine. His interests are in the area of job stress andcross-national issues.

Jesus F. Salgado received his Ph.D. from the Universityof Santiago de Gompostela, Spain. He is a professor ofwork and organizational psychology in the Faculty ofPsychology, University of Santiago de Compostela. Hisresearch focuses on personnel selection, internationalvalidity generalization, economic psychology, and cross-cultural human resource management.

Satoru Shima received his Ph.D. from Keiogijuku Uni-versity, Japan. He is a professor of mental science atTokyo Keizai University. His research focuses on occu-pational stress and occupational mental health.

Oi-ling Siu received her Ph.D. from the University ofLiverpool. She is an assistant professor of psychology atthe Department of Politics and Sociology, Lingnan Uni-versity, Hong Kong. Her research focuses on occupa-tional stress and well-being, organizational climate, andorganizational commitment.

Jean Benjamin Stora received his Ph.D. in law and eco-nomics from the University of Paris as well as his pro-fessional degree in clinical psychology. Retired from theHautes Etudes Gommerciales (HEG) in January 2000, heis an honorary dean of the Faculty and Research ofGroupe HEC. His research focuses on stress management,the psychology of top managers, and cross-cultural psy-chology. He is also a consultant in psychosomatics at Lapiti^-Salpetrifere Hospital in the field of stress and car-diovascular disease.

Mare Teichmann is a researcher at the Tallinn TechnicalUniversity in Estonia.

Tores Theorell, M.D., Ph.D., is a professor of psychoso-cial medicine at the Karolinska Institute in Stockholm.He is also the director of the National Institute for Psy-chosocial Factors in Stockholm. Theorell has publishedmainly in the field of stress and psychosomatic medi-

466 Academy of Management Journal April

cine, specializing on critical life events and work envi-ronment in relation to cardiovascular illness as well asfunctional gastrointestinal and musculoskeletal disorders.

Peter Vlerick is an assistant professor in economic psy-chology at the Faculty of Psychology and EducationalSciences, University of Ghent. He received his Ph.D.from the same university. His research focuses on occu-pational stress, burnout, and consumer behavior.

Mina Westman received her Ph.D. in organizational be-havior from the Faculty of Management at Tel Aviv Uni-versity. She is a senior lecturer at the Graduate School ofBusiness Administration, Tel Aviv. Her current researchinterests are stress and outcomes, work-family issues,crossover, cross-cultural studies, and the impact of vaca-tions on well-being.

Maria Widerszal-Bazyl received her Ph.D. from the Pol-ish Academy of Science. She is a head of the Laboratory

of Work Psychology at Gentral Institute for Labor Protec-tion, Warsaw, Poland. Her research focuses on organiza-tional stress, job control, social support at work, andindividual differences in stress processes.

Paul T. P. Wong received his Ph.D. in psychology fromthe University of Toronto. At present, he is the director ofand a professor in the Graduate Program in GounselingPsychology at Trinity Western University in Ganada.Among his abiding research interests are cross-culturalpsychology in stress and coping, personal meanings, andsuccessful aging.

Shanfa Yu is a professor at the Henan Institute of Occu-pational Medicine in the People's Republic of Gbina.