The Dispositional Approach To Job Attitudes: A Lifetime Longitudinal Test Author(s): Barry M. Staw, Nancy E. Bell, John A.

Clausen Source: Administrative Science Quarterly, Vol. 31, No. 1 (Mar., 1986), pp. 56-77 Published by: Johnson Graduate School of Management, Cornell University Stable URL: http://www.jstor.org/stable/2392766 Accessed: 21/10/2008 16:08
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=cjohn. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

Johnson Graduate School of Management, Cornell University is collaborating with JSTOR to digitize, preserve and extend access to Administrative Science Quarterly.

http://www.jstor.org

The Dispositional ApproachTo Job Attitudes: A LifetimeLongitudinalTest BarryM. Staw, Nancy E. Bell, and John A. Clausen
of University California, Berkeley

Recent debates between the job enrichment and socialinformation-processing perspectives have led to a trend toward greater situationalism in organizational research. This paper, however, argues for a more dispositional approach in which the role of the person is emphasized. Using a longitudinal sample, measures of affective disposition from as early as adolescence were used to predict job attitudes in later life. Results showed that dispositional measures significantly predicted job attitudes over a time span of nearly fifty years. The implications of these findings are discussed in terms of both theories of job attitudes and organizational development activities that attempt to alter employee job satisfactions Not only have hundredsof studies on job attitudesattempted to outlinethe determinantsof job satisfaction,but theories respondto work have been the center of about how individuals resome of the most active controversiesin organizational given the importance search. Thisemphasis is understandable, with welfare and its correlation of job satisfactionto individual of other outcomes (e.g., absenteeism and turnover) concern to organizations. researchrevolves around At present, the focus of job-attitude the debate between job enrichmentand social-informationprocessing theories. Drawingon Maslow's (1954) theoryof needs, advocates of job enrichmenthave arguedthat individual adequatelyor most worktasks failto challenge individuals fulfilltheirdesires for growthand stimulation,resultingin behavior widespreadworkdissatisfactionand unproductive and (e.g., Herzberg,1966; Hackman Oldham,1980; Lawler, 1982). Thus, by redesigningtask characteristics,it is posited, motiindividuals become both satisfied and intrinsically can et vated on the job (Hackman al., 1975). So far,the job supenrichmentapproachhas received considerableempirical port,althoughthe bulkof the data has come fromsurveys using self-reportmeasures of both task characteristicsand job attitudes, ratherthanfrom studies using objective measures of job characteristics,(e.g., Taber,Beehr, and Walsh, 1985) or well-documentedfield experiments (e.g., Goodman,1979). theory has made two key Social-information-processing approach.First,need theories attacks on the job-enrichment as were characterized ambiguousand unnecessary explanaand tions of workattitudes (Salancik Pfeffer, 1977). Then,as a processing social information replacementfor need fulfillment, was suggested as a mechanismthat determines whether individuals respond positivelyor negativelyto task situations and (Salancik Pfeffer, 1978). Ithas been arguedthat individuals'attitudesare not a functionof deep-seated needs but a productof how people sociallyconstructthe worldaround ambiguous,indithem. Thatis, because tasks are inherently them inways that are dictated by the vidualsmay interpret context and meaningof theirown actions. Thus, any information that a task is interesting(perhapsprovidedby others or throughthe observationof one's own behaviorin its social context) could be as strong an influenceon job attitudes as the of objective propertiesof the task. Validation the socialtheory has come most directlyfrom information-processing
56/AdministrativeScience Quarterly,31 (1986): 56-77

? 1986 by Cornell University. 0001-8392/86/31 01-0056/$1.00

Thisresearchwas supportedinpartby Institute grantAG4178 fromthe National of Agingto the Institute Human of Deinvestivelopment(JohnClausen,principal of gator)andby a University California Staw. Corfacultyresearchgrantto Barry this respondenceregarding papershould be sent to Barry Staw, Schoolof BusiM. ness Administration, of University California,Berkeley,CA94720.

Job Attitudes

laboratory studies inwhich manipulated opinionsof others were shown to influencetask satisfaction(O'Reilly Caldand well, 1979; Whiteand Mitchell,1979; Weiss and Shaw, 1979) and indirectly fromexperimentson overjustification show that the interaction intrinsic extrinsicrewardson task attiof and tudes (e.g., Deci, 1975; Lepper,Greene, and Nisbett, 1973; Staw, Calder,and Hess, 1980). The Emerging Situational Approach The debate between job enrichmentand social information to processing has contributed two recent shifts in theories about how people reactto work environments.The first change has been greateremphasis on subjectiveand cognitive of factors, with wider acceptance that interpretation the work situationis at least as important objective reality.Although as job-enrichment theorists have long used perceptualmeasures in assessing task characteristics(e.g., Hackman Oldham, and 1975; Sims, Szilagyi,and Keller,1976), individual cognitionhas historically been treated more as a source of errorvariance than as an important causal factor.Thus,the conflictbetween the social-information-processing job-enrichment and positions has put new life in the search for why a single job can be perceived in such differentways (O'Reilly, Parlette,and Bloom, 1980). The second majorshift injob-attitude theory has been a move towardgreaterenvironmental determinism.Not only have need-based theories come undersevere criticism,but recent differences that attempts to finda coherent set of individual moderatethe effects of job enrichmenthave not been particularlysuccessful (White,1978). Consequently,fromwhat was essentially an interactional positionemphasizingthe proper "fit"between individuals workcontexts (e.g., Hackman and and Oldham,1976), we have now moved furtherin the direction of emphasizingsituationalinfluence.Currently emphaor sized are ways to manipulate change job attitudesexternally,using eitherobjectiveinfluences derivedfromtask design or subjective influences derivedfrom social cues. The best example of such a synthesis is Griffin's (1983) recent field experimentinwhich bothjob enlargementand social labeling were shown to increase the satisfactionof workers in two factorysettings. Restoring the Dispositional Perspective between job-enrichment and Althoughthe confrontation theories has contributedsubsocial-information-processing to stantially our knowledge of job attitudes,there has been, we of believe, an unfortunate byproduct this debate. Because of and interpretive empirical problemswith need theory (Wahba and Bridwell,1976) as well as the recent emphasis on situationalinfluence,the field has very nearlyeliminated individual-level variablesfromthe study of job attitudes.The field is no longeras interested in what the individual bringsto the worksetting interms of behavioral tendencies, traits,and of personality(now commonlysubsumed underthe rubric can personaldispositions)as in how the organization externally prodthe individual evoke more positivejob attitudesand to behavior.
57/ASQ, March1986

The currenttrendtowardsituationalism organizational in research correspondsto similarleaningsin psychologyduring the 1960s and 1970s. Overthese twenty years, dispositional concepts were criticized manygrounds,the most tellingof on which was Mischel's (1968)argumentthat personality scales or traitmeasures account for littlevariancein behavioracross situations. Morerecently,however, a largenumberof counterarguments have surfaced indefense of personalitydeterminants of behavior.Bem and Allen(1974) noted that the behaviorof some but not all individuals consistent across is situations. Block(1977)arguedthat in-depthassessments of personalityby trainedspecialists are much more predictive thanthe paper-and-pencil measures of traitsthat are commonlyused. McGowanand Gormly(1976)and Aries, Gold,and Weigel (1983) posited that personality traitsare more predictive of multipleinstances of behaviorthan behaviorin a single situation.Buss and Craik (1983) noted that personalityconstructs are best used as predictorswhen they are based on aggregate levels of previousbehavior.Monson, Hesley, and Chernick (1982) showed that personalityis more predictiveof behaviorin ambiguoussituationsthan in settings inwhich role demands are so strong that behavioris externallydetermined regardlessof personaldispositions.And,finally,Funderand Ozer(1983) demonstratedthatthe statisticalmagnitudeof manyof the most famous situationaleffects (e.g., forced compliance,bystanderintervention, obedience) is no and greaterthanthat achieved by the more heavilycriticized dispositionalresearch.Thus, a consensus has startedto emerge that dispositionalconstructs, when properly measuredand relatedto behaviorsthat are conceptuallyrelevant,may be useful explanatory variablesafterall. Current personality theory has also moved from a defensive posture, attemptingto justifydispositionalresearch,to a more constructivestance. Not only has there been a resurgence of empiricalresearchon dispositionaldeterminantsof behavior (Buss and Craik, 1985), but a growingnumberof studies in behaviorgenetics are now providing ratherconvincingevidence for the heritability many individual of characteristics (Buss and Plomin,1984). Studies comparingidenticalvs. fraternal twins as well as data on the similarity twins raised of aparthave, for example, shown very strong indicationsof a genetic basis for both personality mentalabilities(Plomin, and DeFries,and McClearn, 1980; Bouchard,1984). Thoughbehaviorgeneticists are carefulnot to deny the role of environment or the interaction heredityand environment,the main of effect of individual dispositionson attitudesand behavioris becoming more firmlyestablished. Inthe context of organizational behavior,dispositionalresearch is also makingsomething of a comeback.AlthoughMaslow's need theory has been largelydiscreditedand neitherdemovariableshave provedto be strong graphicnorpersonality determinantsof organizational behavior,argumentsare starting to be formulatedfor improvementsinthe use of dispositionalconstructs. Weiss and Adler(1984) have noted, for variablesmay have been such example, that dispositional weak predictorsof organizational behaviorbecause they have been used largelyin an atheoreticalmannerto explainadditionalvariancein situationalstudies. Off-the-shelfmeasures of
58/ASQ, March1986

Job Attitudes

personalityas well as easily collected demographicfactors are or usuallyadded to surveys and experimentsas ancillary moderatingvariables.Seldom is much thought given to the variables theoreticalcorrespondence between dispositional and the organizational behaviorsthey are intendedto predict. Norare there manystudies designed inwhich situational factors are variedso as to help develop the constructvalidity of variablesor show underwhat conditionsdisposidispositional tionalfactorswill best explainwork behavior.Thus,as Weiss and Adler(1984) have argued,a byproduct havingputalmost of all ourtheoreticalenergies intodemonstratingsituational effects is thatwe may now mistakenlyinferthatdispositional attitudesand variableshave littleto contributeto organizational behavior. Inthis paperwe outlineevidence for a dispositionalexplanation of job attitudesand attempt to show why it is useful to to restore considerationof the individual researchon job satisfaction.The argumentthatjob satisfactioncan have a dispositionalsource is not necessarilya new idea, but it is one that has been distinctlyout of favorfor at least twenty years. The dispositionalpositioncan be tracedto some of the earliest writingsin organizational psychology. Munsterberg(1913:
198), for example, noted that ". . . the feeling of monotony

kind depends much less uponthe particular of workthan upon the special dispositionof the individual." Hoppock(1935: 5), in the summarizing results of the firstextensive study of job satisfaction,also noted that a multitudeof dispositionaland as demographicfactors ". . . may be just as important the job itself in determining what we tentativelychoose to calljob satisfaction."And,in the 1940s and 1950s there were many survey researchstudies that correlatedmiscellaneous demovariablesto job satisfaction(forregraphicand personality views, see Herzberget al., 1957; Vroom,1964; Locke, 1976). Thoughfew of these previousattempts to explainjob satisfaction with dispositional variablesyielded strong or theoretically meaningfulresults, we would argue, likeWeiss and Adler (1984), that it is time to improvedispositionalexplanations ratherthanabandonthem. Toward a Dispositional Theory of Job Attitudes As a startingpointfor reexamining dispositionalsources of job satisfaction,it is useful to considerwhether job attitudes have some consistency over time. We would not go as faras Buss (1 and Craik 985) inarguingthat stabilityis the sine qua non for dispositions,but we would note that examiningtemporal dispositional stabilityis a valuablefirststep informulating explanations.Evidenceof temporalstabilityprovidesat least a forces may be clue, if not evidence, that some dispositional operatingon attitudesand behavior.And, more importantly, withoutconsideringtemporalconsistency, dispositional theories must resortto complicatedstage and interactional models to explainjob attitudes. Severalfindingssuggest thatjob attitudesdo have some temporalstability.Schneiderand Dachler(1978),for example, found strong temporalconsistency in satisfactionscores, with the correlationsbetween pre-and post-measures of job attitudes averaging.56 for managersand .58 for nonmanagersin a
59/ASQ, March1986

16-monthlongitudinal study. Pulakosand Schmitt(1983) also found that highschool students' pre-employmentexpectations of satisfactionwere a significantpredictor subsequent job of attitudes. Finally, some indirect supportfor temporalconsistency comes fromthe fact that national survey data have shown ratherstable levels of job satisfactioneven though there have been manyobjectivechanges in the qualityof worklifeover the past decades (Quinn,Staines, and McCollough, 1974; Quinnand Staines, 1979; Weaver, 1980). Inorderto demonstratedispositionalsources of job satisfaction it is also important establish cross-situational to consistency - to show that behaviorcan be predictedacross disparate contexts as well as withinthem (cf. Mischel, 1968). So far, the strongest evidence for the cross-situational consistency of job attitudes comes froma recent study by Staw and Ross (1985). They reanalyzed data fromthe NationalLongitudinal Survey(originally designed to study laboreconomics issues) and found stabilityinjob attitudes over time and situation.Over both three- and five-yeartime intervalsStaw and Ross found significantconsistency injob satisfactionwhen people worked for the same employerand in the same occupation.More theirreanalysesof the NationalLongitudinal importantly, Survey data showed significantconsistency injob satisfaction when individuals changed both the employerforwhom they worked and theiroccupation. The findingsof temporaland cross-situational stabilityinjob attitudes are important preliminary yet a steps in formulating dispositionalapproachto job attitudes.They create a basis for what Buss and Craik (1983) have identifiedas an actuarial approachto dispositionalresearch, providing summaryof a past attitudesthat is useful for predictionsover time and situation.Such consistency datado not, however, constitute a dispositionaltheoryof job attitudes, since they have littleto say about why individuals may show stabilityinjob satisfaction. To formulatea dispositionaltheorywe must begin to understandsome of the mechanisms that underliethese consistency effects. Job Attitudes As a Function of Affective Disposition One clue about dispositionalsources of job satisfactionwas providedlong ago by Fisherand Hanna(1931).They argued that job dissatisfactionmay be caused as much by the ongoing emotionalstate of the individual by any objective as propertiesof the job. Intheirdispositionalexplanation job of attitudes, Fisherand Hanna(1931: vii-viii)proposed an attributionalprocess not too differentfromthe social-informationprocessing ideas now popularin organizational behavior:
The emotionallymaldevelopedindividual almost certainto become is maladjusted sooner or laterto some or several of the variousmajor apects of his everydaylife. Inasmuchas his feelings and emotions are inherentaspects of himself,he carriesthem with him,so to speak, into every situationwhich he enters. Now, since he does not usually know the reason of his dissatisfaction,does not understand the whyforand natureof his maladjustment, is not surprising it that he very frequentlyattaches or attributesit (hisdissatisfaction) his work to or his workingsituation.

Althoughwe would not labeljobdissatisfactionas a pathology, we do agree with Fisherand Hannathat the person's internal
60/ASQ, March 1986

Job Attitudes

state can serve as an important stimulusfor the interpretation of job information. People may bringa positiveor negative dispositionto the work setting, process information about the job in a way that is consistent with this disposition,and then experience job satisfactionor dissatisfactionas a result.Thus, likesocial-information-processing theory,we would posit that jobs are ambiguousstimulisubject to the cognitive manipulation of meaning. However,we would argue that the formulation of task attitudes can come as much fromthe internal state of the individual fromany externalcues. as The hypothesis that personaldispositionsaffect job attitudes could draw at least indirectsupportfromthree separate findin ings. First,there is substantialindividual variation the perception of tasks that have identicalformaljob descriptions (O'Reilly, Parlette,and Bloom, 1980). Thus, there is probably enough ambiguityin most job situationsto allow individuals to interpret work inways that fit theirown dispositions. the beSecond, there is substantialevidence of a relationship tween job and life satisfaction(e.g., Orpen,1978; Near, Rice, and Hunt,1978; Weaver, 1978) as well as a linkbetween job attitudesand mental health(Kornhauser, 1965; Gechmanand Wiener, 1975; Kahn,1981). Althoughthese data are usually interpretedas evidence of the effects of job satisfaction,they can also reflecta more generaldispositionalsyndrome (Schmittand Pulakos,1986). A thirdpiece of data indirectly supportingthe dispositionalperspective is the fact that few well-documentedfield experiments have shown long-term changes injob attitudes. Oldhamand Hackman(1980) have arguedthat most organizational changes are eithertoo weak to have an effect or are resisted by traditional practices inorganizations.What has not been noted, however, is the possibility that attitudechanges may not persist over time because of a to state tendency for satisfactionto return an equilibrium determinedby one's affective disposition(cf. Landy,1978). Inourview, a primecandidatefor explaining dispositional sources of job satisfactionis the affective state of the individual. Historically there has been substantialinterest in the psychologicalstates of mood and depression (e.g., Beck, 1967; Zuckerman, 1977). However,only recentlyhas it been noted that positiveor negative affectivitymay constitute an important on-goingdispositionof individuals, perhapsintegrating manydisparatepersonalitydimensions (Watsonand Clark, 1984). Recent researchon identicaltwins raisedaparthas also revealedthat affectivity,likemanyother individual differences, could have a genetic basis that persists over one's lifetime (Bouchard, 1984, 1985). Thus, people's affective dispositions can be thoughtof as generaltendencies toward positive or negative evaluationof life stimuli- tendencies that should influence-theway people perceive work environmentsduring their lives. We do not, of course, expect that one's affective dispositionwill dominateperceptionsto the extent thatwork in a grossly negative setting will be interpreted a highly as pleasantjob but, rather,that affective dispositionswill constitute a significant(andpreviouslyoverlooked)determinantof work attitudes. The present study examines the influenceof affective disposition on job attitudesover long periodsof time. Thisresearch
61/ASQ, March1986

takes advantageof a very unusualdata set, one that extends who have over much of the lifetimeof a groupof individuals been extensively tested at several points intime. We used these data to test whether the affective dispositionof individuals,fromas farbackas junior senior highschool years, and can predictjob satisfactionin laterlife. METHOD Subjects Datafor this researchwere taken from the Intergenerational Studies (IGS) conducted by the Instituteof HumanDevelopment at the University California, of Berkeley.The IGSdataare an aggregationof three separate longitudinal studies that have for investigatedthe lives of selected individuals over fifty years. Begun inthe late 1920s and early1930s, the three studies each startedwith differentprincipal investigators,samples, and researchgoals, yet there was enough commonalityin measurement to allow IGSstaff subsequently to combine the samples intoa single data pool. The firstof the IGSwas the GuidanceStudy initiatedby Jean WalkerMacfarlane 1928. Althoughthis projectwas originally in plannedas a six-yearstudy of children'sproblembehavior,it was latercontinuedto researchthe interaction psychologiof cal, social, and biologicalfactors in personalitydevelopment. The study's 248 original participants were drawnfrom a survey of every thirdbirthin Berkeleybetween January1, 1928 and June 30, 1929. Since one of the study's goals was to assess the impactof guidanceof parentsby professionalstaff, the were also dividedintoguidanceand control participants groups. Extensivepersonalityassessments were taken for the guidance group;fewer data were gatheredforthe control subjects untilearlyadulthood. The BerkeleyGrowthStudy,the second IGS,was started by Nancy Bayleywith 61 infantsbornduring1928 and 1929. The initial sample was located by requestingpermissionof obstetricians to visit parentswith new babies and was laterextended to 74 subjects by including13 additional babies bornin the subsequent three years. Althoughthe original purposeof the BerkeleyGrowthStudywas to trace the intellectual,motor, and physicaldevelopment of infantsand young children,the projectwas extended over time with the inclusionof a wide rangeof psychologicaland social assessments. GrowthStudy. Begun in The thirdof the IGSwas the Oakland E. 1931 by Harold Jones, MaryC. Jones, and HerbertR. Stolz, the Oakland GrowthStudywas designed as a study of normal adolescence. The projectoriginally included212 subjects recruitedfromthe fifthand sixth grades of five elementary and schools in northeasternOakland subsequently enrolledin a single juniorhighschool. The medianbirthmonth for this Growth sample is May 1921, thereby makingthe Oakland Studysubjects an average of nearlyeight years olderthan subjects in the Guidanceand BerkeleyGrowthStudies. When Growth initial measures were taken, members of the Oakland Studywere 10-12 years in age, butthe most extensive psychologicalassessments of them were made duringjunior and senior highschool years.
62/ASQ, March1986

Job Attitudes

and the Demographically, Oakland BerkeleyGrowthStudies were limitedto Caucasians,whereas the Guidanceand Control minorities.About samples did includesome (English-speaking) GrowthStudysubjects came from 40 percent of the Oakland working-classhomes, whereas about 33 percent of the participants of the Guidanceand BerkeleyGrowthStudies had Thisdifference is a reflectionof working-classbackgrounds. and the demographiccharacteristicsof Oakland Berkeleyat 90 the time of the studies. At adulthood,approximately peras cent of allthree samples could be characterized middle occurredin the class. Overtime, some selective attrition of studies, with working-classmen and lower IQparticipants both sexes havingslightlyhigherdrop-outrates (see Eichorn, 1981, for a more thoroughdescriptionof the samples in the IGS). Q-Sort Data and Methodology is Because attrition a majorproblemwith any longitudinal study, let alone one that spans several decades of people's lives, it was necessary for IGSresearchersto set up a procedurefor poolingdata across studies. Althoughallof the IGS measures development, the particular investigatedpersonality variedacross the projects.An used to assess individuals goal, therefore,was to aggregate availabledata on a important set of common dimensions across the three studies. with Haan(1971),developed a Q-sort Block,in collaboration methodologyas a way of combiningdiverse data fromvarious assessments intoa set of common personalitycharacteristics. from the three studies were Althoughthe materialsavailable heterogenous (consistingof interviews,tests, and case recon ords),there was enough information most cases (interms of breadthand depth)for a set of common Q-sortdimensions to be derived.Using the data archivesfor each subject, trained using a set personalityraterswere asked to assess individuals of descriptivestatements such as "favorsconservativevalues and in a varietyof areas," "initiateshumor,""irritable," "seeks reassurancefromothers." Thejudges placed allof the items in in a forced normaldistribution which lower categories (1-3) of represented qualitiesor traitsleast characteristic the perscores (4-6) represented moderatelycharson, middle-range acteristictraits,and higherscores (7-9) were most characteristic of the person being rated.The procedureresulted in an ipsativemeasure of the salience of variousdimensions of each individual's personalityratherthana normativeratingof individualsagainst age or sex peers. Q-sortswere done separatelyfor each of five time periods. Two adolescent-subjectsorts were conducted, one for ages 12-14 (earlyadolescence) and anotherfor ages 15-18 (late mateadolescence), based on a case assembly of the archival A rialavailablefor each study participant. sort for the firstadult period(aboutage 30 for Guidancesubjects, 36 for Berkeley Growth in Growthsubjects, and 38 for participants the Oakland Study)was based on interviews rangingfromfourto six hours for BerkeleyGrowthsubjects to an average of 12 hoursover several days for members of the other studies-.Sorts for the second adultwave (when Guidancesubjects were about age Growthsubjects about 48) were based on 40 and Oakland tape-recordedinterviewsaveragingfourhours in length. Due
63/ASQ, March1986

BerkeleyGrowthsubjects were not to fundinglimitations, interviewedfor the second adultwave. Sorts for the thirdadult period(aboutage 54 for Guidanceand BerkeleyGrowthsubwere also based Growthparticipants) jects and 62 for Oakland on tape-recordedinterviewslastingabout two hours. At least three judges ratedeach adolescent and at least two of ratedeach adultfor the varioustime periods. Combinations judges were systematicallyvariedso that no judge ratedthe same subject at more than one time period,nordidjudges fromtime periodsother than the have access to information one they were rating.The judges were highlyqualified,consocial workers, sisting of clinicalpsychologists and psychiatric allwith substantialclinicaland/orresearchexperience. Thorin ough training the Q-sortmethod was also provided,and judges' performanceduringthe ratingswas closely monitored. Judges were given feedback if they appearedto be using any manner(e.g., by consisof the Q-items ina nondifferentiating tently placingan item highin allsubjects' profiles),and several tests were performedto provideassurance against the use of placement of items and stereotypic ratings.Nondifferentiating stereotypingdid not appearto be problemsin the Q-sorts,and extra raterswere used if the Q-sortdata were not sufficiently of reliableacross the multiplejudges. The mean reliabilities the Q-sortsat the varioustime periodsrangedfrom .72 to .78. data Additional reliability as well as more extensive explanation of the Q-sortproceduresused in the IGScan be found in Block et (with Haan,1971) and in Eichorn al. (1981). Sample Size Althoughthe Q-sortprocedureallowed the IGSresearchersto studies, sample sizes are still modcombine three longitudinal est forthe present investigation.The firstlimiting factoris that female subjects of the IGS(halfof the three samples) hadvery limitedwork experience and thus hadto be excluded from the present analyses. The second problemis that relatively few IGSparticipants were assessed for allof the five time and periods. Early late adolescent datawere not availablefor in subjects who were originally the controlgroupof the Guidance Study,since there were insufficientmaterialsfor 0-sorts to be performedfor these cases. BerkeleyGrowthsubjects were not assessed in the second adultwave. Finally, although over any extensive efforts were made not to lose participants of the data collectionwaves (whichsometimes involvedinterviewers travelinglong distances to assess those unableto are returnto Berkeley)data for manyindividuals missing for at to least one of the time periods. Fora study participant be includedin every wave of ouranalyses, sufficientdatawould be necessary forfive personalityQ-sorts,rangingover a period of 40 to 50 years, as well as two employmentquestionnaires and an employment interviewadministeredover the last 12 we years of the study. Because of these sample limitations, present analyses using both the largestavailabledata set as well as a smallerset of complete longitudinal data. Measurement of Affective Disposition Q-sortitems were used consistently in the IGS Eighty-three across the five waves of data collection. Fromthese 83 personalitydescriptionswe isolatedallthose statements that
64/ASQ, March1986

Job Attitudes

appearedto capturesome aspect of affect on a priori theoretical grounds. Some of these Q-sortstatements pertainedto intrapersonal aspects of affect, such as "cheerful,""feels lack of personalmeaning"(reverse scored), and "irritable" (reverse scored). Otheritems pertainedto interpersonal aspects of affect, such as "behaves in a givingway," "has hostility towardothers" (reversescored), and "feels victimized"(reverse scored). Factoranalyses showed that 17 items pertaining to both intra- interpersonal and affect formeda single affective factor,in a similarmanneras would be predictedby Watson and Clark 984). (1 The specific affective items and theirfactorloadingsare shown inTable 1, alongwith the sample size on which each of the factoranalyses was based. Also shown in the table are the of squared multiplecorrelations the variableswith the factors, eigenvalues, Kaiser'smeasure of samplingadequacy (MSA), As and Cronbach's alphameasure of scale reliability. illustrated in the table, factorloadingswere ratherconsistent across the variouswaves of data collection,averaging.65 across the five time periods.The averagefactor loadingrangedfroma highof .71 forthe late-adolescence periodto a low of .62 for the Adult 1 wave. Althoughit would have been preferableto have had a largersample on which to base the factoranalyses, Kaiser's measure of samplingadequacy indicatedthatthe datawere and Mueller,1978). adequate for the factoringtechnique (Kim Since reliabilities (Cronbach's alpha)rangedfrom .90 to .93 and the squaredmultiplecorrelations the items with the factor of never fell below .95, we can conclude that the 17-itemaffective measure was a reliableindicator that retainedits factor structureover time.
Table1 Items and Factor Loadings for The Affective Disposition Scale Early adolescence .79 -.39 .78 -.61 -.49 .81 .75 -.81 - .74 -.70 -.67 -.76 -.65 .49 -.74 -.46 .64 Time Period Late Adult 1 adolescence .76 -.53 .80 -.66 -.53 .83 .70 -.77 -.76 -.70 -.80 -.75 -.70 .55 -.76 -.69 .75 .73 -.62 .83 -.65 -.35 .68 .70 -.65 -.72 -.49 -.67 -.69 -.52 .36 -.66 -.55 .68

Item Giving Thin-skinned Sympathetic Punitive Condescending Likeable Warm Hostile Distrustful Botheredby demands Irritable Negative Self-defeating Satisfiedwith self Feels victimized Moody Cheerful Squaredmultiple of correlation with factor variables Eigenvalue Kaiser'sMSA N Cronbach's alpha

Adult 2 .68 -.52 .72 -.57 -.35 .74 .65 -.79 -.67 -.69 -.54 -.67 -.69 .37 -.71 -.48 .64

Adult 3 .78 -.62 .84 -.78 -.63 .78 .85 -.88 -.83 -.62 -.53 -.80 -.58 .26 -.66 -.07 .75

.96 7.78 .88 101 .92 65/ASQ, March1986

.97 8.64 .89 94 .93

.95 6.81 .87 130 .91

.95 6.72 .87 104 .90

.97 8.21 .92 128 .92

As expected, the temporalstabilityof affective disposition depended on the time periodseparatingdata collectionwaves. Stability was highest (averaging.48) for the temporally adjacent periods (e.g., Adult1 and 2 waves) and somewhat lower for data collections separated by two time periods(averaging .31) and for measurements dividedby three time periods (averaging.29). These temporalrelationships in linewith are previouslongitudinal researchon individual affect reviewed by Watson and Clark (1984), as well as with relatedpersonality scales (e.g., the nurturant/hostile dimension)developed in previousresearchusing the IGSsamples (Haan,1981). Measurement of Job Attitudes IGSmeasures of job attitudeswere available froma questionnairethat was administeredto study participants duringthe Adult2 wave of data collectionand fromboth a questionnaire and structuredinterviewthatwere partof the Adult3 wave of assessments. Because the questionnairemeasures of job attitudeswere self-reportsthey were independentof the Q-sortassessment of personality. The interviewmeasures were also largelyindependentof the Q-sortdata, since the staff that conductedthe interviewsdid not have access to any of the Q-sortassessments, nordid ratersforthe Q-sorts have access to any specific interviewresponses. Thus, one strength of this study is an independence of measurement, avoiding problemsof common-methodbias that so frequentlyplague researchon job attitudes (cf. Robertsand Glick,1981). The IGSmeasures of job attitudefor the Adult2 time period consisted of survey questions on bothjob and careersatisfaction. One measure was a single item assessing careersatisfaction with five response alternatives,rangingfrom "it is a career Idislikeand wish Icould leave for another"to "ittruly represents what Ihave wanted to do." A second indicator of job attitudeswas a 14-itemfacet satisfactionmeasure inwhich subjects were asked to ratevariousaspects of theirwork on a five-pointscale, rangingfrom "likeit very much"to "dislikeit very much."Thejob facets includedin the measure were workinghours,amountof tension and stress, income, degree to which the work involves interests, generalconvenience for the family,amountof leisuretime, opportunity advancefor ment, natureof supervision,meeting and being with people, use of skillsand abilities,supervisionof others, respect that others give to the job, freedom to develop ideas, and security of the job. These items were aggregated intoa facet satisfaction measure. Because facet satisfactionand careersatisfaction were strongly intercorrelated (r = .51, N = 75, p<.001) index. they were also combined intoan overalljob-attitude Forthe Adult3 time periodthere were five measures of job attitudes availablein the IGSdata.The Michigan facet-specific measure of job satisfaction(Quinn Staines, 1979) was and administered,in which subjects respondedto 18 positive statements abouttheirjobs, such as "the job securityis good," "the chances for promotionare good," and "Ihave an to opportunity develop my own special abilities."Responses to these items were aggregated intoa facet job-satisfaction measure. Inadditionto this facet satisfactionmeasure the IGS data also includedseveral single-itemscales relatingto job attitudes.The Michigan facet-freejob-satisfaction item was
66/ASQ, March1986

Job Attitudes

included,which simplyasked, "allinall, how satisfied would you say you are with yourcurrentjob?"Therewas also a item with slightlydifferentwordingfrom career-satisfaction that administeredduringthe Adult2 wave, and a related whether one would take the job questionnaireitem inquired would again,with three alternatives,rangingfrom "definitely take the same job"to "Idefinitelywould not take the same the job." Finally, structuredinterviewsconducted at the Adult 3 time periodincludeda question about how one felt about going to work,with alternativescoded from "strongpositive responses" to "negativefeelings." When these five measures of job attitudeswere subjected to a factoranalysis,they revealeda single factorwith loadingsrangingfrom .50 to .78. Factorscores were used as an overallmeasure of job attitudes for the Adult3 time period. RESULTS The majordata analyses to be reportedby this study are correlationsbetween affective dispositionandjob attitudes.The firstanalyses displayedare those based on the maximumsamples availablefor each pointintime. The samples differfor few each of the waves of data collection,since comparatively of the subjects were assessed at allfive time periodson both affect and job attitudes. Table2 shows the correlatesof Adult2 job attitudes.The smallest sample availableforthe correlationsconsisted of 51 men who were assessed on affect duringlate adolescence and also forAdult2 job attitudes.The largestsample was 76 men who were assessed on both affect and job attitudes during the Adult2 time period.As shown in the table, the results as revealedseveral significantrelationships well as a conwere sistent trend.As mightbe expected, the correlations progressivelystrongeras the time between measurements resultedfromthe condecreased. The strongest relationship currentmeasurement of affect and job attitudes, both assessed at the Adult2 time period.The weakest relationship, of significant,was the prediction Adult2 though still marginally job attitudesfromaffect duringearlyadolescence.
Table 2 Relationships between Affective Disposition and Adult Job Attitudes Job Attitudes (Adult 2) Affective Disposition Early adolescence Late adolescence Adult 1 Adult 2 p<. 10, p< 05, *p<.01; Facet job satisfaction .14 Career satisfaction .19 Overall job attitude (average score) .20

(N = 59)
.16

(N = 58)
.270

(N = 59)
.260

(N = 52)
.310 (N= 70) .35000 (N= 76) one-tailed tests.

(N = 51)
.230 (N= 69) *35000 (N= 75)

(N = 52)
.30000

(N= 70) .40000 (N = 76)

67/AS0, March1986

Table3 shows the relationship between affective disposition and job attitudesfromthe Adult3 time period.The sample sizes for the variousanalyses rangedfrom 51 for the adolescent years to 101 for the Adult3 time period.As shown inthe table, the affective dispositionscores from earlyand late adolescence were generallyrelatedto the dependent measures. The facet job satisfaction,generaljob satisfaction,and career satisfactionmeasures were allsignificantly predictedby affect fromthe adolescent years. The interviewitem measuring workfeelings and the questionnaireitem on whether one would take the job againdid not show significantrelationships with adolescent affect. However,the overalljob attitudemeasure (usingfactorscores) was significantly predictedby affect from both earlyand late adolescence. Affective dispositionsfromAdult1 were stronglyrelatedto job attitudes at the Adult3 time period.Allof the dependent measures at Adult3 were significantly predictedby Adult1 affect, and the overalljob attitudemeasure was stronglyrelatedto affect fromthe Adult1 period.
Table3 Relationships between Affective Disposition and Adult 3 Job Attitudes Job Attitudes (Adult3) Affective Disposition Early
adolescence

Career satisfaction .20

Facet-free satisfaction .20

Facet satisfaction .240

Take job again .12

Overalljob Feelings attitude about work (factorscore) .11

370
(N = 46)

Late
adolescence

*35
(N = 45)

.21*35g
(N = 45)

.09
(N = 45)

.15
(N = 52)

39g
(N = 40)

Adult1 Adult2 Adult3 *p<.10, *p<.05, *p<.01;

*45 (N = 72) .19 (N = 68) .25g (N = 90) one-tailedtests.

*370 (N = 72) .00 (N = 68) .200 (N = 90)

260 (N = 72) .04 (N = 68) .230 (N = 90)

330 (N = 72) .13 (N = 68) .19 (N = 90)

.200 (N = 81) .08 (N = 75) .10 (N = 101)

48 (N = 67) .12 (N = 63) .230 (N = 81)

Of allthe relationships shown inTable3, the correlationsbetween Adult2 affect and Adult3 job attitudeswere the weakest. Onlyone dependent measure was even marginally relatedto Adult2 affect, and the overalljob attitudemeasure did not show a significantcorrelation. concurrentrelaThe tionships between Adult3 affect and Adult3 job attitudes were somewhat stronger.These correlations were modest in magnitudebut statisticallysignificantdue to the largersample sizes associated with the data. Panel Data One of the interestingfeatures of the results inTable3 is the appearancethat earlymeasures of affect are strongerpredictors of Adult3 job attitudesthan lateraffective measures. However, one cannot be sure of such a trendfromthe array shown in the table. Because differentsubjects were used to test each relationship (withmissing datafor manyof the waves), any comparisonof correlationsinvolvesdifferentindi68/ASQ, March 1986

Job Attitudes

vidualssampled. Therefore,the correlationsreportedinTables 2 and 3 should be interpreted separate tests of the hypothas esis thatjob attitudes can be predictedfromaffective disposition ratherthan an answer to the question of whether affect measured from one life stage is a strongerpredictor thanthat fromanotherperiod. Table4 presents data froma constant sample across the five time periods, using the overalljob attitudemeasures as the primary dependent variable.Results were consistent with the data presented inTables 2 and 3. Adult2 job attitudeswere shown to be most stronglypredictedby Adult2 affect, with adolescent affect the weakest predictor Adult2 attitudes. In of contrast,Adult3 job attitudeswere more stronglypredictedby adolescent and Adult1 affect than by affect measured during the Adult2 and 3 waves. Thus,the datawere inconclusivein establishingwhether earlyor late affective measures are the strongest predictorsof subsequent job attitudes; overall,they appearapproximately equal in strength.
Table4 Relationships between Affective Disposition and Job Attitudes During Adult 2 and Adult 3 Time Periods Using a Constant Sample Job Attitudes Adult 2 Adult 3 overall job overall job attitude attitude
Affective Disposition (N = 52) (N = 31)

Early adolescence Lateadolescence Adult1 Adult2

.270 .260 .310 .380

.34a .360 .480 .01

*p<. 10, *p<.05, 00p<.01; one-tailedtests.

DISCUSSION The data fromTables 2 and 3 show distinctevidence that of affective dispositionis a significantpredictor job satisfacare tion. The correlations not of such magnitudeas to deterof mine more thana minority the variance,but they are at least as strong as those routinely reportedfromcross-sectional studies using a single questionnaire.This is by itself surprising, given that the data reportedhere span almost fiftyyears in time, with nearlyall of the data collections at least ten years apart. The second surprising findingfromthe relationshipsreported in Tables 2 and 3 is the lackof a consistent patternof decay. Onlythe antecedents of Adult2 attitudes showed the normally expected patterninwhich earlymeasures of affect would be weaker than subsequent indicatorsin predicting satisfacjob for tion. We do not have any firmexplanation these results except to note that they are not unusualfor research using the Studies. Clausen(1968),for example, found Intergenerational that adolescent personality variableswere strong predictorsof some forms of adultbehavior(e.g., smoking),with at least as
69/ASO, March 1986

much explanatory power as personality traitsmeasured from adultyears. Ininterpreting such a patternit is possible to argue that adolescent dispositionsset in motiona set of behaviorsor choices that laterproduceimportant consequences for the individual.Itis also possible that adolescent assessments are somehow more validindicatorsof individual dispositionthan data taken from latertime periods, perhapsbecause adolescent behaviorsare less disguised or socializedintoacceptable forms thanthose of adults. Finally, because the adolescent and Adult1 personality assessments were based on a greater amount of personality data (inthe case files) than the Adult2 and 3 measurements, they may be more representativeof the affective dispositionof individuals. Sources of Affective Disposition Ourresults showed that it is possible to constructinternally consistent measures of affective disposition,even when these measures are constructed post hoc from priorpsychological assessments. LikeWatson and Clark (1984),we found a range of personaland interpersonal dimensions forminga coherent set of individual characteristicsthat retainedits factorstructure over time and that possessed some temporalstability(being statisticallysignificantfor up to 40-year intervals). Evidenceof stabilityin affect does not mean that this disposition is unchangeableor not subject to contextualinfluence. Itis unclearfromourdata how the affect of individuals originated, from either genetic or social sources, and how it may be influenced by externalfactors over one's lifetime. Itis also unclear whether the constructof affect remainsa constant entity over time. When assessing personalityin a longitudinal study one cannot, as noted by Haan(1981), assume that scores for a personalityvariableare necessarily equivalentfor childhood and adultyears. Since the Q-sortswere based on disparate sets of interviewdata, it is possible that what was interpreted by judges as affect fromthe adultyears is not a simple extension of affect fromadolescence. Therefore,it may be useful to the interpret present results as showing relationshipsbetween job attitudesand separate (thoughrelated)affective constructs. Of course, even with such a restrictedinference, it is still possible to conclude fromthe IGSdatathat affective measures, fromas farbackas adolescence, are useful predictorsof subsequent work attitudes. While it is possible to speculate about alternativeinterpretations of these results, it is important realizethat few organito zationalstudies have presented longitudinal data beyond several years in duration(Bray,Campbell,and Grant,1974 is a rareexception). It is also important note that the measureto ment of affect for each life stage was, inthis study, assessed by a differentraterand that satisfactionwas measured on entirelydifferentinstrumentsfrom any of the affective measures. Too often organizational researchstems from measurements taken from single self-reportinstruments,thereby inflating the magnitudeof results, as in the correlation job and life of satisfactionor the common practiceof relatingperceivedtask characteristics job attitudes (cf. Robertsand Glick,1981). to Thus, the present findingscan be viewed as an extremely conservativetest of the dispositionalapproachto job attitudes.
70/ASO, March1986

Job Attitudes

Cognitive Reevaluation of Jobs Intesting the hypothesis that affective dispositionpredictsjob attitudes, it has been assumed that cognitivedistortionmay underliethe consistency between affect and job attitudes. It should be stressed, however, that this is an hypothesizedexplanation the data, since we have no directinformation of on perceptualprocesses. A plausiblealternative, example, is for that individuals with positive dispositions have sought out job environments(orhave been selected intojobs) that are objectively more pleasantthanthe employment situationof those with more negative dispositions.Thisalternativecan be eliminated by controlling objective differences injob conditions. for Althoughthe only measure of objectivejob qualityavailablefor both the Adult2 and Adult3 waves of IGSdata is the socioeconomic status (SES)of the individuals' SES should be at job, least a roughproxyfor the overalllevel of pay, responsibility, and the approval others. Therefore,we computed partial of correlationsbetween affect at the variouspoints in individuals' lives with subsequent job attitudes, holdingSES constant. The results of these analyses showed no material changes from the data reportedinTables 2 and 3. Forthe Adult3 periodonly, some additional controlmeasures were available.Recently, IGSstaff completed a coding of the substantive complexityof participants' jobs in terms of dealing with data, people, and things. The substantive complexity of scales were taken fromthe Dictionary Occupational Titles and probably providesomewhat better measures of objective job qualitythan SES. Therefore,we againcomputed partial correlationsbetween affect at the varioustime periodswith job attitudesduringthe Adult3 period,holdingthe three aspects of substantivecomplexityconstant. The results once again showed no significantdeparturefromthose of Tables2 reand 3, suggesting that affect may influenceindividuals' sponses to jobs, regardlessof at least some dimensions of task content. Itwas not possible with the IGSdata to sort out allthe possible the processes underlying relationshipsbetween affect and job satisfaction,since this would have taken both a largersample and measures thatwere unavailable this research. Norwas for it possible with the IGSdata to answer questions about whether affective dispositionis a functionof heredityor environment.Affect could resultfrom inheritedbiologicalfactors, and these inheriteddispositionscould inturnbe associated with other mentalor emotionaltraitsunderlying laterachievement inoccupationalsettings. Alternatively, affect could be a productof earlysocialization experiences that conditionor are correlatedwith futurejob opportunities the likelihood or of success. The present data cannot (andwere not intendedto) between job conditions untangleallthe possible relationships and and attitudes northe web of hereditary environmental influences. We have instead sought to show that dispositional however constituted or developed propertiesof individuals, overtime, can be significantpredictorsof the satisfactionof we with theirwork. Fortunately, can have some individuals confidence inthis limitedinference.The IGSvariableswere much more independentlymeasured and showed relationships over a much longertime span than is usuallythe case
71/ASO, March1986

in organizational research.Ourconfidence in a dispositional explanationof job attitudes is also strengthened by the fact that adultjob satisfactionwas significantly predictedby affective dispositions froman age before formalemployment had begun (i.e., earlyand late adolescence). These relationships are difficult explainby any presumed effects of job experito ence on individuals' affective disposition. Theoretical Implications Inevaluatingthe theoreticalimplications this study, one of should keep in mindthat there are several potentiallevels of dispositionalexplanation. The most fundamentalbut also most primitive knowledge that attitudes have some source of conis sistency over time and context, since this simplyprovidesa clue that dispositionalcharacteristicsare a source of influence. At a second level of explanation reportedrelationshipsbeare tween personality variablesandjob attitudes that invokeuntested or unspecifiedtheoreticallinkages.Here,the causal mechanisms underlying such relationshipsmust be inferred froma wide rangeof theoreticalpossibilities(e.g., those with high-achievementneeds being less satisfied because they might possibly have higheraspirations). somewhat higheror A thirdlevel of dispositionalexplanation would be one where there is theoreticalcorrespondence between an individual characteristic an attitudinal behavioral and or construct.When, for example, power needs correlatewith promotionor, as in this study, affective dispositionrelates to job satisfaction,the theoreticallinkagebetween variablesis more specified or obvious thanwith the almost arbitrary of personality use constructs in much of the organizational literature Weiss and (cf. Adler,1984). Studies using time lags, such as when early needs for power or affective dispositionrelateto subsequent and correspondentdependent variables,are also helpfulin a building convincingdispositionalinference, since they show both causal directionand potency. A fourthlevel of explanation, and one not providedby this study, involves knowledge about the exact derivation dispositionsand aboutthe preof cise mechanisms by which dispositions can affect attitudes or behavior.Inthis researchwe hypothesizedthat affective dispositions may constitute internal cues that influencethe way people perceive or reactto job environments.However, we had no data on the cognitiveprocesses of the IGSsample, nor couldwe estimate the inherited social derivation affecvs. of tive dispositions.These difficultand detailedquestions must therefore await furtherresearchthat logicallyfollows from earlierlevels of dispositionalunderstanding. Althoughwe do not now fullyunderstanddispositionalsources of job attitudes,the generalfindingthatjob attitudes can be predictedby affective dispositionstill has some important implications.Interms of job attitudetheory, it means that we have probably erredtoo faron the side of situational determinism, throwingaway valuabledispositionalpredictorsin our with Maslow's need theory. Italso general disillusionment means that, in orderto understandthe originand change of job and attitudes, it may be useful to understandthe background understudy. As Weiss and development of the individuals Adler(1984) so aptlynoted, we need studies focused on dispositionalvariables,perhapsusing situational variablesas
72/ASO, March1986

Job Attitudes

moderatorsof dispositionalinfluence,ratherthan additional as characteristics simple moderators studies using individual effects. Forexample, if we were to explainjob of situational attitudes in dispositionalterms, we would want to identify those contextualfeatures (e.g., workingconditions,supervision, etc.) that eitherallow people with positive affect to enjoy with negative affect from theirwork or preventindividuals unnecessarilyrejectingtheirjobs. This is a very different researchdirectionthanwe have now and one that deserves furtherconsideration. The argumentthat job satisfactionhas dispositionalcauses also suggests some necessary revisionof our models of job attitudes. One possibilityis that objectivefeatures of the job, aboutthe work, and affective dispositionare social information three separate or independentdeterminantsof job satisfaction. Anotherpossibilityis that these three sources of job attitudes are highlyinteractive.Futureresearchmay find,for example, that the perceptionof objectivejob characteristicsis cues from both internal (disposiconditionedby attitudinal we tional)and external(social)sources. Alternatively, may find and that internal externalcues are moderatedby the realitiesof work life, such that neithera positive dispositionnorshared enthusiasm for the job can influencework satisfactionwithout the presence of objectivelypositive task conditions(so as to validateperceived cues fromoneself and others). Additional research is needed to sort out whether such interactional models are better explanationsof job satisfactionthan a simple main-effects model. of Sortingout the interaction dispositionalandjob characteristics is likelyto be a complex researchtask, as alreadyillustratedby some recent studies in occupationalsociology. For and example, researchby Kalleberg Loscocco (1983)and Mor(1985) suggests that not only timer, Finch,and Maruyama does job satisfaction increase with aging, but that the facets of the jobthat most directlydetermine satisfactionalso change overtime. Thus,we mighteventuallyneed to formulatecareer and life-stage models of dispositionaland task influences in or lieu of simple personality situationalpredictions.Itis also to difficult escape fromthe fact that personalitynot only affects the choice of job situation,but that the job situationcan thus creating also have some lastingeffects on the individual, to causal loops that are difficult understandand reciprocal study (e.g., Kohnand Schooler, 1978). These and related questions should be idealfor a mergingof the psychological behavior. and sociologicaltraditionsof organizational Practical Implications of Interms of practical issues, the most obvious implication our findingsconcerns the usefulness of organizational development activities.The stabilityof job attitudes identifiedby Staw and Ross (1985) and the present findingthat job attitudesare influencedby affective dispositionboth implythat it may be to extremely difficult improvejob attitudesvia externalinterventions.Job redesign efforts may, for example, be prone to failure,since they must contend with strong forces for resistantto stability-forces that either make individuals to to change or promptindividuals return their baseline states over time.
73/ASQ, March1986

our Interpreting results in a conservativeway, one mightinfer that job redesign is seldom worth the effort. The objective features of tasks are seldom concrete enough to dictate a positive or negative response, and as a consequence, indithe vidualsmay generallyinterpret workenvironmentinways that are consistent with theirown psychologicalstates. A more and liberalinterpretation, one that is just as defensible, is that interventionsneed to be strongerand more organizational changes, be they job redesign, improvepervasive.Situational or ments in physicalsurroundings, the governance of the may need to be strong andwidespread enough to organization, overwhelm dispositionalforces for stability. development A compromise positionon the organizational 1981; perspective (Terborg, issue would be the interactionist mightconclude that intervenSchneider, 1983). Interactionists differences tions such as job redesign need to take individual more fullyintoconsideration.Throughbetter or more extendispositions,job assignments sive measurement of individual characteristics. could be more preciselytailoredto individual on it Alternatively, mightbe possible to divideindividuals the for basis of theirlikelihood change. As Bem and Allen(1974) have noted, some people may simplybe more consistent in changes such as theirbehaviorthan others. Thus, situational job redesign may be best suited, ratherthan forthe chronically happyor unhappy,forthose whose affective dispositionvaries over time and across situations. Ourdata on the influenceof affective dispositionmay also for have direct implications personnel selection practices. Ifjob disposiattitudes are ratherstable and governed by individual morale tions, it may be easier to improveorganizational dethroughemployee selection thanthroughorganizational velopment activities. Intheirsimplest form,such personnel implicationsmay not necessitate changes in practice,since alreadybehindmany personalcharacteristicsare probably (e.g., Ross selection and promotiondecisions in organizations and Ferris,1981). The common argumentagainst using perand sonal factors such as likability positiveaffect for evaluation purposes is that they would not be performancerelated. This argumentmay not holdin service occupations, however, can since one of the majorproductsof a service relationship be the conveyance of positiveaffect (e.g., Schneider, Parkington, and Buxton, 1980; Smith, Organ,and Near, 1983). Ifwe go beyond the simplest personnel issues, our disposiand tionalfindingshave additional perhapsmore important When we consider, for example, the research implications. findingsthat depressives are more realisticintheirjudgments of riskand causationthan others (e.g., Alloyand Abramson, 1979), we are led to the conclusionthat positive affect could and hinderperformancein some managerial staff roles. As a role result,we may need to examine each organizational for thinkinginvolvedin the amountof enthusiasm versus critical affective dispositions. orderbest to fit jobs with individuals' often Likewise,if we note that workgroups (andorganizations) to requirea varietyof individuals serve multipleroles, ranging from being a source of social supportto being a devil's to advocate, it may be important search for the propermix of These are the kindsof quespersonalitiesin the organization.
March1986 74/ASQ0,

Job Attitudes

tions that are raisedas one begins to probethe role of individual dispositionin organizational behavior.As we pursue the dispositionalperspective further,new theoryas well as revisionsto managerial practiceare likelyto be suggested.
REFERENCES Alloy, Lauren B., and Lynn Y. Abramson 1979 "Judgment of contingency in depressed and nondepressed students: Sadder but wiser?" Journal of Experimental Psychology: General, 108: 441-485. Aries, Elizabeth J., Conrad Gold, and Russell H. Weigel 1983 "Disposition and situational influences on dominance behavior in small groups." Journal of Personality and Social Psychology, 44: 779-786. Beck, Aaron T. 1967 Depression: Clinical, Experimental, and Theoretical Aspects. New York: Hoeber. Berm, Daryl J., and Andrea Allen 1974 "On predicting some of the people some of the time: The search for cross-situational consistencies in behavior." Psychological Review, 81: 506-520. Block, J. 1977 "Advancing the science of personality: Paradigmatic shift or improving the quality of research?" In David Magnusson and Norman Endler (eds.), Personality at the Crossroads: Current Issues in Interactional Psychology: 37-63. Hillsdale, NJ: Erlbaum. Block, Jack, in collaboration with Norma Haan 1971 Lives through Time. Berkeley, CA: Bancroft Books. Bouchard, Thomas J. 1984 "Twins reared together and apart: What they tell us about human diversity." In SidneyW. Fox (ed.), Individualityand Determ'inism, New York: 147179. Plenum. 1985 Personal communication with the authors. Bray, Douglas W., Richard J. Campbell, and Donald L. Grant 1974 Formative Years in Business. New York: Wiley. Buss, Arnold H., and Robert Plomin 1984 Temperament: EarlyDeveloping PersonalityTraits. Hillsdale, NJ: Erlbaum. Buss, David M., and Kenneth H. Craik 1983 "The act-frequency approach to personality." Psychological Review, 90: 105-126. 1985 "Why not measure that trait? Alternative criteria for identifying important dispositions." Journal of Personality and Social Psychology, 48: 934-946. Clausen, John A. 1968 "Adolescent antecedents of cigarette smoking: Data from the Oakland Growth Study." Social Science and Medicine, 1: 357-382. Deci, Edward L. 1975 Intrinsic Motivation. New York: Plenum Press. Eichorn, Dorothy H. 1981 "Samples and procedures." In Dorothy H. Eichorn, John A. Clausen, Norma Haan, Marjorie P. Honzik, and Paul H. Mussen (eds.), Present and Past in Middle Life: 33-54. New York: Academic Press. Eichorn, Dorothy H., and John A. Clausen, Norma Haan, Marjorie P. Honzik, and Paul H. Mussen (eds.) 1981 Present and Past in Middle Life. New York: Academic Press. Fisher, V. E., and Joseph V. Hanna 1931 The Dissatisfied Worker. New York: MacMillan. Funder, David C., and Daniel Ozer 1983 "Behavior as a function of the situation." Journal of Personality and Social Psychology, 44: 107-112. Gechman, Arthur S., and Yoash Wiener 1975 "Job involvement and satisfaction as related to mental health and personal time devoted to work." Journal of Applied Psychology, 60: 521-523. Goodman, Paul S. 1979 Assessing Organization Change: The Rushton Quality of Work Experiment. New York: Wiley-lnterscience. Griffin, Ricky A. 1983 "Objective and social sources of information in task redesign: A field experiment." Administrative Science Quarterly, 28: 184-200. Haan, Norma 1981 "Common dimensions of personality development: Early adolescence to middle life." In Dorothy H. Eichorn, John A. Clausen, Norma Haan, Marjorie P. Honzik, and Paul H. Mussen (eds.), Present and Past in Middle Life: 117-153. New York: Academic Press. Hackman, J. Richard, and Greg R. Oldham 1975 "Development of the Job Diagnostic Survey." Journal of Applied Psychology, 60: 159170. 1976 "Motivation through the design of work: Test of a theory." Organizational Behavior and Human Performance, 16: 250279. 1980 Work Redesign. Reading, MA: Addison-Wesley. Hackman, J. Richard, Greg R. Oldham, Robert Janson, and Kenneth Purdy 1975 "A new strategy for job enrichment." CaliforniaManagement Review, Summer: 57-71. Herzberg, Frederick 1966 Work and the Nature of Man. Cleveland: World. Herzberg, Frederick, Bernard Mausner, Richard 0. Peterson, and Dora F. Capwell 1957 Job Attitudes: Review of Research and Opinion. Pittsburgh: Psychological Service of Pittsburgh. Hoppock, Robert 1935 Job Satisfaction. New York: Harper. Kahn, Robert L. 1981 Work and Health. New York: Wiley. Kalleberg, Arne L., and Karyn A. Loscocco 1983 "Aging, values, and rewards: Explaining age differences in job satisfaction." American Sociological Review, 48: 7890. Kim, Jae-On, and Charles W. Mueller 1978 FactorAnalysis: Statistical Methods and Practical Issues. Beverly Hills, CA: Sage.

75/ASQ, March 1986

Kohn, Melvin L., and Carmi School er 1978 "The reciprocal effects of the substantive complexity of work and intellectual flexibility: A longitudinal assessment." American Journal of Sociology, 84: 24-52. Kornhauser, Arthur W. 1965 Mental Health of the Industrial Worker: A Detroit Study. New York: Wiley. Landy, FrankJ. 1978 "An opponent process theory of job satisfaction." Journal of Applied Psychology, 63: 533547. Lawler, Edward E., IlIl 1982 "Increasing worker involvement to enhance organizational effectiveness." In Paul Goodman (ed.), Change in Organizations: 280-31 5. San Francisco: Jossey-Bass. Lepper, Mark R., David Greene, and Richard E. Nisbett 1973 "Undermining children's intrinsic interest with extrinsic reward: A test of the 'overjustification hypothesis'." Journal of Personality and Social Psychology, 28: 129-1 37. Locke, Edwin A. 1976 "The nature and causes of job satisfaction." In Marvin D. Dunnette (ed.), Handbook of Industrialand Organizational Psychology: 1297-1 349. Chicago: Rand McNally. Maslow, Abraham H. 1954 Motivation and Personality. New York: Haper& Row. McGowan, John, and John Gormly 1976 "Validationof personality traits: A multicriteria approach." Journal of Personality and Social Psychology, 34: 791-795. Mischel, Walter 1968 Personality and Assessment. New York: Wiley. Monson, Thomas C., John W. Hesley, and Linda Chernick 1982 "Specifying when personality traits can and cannot predict behavior: An alternative to abandoning the attempt to predict single act criteria." Journal of Personality and Social Psychology, 43: 385-399.

Mortimer, Jeylan T., Michael D. Finch, and Geoffrey Maruyama 1985 "Work experience and job satisfaction: Variationby age and gender." Paper presented, the Annual Meeting of the American Association for the Advancement of Science, Los Angeles. Munsterberg, Hugo 1913 Psychology and IndustrialEfficiency. Boston: Houghton Mifflin. Near, Janet R., Robert W. Rice, and Raymond G. Hunt 1978 "Work and extra-work correlates of life and job satisfaction." Academy of Management Journal, 21: 248-264. Oldham, Greg R., and J. Richard Hackman 1980 "Work design in the organizational context." In BarryM. Staw and LarryL. Cummings (eds.), Research in Organizational Behavior, 2: 247-278. Greenwich, CT: JAI Press. O'Reilly, Charles A., and David F. Caldwell 1979 "Informational influence asa determinant of perceived task characteristics and job satisfaction." Journal of Applied Psychology, 64: 157-1 65. O'Reilly, Charles A., G. Nicholas Parlette, and Joan R. Bloom 1980 "Perceptual measures of task characteristics: The biasing effects of differing frames of reference and job attitudes." Academy of Management Journal, 23: 118-131. Orpen, Christopher 1978 "Work and nonwork satisfaction: A causal-correlational analysis." Journal of Applied Psychology, 63: 530-532. Plomin, Robert, J. C. DeFries, and Gerald E. McClearn 1980 Behavioral Genetics: A Primer. San Francisco: Freeman. Pulakos, Elaine D., and Neal Schmitt 1983 "A longitudinal study of a valence model approach for the prediction of job satisfaction of new employees." Journal of Applied Psychology, 68: 307312. Quinn, Robert P., and Graham L. Staines 1979 The 1977 Quality of Employment Survey. Ann Arbor, Ml: Institute for Social Research.

Quinn, Robert P., Graham L. Staines, and Margaret R. McCollough 1974 "Job satisfaction: Is there a trend?" Washington, DC: U.S. Department of Labor. Roberts, Karlene H., and William Glick 1981 "The job characteristics approach to task design: A critical review." Journal of Applied Psychology, 66: 193-217. Ross, Jerry, and Kenneth R. Ferris 1981 "Interpersonal attraction and organizational outcomes: A field examination." Administrative Science Quarterly, 26: 61 7-632. Salancik, Gerald R., and Jeffrey Pfeffer 1977 "An examination of need satisfaction models of job attitudes." Administrative Science Quarterly, 22: 427-456. 1978 "A social information processing approach to job attitudes and task design." Administrative Science Quarterly, 23: 224-253. Schmitt, Neal, and Elaine D. Pulakos 1986 "Predicting job satisfaction from life satisfaction: Is there a general satisfaction factor?" International Review of Applied Psychology (in press). Schneider, Benjamin 1983 "Interactional psychology and organizational behavior." In L. L. Cummings and BarryM. Staw (eds.), Research in Organizational Behavior, 5: 1-31. Greenwich, CT: JAI Press. Schneider, Benjamin, and H. Peter Dach ler 1978 "A note on the stability of the job description index." Journal of Applied Psychology, 63: 650-653. Schneider, Benjamin, John J. Parkington, and Virginia M. Buxton 1980 "Employee and customer perceptions of service in banks." Administrative Science Quarterly, 25: 252-267. Sims, Henry P., Andrew D. Szilagyi, and Robert T. Keller 1976 "The measurement of job characteristics." Academy of Management Journal, 19: 195-212.

76/ASQ, March 1986

Job Attitudes
Smith, C. Ann, Dennis W. Organ, and Janet Near 1983 "Organizational citizenship behavior: Its nature and antecedents." Journal of Applied Psychology, 69: 653-663. Staw, Barry M., Bobby J. Calder, and Randall K. Hess 1980 "Intrinsic motivation and norms about payment." Journal of Personality, 48: 1-14. Staw, Barry M., and Jerry Ross 1985 "Stability in the midst of change: A dispositional approach to job attitudes." Journal of Applied Psychology, 70: 469-480. Taber,Tom D., Terry A. Beehr, and Jeffrey T. Walsh 1985 "Relationships between job evaluation ratings and selfratings of job characteristics." Organizational Behavior and Human Decision Processes, 35: 27-45. Terborg, James R. 1981 "Interactional psychology and research on human behavior in organizations." Academy of Management Review, 6: 569576. Vroom, Victor H. 1964 Work and Motivation. New York: Wiley. Wahba, Mahmoud A., and Lawrence G. Bridwell 1976 "Maslow reconsidered: A review of research on the need hierarchy theory." Organizational Behavior and Human Performance, 15: 212-240. Watson, David, and Lee Anna Clark 1984 "Negative affectivity: The disposition to experience aversive emotional states." Psychological Bulletin, 96: 465-490. Weaver, Charles N. 1978 "Job satisfaction as a component of happiness among males and females." Personnel Psychology, 31: 831-840. 1980 "Job satisfaction in the United States in the 1970's." Journal of Applied Psychology, 65: 364-367. Weiss, Howard M., and Seymour Adler 1984 "Personality and organizational behavior." In BarryM. Staw and L. L. Cummings (eds.), Research in Organizational Behavior, 6: 1-50. Greenwich, CT: JAI Press. Weiss, Howard M., and James B. Shaw 1979 "Social influences on judgments about tasks." Organizational Behavior and Human Performance, 24: 126-140. White, J. Kenneth 1978 "Individualdifferences and the job quality-worker response relationship: Review, integration, and comments." Academy of Management Review, 3: 267280. White, Sam E., and Terence R. M itchel l 1979 "Job enrichment versus social cues: A comparison and competitive test." Journal of Applied Psychology, 64: 1-9. Zuckerman, Marvin 1977 "Development of a situationspecific trait-state test for the prediction and measurement of affective responses." Journal of Consulting and Clinical Psychology, 45: 513-523.

77/ASQ, March 1986