Journal of Applied Psychology1983, Vol 68, No 4, 653-663Copyright 1983 by theAmerican Psychological Association, incOrganizational Citizenship Behavior: Its Nature and AntecedentsC. Ann Smith, Dennis W. Organ, and Janet P. NearSchool of Business, Indiana UniversityIt is argued here that a category of performance called citizenship behavior isimportant in organizations and not easily explained by the same incentives thatinduce entry, conformity to contractual role prescriptions, or high production Astudy of 422 employees and their supervisors from 58 departments of two bankssought to elaborate on the nature and predictors of citizenship behavior Resultssuggest that citizenship behavior includes at least two separate dimensions Altruism,or helping specific persons, and Generalized Compliance, a more impersonal formof conscientious citizenship Job satisfaction, as a measure of chronic mood state,showed a direct predictive path to Altruism but not Generalized Compliance Ruralbackground had direct effects on both dimensions of citizenship behavior Thepredictive power of other variables (e g , leader supportiveness as assessed independently by co-workers, personality measures) varied across the two dimensionsof citizenship behaviorNearly two decades ago, Katz (1964) identified three basic types of behavior essentialfor a functioning organization (a) People mustbe induced to enter and remain within thesystem, (b) they must carry out specific rolerequirements in a dependable fashion; and (c)there must be innovative and spontaneous activity that goes beyond role prescriptions.Concerning this third category, Katz noted,"An organization which depends solely uponits blue-prints of prescribed behavior is a veryfragile social system" (p. 132). Every factory,office, or bureau depends daily on a myriadof acts of cooperation, helpfulness, suggestions,gestures of goodwill, altruism, and other instances of what we might call citizenship behavior.Citizenship behaviors comprise a dimensionof individual and group functioning thatRoethhsberger and Dickson (1964) seemed tohave in mind when they used the term cooperation A close reading of the concludingchapters (22-26) in Management and theWorker reveals that cooperation refers tosomething other than productivity. The latterThe data reported m this article are drawn from a doctoral dissertation project conducted by the first author,who died on April 27, 1982 She was posthumouslyawarded the degree of Doctorate in Business Administration m August, 1982Requests for reprints should be addressed to DennisW Organ, School of Business, Indiana University, Bloommgton, Indiana 47405was regarded as a function of the formal organization (the authority structure, role specifications, technology) and the "logic of facts "Cooperation, on the other hand, referred toacts that served more of a maintenance purpose, to "maintain internal equilibrium " Cooperation thus included the day-to-day spontaneous prosocial gestures of individual accommodation to the work needs of others (e.g.,co-workers, supervisor, clients in other departments), whereas productivity (or efficiency) was determined by the formal or economic structure of the organization. Roethhsberger and Dickson viewed cooperation asa product of informal organization and, significantly, the "logic of sentiment." The latterwas seen as influenced both by the quality ofwork experience and by previous social conditioningFailure to recognize these subtle distinctionsmay account for the often-voiced criticism(e.g., Lawler & Porter, 1967) of the Hawthorneresearch and "Human Relations" school asnaively proposing that satisfaction causes performance. To be sure, Roethlisberger andDickson implied that, at the aggregate level ofanalysis (e g., the firm) and over the long run,efficiency and cooperation were interdependent with each other But at the individuallevel of analysis, the emphasis on sentimentwas due to the presumed connection to cooperation, or citizenship behavior.Substantively, citizenship behaviors are lm-653
654C SMITH, D ORGAN, AND J. NEARportant because they lubricate the social machinery of the organization. They provide theflexibility needed to work through many unforeseen contingencies, they enable participants to cope with the otherwise awesomecondition of interdependence on each other.Theoretically, they are of interest precisely because they cannot, as Katz noted, be accountedfor by the same motivational bases as thosethat induce people to join, stay, and performwithin contractual, enforceable role prescriptions. Because citizenship behavior goes beyond formal role requirements, it is not easilyenforced by the threat of sanctions. Furthermore, much of what we call citizenship behavior is not easily governed by individual incentive schemes, because such behavior is oftensubtle, difficult to measure, may contributemore to others' performance than one's own,and may even have the effect of sacrificingsome portion of one's immediate individualoutput To be sure, frequent acts of citizenshipbehavior will often be noted by organizationofficials (e.g., supervisors), and undoubtedlythis has some influence on subjective appraisals of individual performance. But given themfrequency and unsystematic nature of mostappraisal systems, coupled with the fact thatmany supervisors have limited control overformal rewards, it seems unlikely that mostof the variance in "good citizen" behavior isexplained by the calculated anticipation thatthey will pay off in largesse for the person.Consider, by analogy, the larger social order.A society functions for better or for worse asa consequence of the frequency of many actsof citizenship (i.e., spontaneous charitable actsto specific others, as well as more impersonalprosocial conduct) that are either not requiredby law or are essentially unenforceable by theusual incentives or sanctions Thus, it wouldseem to be a worthwhile exercise to inquireinto the antecedents of such behavior in organizations.Determinants of Citizenship BehaviorBecause much of what we call citizenshipbehavior has an altruistic character, it seemedworthwhile to explore the social psychologyliterature for determinants of altruism Theresults of a number of studies (e.g., Berkowitz& Connor, 1966; Isen, 1970; Isen & Levin,1972; Levin & Isen, 1975) can be summarizedby concluding that mood state influences theprobability of prosocial gestures. Subjects mwhom a mood of positive affect had been induced—whether by prior success on a challenging task, the good fortune of receivingsome windfall, or simply quiet meditation onpast enjoyable experiences—were more likelyto behave altruistically Conversely, subjects inwhom a negative mood (e.g , of frustration,disappointment, anger) had been aroused wereless likely to show prosocial behavior. Thus,we might tentatively propose that job satisfaction, to the extent that it represents a characteristic or enduring positive mood state,would account for some portion of citizenshipbehavior.Bateman and Organ (Note 1) found thatjob satisfaction, as measured by the Job Descriptive Index, did correlate with the extentof citizenship behavior as independently ratedby supervisors However, this study, which wascharacterized by a longitudinal, two-wavepanel design, found the association to be limited to the concurrent correlations; there wasno significant difference in the cross-laggedcorrelations. Thus, the correlation did notreally pass the test of spuriousness. The authorssuggested that other environmental factors(e g., leader supportiveness) or individual attributes (e.g., traits, such as neuroticism) mightindependently affect both satisfaction and citizenship behavior.On a general level, three alternative models,as shown in Figure 1, invite scrutiny. The firstof these is consonant with a mood explanationof citizenship behavior. In this scheme, characteristic level of job satisfaction predicts citizenship behavior. Environmental and individual difference variables affect citizenshipbehavior only indirectly via satisfaction. Thesecond model would accord a direct linkagefrom environmental and personality factors tocitizenship behavior with concurrent, independent effects of those factors on satisfaction,thus rendering satisfaction and citizenship behavior correlated but functionally unrelated.The third model would account for citizenshipbehavior by a combination of direct effectsfrom environmental and personality variablesas well as indirect effects through satisfaction.What environmental dimensions have directimplications for citizenship behavior and why?
655ORGANIZATIONAL CITIZENSHIP BEHAVIOR(a)WorkplaceEnvironment.Satisfaction- CitizenshipBehavior ty(c)WorkplaceEnvironmentPersonalityFigure 1 Competing models of prosystem maintenance behaviorLeader supportiveness may represent one suchdimension, for two reasons First of all, muchof supervisor consideration is, m itself, citizenship behavior (i.e., discretionary acts aimedat helping others) Thus, the supervisor servesto some extent as a model, and social psychological studies (e.g., as reviewed by Krebs,1970, and Berkowitz, 1970) strongly suggestthat many forms of prosocial behavior are influenced by models. Models provide cues forwhat behavior is appropriate and make salientthe situational needs for prosocial gestures.Second, at some point leader supportivenessinitiates a pattern of exchange that is socialand noncontractual in character (Dansereau,Graen, & Haga, 1975). The exchange becomessubject to broader norms of reciprocity(Gouldner, 1960) or equity (Adams, 1965)Subordinates may choose citizenship behavioras a means of reciprocation to superiorsMoreover, they may choose citizenship behavior, as opposed to increased productivity,because variation in the latter is more constrained by ability, work scheduling, or taskdesign. And such reciprocity may occur netof any effect of supportiveness on general jobsatisfaction, because the latter may be largelydetermined by factors (e.g., pay) beyond thesupervisor's controlA second environmental variable possiblydirectly affecting citizenship behavior is taskinterdependence. Specifically, task groupscharacterized by reciprocal interdependenceshould display more citizenship behavior thangroups m which independence or sequentialdependence is the rule. Reciprocal interdependence, according to Thompson (1967), requires frequent instances of spontaneous mutual adjustment in order to effect coordination.This requirement presumably fosters socialnorms of cooperation, helping, and sensitivityto others' needs and makes salient a collectivesense of social responsibility (Krebs, 1970). Atthe same time, it tends to promote, ceteris
656C SMITH, D ORGAN, AND J NEARpanbus, higher levels of group cohesion thanother task environments (Seashore, 1954);thus, because cohesion influences satisfaction,task interdependence is a potential source ofcommon variance in both mood and citizenship behavior.Citizenship behavior may represent just onemanifestation of a broader disposition towardprosocial behavior. Insofar as individual attributes are concerned, the social psychologyliterature on altruism (Krebs, 1970) suggeststhat extraversion is positively correlated withprosocial behavior and that neuroticism(emotional instability) bears a negative relationship with such behavior. Extraverts tendto be more sensitive to their external environments, more sensitive to social stimuli, andmore prone to spontaneity in behavior Thosewho score high in neuroticism tend to be morepreoccupied with their own anxieties and presumably do not have the emotional staminato concern themselves with others' problemsor general system requirements unrelated totheir own immediate needs.More recent research attests to the existenceof a stable dimension of individual differencesconcerning "belief in a just world" (Lerner &Miller, 1978). And there is evidence (Zuckerman, 1975) that persons strongly endorsingthis belief system are more likely to exhibitprosocial behavior. Presumably, individualswho have such beliefs are confident that theywill, somehow and in good time, be rewardedfor any charitable or responsible conduct (i.e.,they accept the admonition from Ecclesiastes11:1, "Cast thy bread upon the waters: forthou shalt find it after many days").Finally, certain demographic variables have,with varying degrees of consistency, been citedas predictors of altruism or other forms ofprosocial behavior. Ordinal birth position appears to have some significance in this regard,with firstborn subjects found to show morealtruism (Krebs, 1970). A review by Gergen,Gergen, and Meter (1972) suggests educationallevel to be positively correlated with generalsocial responsibility. The significance of urbanversus rural origins with respect to altruismis more tenuous (Hansson, Slade, & Slade,1978). In a work setting, one might expect tosee more citizenship behavior displayed bythose of rural or small-town origins, if we accept the hypothesis that such persons morereadily endorse a work ethic (Hulm & Blood,1968) such that citizenship behavior is an endin itself.In summary, this study attempted to assessthe extent to which "good citizenship" behavior could be accounted for by characteristicmood state and the extent to which certainenvironmental forces and individual differencevariables could independently predict citizenship behavior.MethodSampleEmployees and their supervisors, representing 77 departments of two banks of a large midwestern city, participated in the study During working hours, the subordinates in each department responded to a questionnairethat contained measures of general job satisfaction, personality, leader supportiveness, task interdependence, anditems asking for demographic information Within eachdepartment, employees were randomly divided into twogroups (a) subjects, from whom only the satisfaction, personality, and demographic data were used, and (b) descriptors, from whom only the responses concerning leadersupportiveness and task interdependence were used Theresponses of the descriptors within each department wereaveraged to provide indices of supervisor supportivenessand task interdependence as environmental factors, procedurally independent of the subjects' responses to satisfaction and individual difference measures Supervisorsresponded to a separate questionnaire that asked them toassess each of their subordinates on items comprising ameasure of citizenship behavior The complete set of supervisory assessments of all subordinates was used onlyin preliminary analyses in order to assess psychometricproperties of the citizenship behavior measure Their assessments of the randomly selected subjects only were usedin the major analyses of relationships among variablesWork units with fewer than four subordinates wereeliminated so that the subject and descriptor groups ineach department would contain at least two persons Inunits with an odd number of employees, the extra respondent was assigned to the subject group Ultimately,usable data were obtained from 58 departments with 422respondents, of whom 220 served as subjects and 202 asdescriptorsMeasuresCitizenship behavior was denned by the 16 items shownin Table 1 These items represent the final product ofsemistructured interviews with a number of managersrepresenting other organizations not included in the studyThese managers were asked to identify instances of helpful,but not absolutely required, job behavior A pool of suchitems was pilot tested with a group of 67 full-time employedmanagers enrolled in evening business classes in an urbancampus setting Respondents were asked to think of anemployee who worked or had worked for them and torate, on a 5-point scale, how characteristic each statement
657ORGANIZATIONAL CITIZENSHIP BEHAVIORTable 1Principal-Factor Analysis With Varimax Rotation Citizenship BehaviorFactor 1AltruismItem12345678910111213141516Helps others who have been absentPunctualityVolunteers for things that are not requiredTakes undeserved breaks *Orients new people even though it is not requiredAttendance at work is above the normHelps others who have heavy work loadsCoasts towards the end of the day*Gives advance notice if unable to come to workGreat deal of time spent with personal phone conversations*Does not take unnecessary time off workAssists supervisor with his or her workMakes innovative suggestions to improve departmentDoes not take extra breaksAttend functions not required but that help company imageDoes not spend time in idle conversationEigenvaluePercent variance explainedCumulative percentage of variance explained1123J78211121Ik33221207JOFactor 2GeneralizedCompliance24kl28 1104.5931.3911 likl06392508.63.34Ji155 4038 638 62 1715.554 1IkNote Factor loadings of 50 and above are underscoredaReversed scoringwas of the employee The responses were submitted tofactor analysis, with communahty estimates in the diagonalsand using orthogonal vanmax rotation Results suggestedtwo fairly interpretable and distinct factors, and the samefactor structure emerged from responses of the 422 personsin the present studyFactor 1 (see Table 1) appears to capture behavior thatis directly and intentionally aimed at helping a specificperson in face-to-face situations (e g, orienting new people,assisting someone with a heavy workload) The elicitingstimulus, in other words, is someone needing aid, as mthe fashion of social psychological studies of altruism Thus,this dimension is referred to as AltruismFactor 2, by contrast, pertains to a more impersonalform of conscientiousness that does not provide immediateaid to any one specific person, but rather is indirectlyhelpful to others involved in the system The behavior(e g , punctuality, not wasting time) seems to representsomething akin to compliance with internalized normsdefining what a "good employee ought to do " This factorwill be referred to as Generalized Compliance Only thoseitems loading above 50 on one factor and less than 50on the other were scored Coefficient alpha reliability estimates were 88 and 85, respectively, for the two factorsFor data analytic purposes in the major part of thisstudy, factor scores were estimated by simple summationof subjects' scores on the items loading on each factor Asshown below, this resulted in a correlation of 45 (p 001) between the two supposedly independent factorsThis result prompted a reexamination of the factor structure, this time using an oblique rotation The factor structure that emerged was identical to that produced by orthogonal rotation, with virtually identical loadings, andthe factor pattern correlation was 43, virtually identicalto that obtained when using unit weights to estimate factorscores Thus, given the method used here to estimate factorscores, the question of orthogonality did not make muchdifference empirically, because either assumption ultimatelyled to the same degree of relatedness Conceptually, itseems more plausible to assume some degree of associationbetween the factors, given the nature of what the constructsare intended to representJob satisfaction Scott's (1967) semantic differentialscale of the concept "me at work" was used This measureconsists of 20 bipolar adjectives separated by 7-pomt scales.Only those adjectives (e g, satisfied-dissatisfied, penakzedrewarded, interested-bored) shown to load on the factorAffective Tone were used as the measure of facet-free generalized job satisfaction as a chrome mood stateLeader supportiveness The measure used in this studywas the Supportive Leadership Behavior scale, developedby House and Dessler (1974) It consists of 10 questionsdrawn from Form XII of the Ohio State Leader BehaviorDescription Questionnaire (Stogdill, 1965) Seven itemstapping other dimensions of leader behavior were usedsolely as filler itemsTask interdependence The instrument used was anindex developed by Van de Ven, Delbecq, and Koenig
behavior will often be noted by organization officials (e.g., supervisors), and undoubtedly this has some influence on subjective apprais-als of individual performance. But given the mfrequency and unsystematic nature of most appraisal systems, coupled with the fact that many supervisors have limited control over formal rewards, it seems unlikely that most of the variance in "good citizen .