Do Other-Reports of Counterproductive Work Behavior Provide an

Do Other-Reports of Counterproductive Work Behavior Provide anIncremental Contribution Over Self-Reports? A Meta-Analytic Comparison

Christopher M. Berry, Nichelle C. Carpenter, and Clare L. BarrattTexas A&M University

Much of the recent research on counterproductive work behaviors (CWBs) has used multi-item self-report measures of CWB. Because of concerns over self-report measurement, there have been recent callsto collect ratings of employees’ CWB from their supervisors or coworkers (i.e., other-raters) asalternatives or supplements to self-ratings. However, little is still known about the degree to whichother-ratings of CWB capture unique and valid incremental variance beyond self-report CWB. Thepresent meta-analysis investigates a number of key issues regarding the incremental contribution ofother-reports of CWB. First, self- and other-ratings of CWB were moderately to strongly correlated witheach other. Second, with some notable exceptions, self- and other-report CWB exhibited very similarpatterns and magnitudes of relationships with a set of common correlates. Third, self-raters reportedengaging in more CWB than other-raters reported them engaging in, suggesting other-ratings capture anarrower subset of CWBs. Fourth, other-report CWB generally accounted for little incremental variancein the common correlates beyond self-report CWB. Although many have viewed self-reports of CWBwith skepticism, the results of this meta-analysis support their use in most CWB research as a viablealternative to other-reports.

Keywords: counterproductive work behavior, workplace deviance, self-report, meta-analysis

Counterproductive work behaviors (CWBs) refer to voluntaryemployee behaviors that are viewed by the organization as con-trary to its legitimate interests, violate significant organizationalnorms, and threaten the well-being of the organization or itsmembers (e.g., theft, withholding effort, interpersonal aggression,poor attendance; Bennett & Robinson, 2000; Sackett & DeVore,2002). Studies have estimated that CWBs not only cost organiza-tions billions of dollars annually (Camara & Schneider, 1994;Murphy, 1993; Vardi & Weitz, 2004), but they have negativeconsequences for employees as well. For instance, being the targetof CWB can lead to decreased job satisfaction and increased stressand intentions to quit, among other things (Budd, Arvey, & Law-less, 1996; Glomb, 2002). As a result, a great deal of research hasbeen devoted to the measurement and prediction of CWBs (Berry,Ones, & Sackett, 2007).

Until recently, there was a tendency to treat each type ofCWB as discrete, resulting in separate literatures focused on themeasurement of specific CWBs such as theft or harassment(Sackett, 2002). Although separate literatures devoted to indi-vidual CWBs still exist and flourish, recent years have seen anincreased interest in more integrative measurement of CWB.Much of the recent research on CWB uses multi-item measures,such as the Workplace Deviance Scale (Bennett & Robinson,2000) and Counterproductive Work Behavior Checklist (Fox,

Spector, & Miles, 2001), that include a wide range of CWBs(Rotundo & Spector, 2010). These multi-item measures ofCWB acknowledge the positive manifold across individualCWBs, and the theoretical perspective that these individualbehaviors reflect a broad, hierarchical construct with a generalCWB factor at the top level, group factors such asinterpersonal- and organizational target CWB beneath the gen-eral factor, and specific CWBs below the group factors (Berry,Ones, & Sackett, 2007; Sackett & DeVore, 2001). Furthermore,this broad CWB construct has been shown to relate to variousorganizationally relevant variables such as Big Five personal-ity, negative affectivity, organizational justice, and job satis-faction, among others (Berry, Ones, & Sackett, 2007; Dalal,2005; Hershcovis et al., 2007). Thus, there is value in measur-ing and predicting a broad CWB domain (Sackett, 2002). Thepresent meta-analysis concentrates on studies with such a focuson broad measurement of CWB via multi-item measures.

A number of recent meta-analyses have used broad, multi-itemCWB measures to address important research questions surround-ing CWB. For instance, Berry, Ones, and Sackett’s (2007) meta-analysis supported Sackett and DeVore’s (2001) proposition thatCWB is a hierarchical construct consisting of separateinterpersonal- and organizational target lower order factors. Meta-analyses have also increased understanding of CWB’s nomologi-cal net by summarizing the relationships that CWB has with a widerange of organizationally relevant variables (e.g., Berry, Ones, andSackett, 2007; Dalal, 2005; Hershcovis et al., 2007). Althoughthese meta-analyses have advanced knowledge related to the na-ture of CWB, these meta-analyses have in common one importantlimitation. Specifically, the primary studies included in each of theabove meta-analyses relied almost completely on single-source

This article was published Online First December 26, 2011.Christopher M. Berry, Nichelle C. Carpenter, and Clare L. Barratt,

Department of Psychology, Texas A&M University.Correspondence concerning this article should be addressed to Christo-

pher M. Berry, Department of Psychology, Texas A&M University, 4235TAMU, College Station, TX 77843-4235. E-mail: [email protected]

Journal of Applied Psychology © 2011 American Psychological Association2012, Vol. 97, No. 3, 613–636 0021-9010/12/$12.00 DOI: 10.1037/a0026739

613

self-reports of CWB.1 This is not surprising, considering that themost popular multi-item CWB measures were originally designedas self-report measures (e.g., Bennett & Robinson, 2000; Fox etal., 2001). However, there have been recent calls to use suchmulti-item measures to collect supervisor and/or coworker ratingsof CWB (we hereafter refer to supervisor and coworker ratings asother-ratings) as alternatives or supplements to self-ratings (e.g.,Fox, Spector, Goh, & Bruursema, 2007; Stewart, Bing, Davison,Woehr, & McIntyre, 2009). This highlights two gaps in the presentunderstanding of CWB. First, it is not clear whether the abovemeta-analyses’ results may have been affected by a range of issuesgermane to the use of self-report measures such as commonmethod variance and social desirability bias (Podsakoff, MacKen-zie, Lee, & Podsakoff, 2003). Second, it is unclear the extent towhich the present understanding of CWB—obtained primarilyfrom self-report measures—generalizes to CWB assessed throughnon-self-report measures.

We address these gaps in the present study by empiricallyinvestigating a number of important issues regarding the degree towhich using self- versus other-ratings makes a difference in CWBresearch. For instance, what is the correlation between self- andother-reports of CWB? Do self- and other-reports of CWB relatedifferently to theoretically important variables? Do self- and other-raters report different mean levels of CWB for employees? Does apattern of results emerge suggesting one method of measurementis preferable? We use meta-analysis in the present study to draw onmultiple samples and thousands of participants, which thereforerepresents a powerful test of the above research questions.

Review of the Literature and Development ofHypotheses

Most fundamentally, we ask in the present study: Are self-reportCWB measures a problem, and are other-reports the solution?Self-report has long been a common mode of measurement forCWB researchers (Fox et al., 2007). This is in part due to a numberof important advantages that self-report measures of CWB haveover other-report CWB measures. First, because many CWBsare relatively covert behaviors that employees engage in with theintention of not getting caught, the only source that has completeknowledge of an employee’s engagement in CWB is the employee.Second, having employees anonymously report their own CWBcircumvents some ethical concerns associated with supervisor orcoworker reports (Fox & Spector, 1999). That is, drawing attentionto employees’ negative organizational behaviors by asking theirsupervisors or coworkers to rate the employees’ CWB could havenegative consequences for the employees, calling into questionwhether risk is adequately minimized. Third, publicly available,multi-item, and validated self-report measures of CWB have ex-isted for years (e.g., Bennett & Robinson, 2000; Fox & Spector,1999; Marcus, Schuler, Quell, & Humpfner, 2002), and it is easierto administer such measures to employees themselves than toenlist supervisor or coworker raters.

Regardless, self-report measures of CWB are not without theirdisadvantages. Unlike many variables in applied psychology,CWB involves highly sensitive inquiries about potentially self-incriminating information. So, one concern is the possibility thatemployees underreport the extent to which they engage in CWB(e.g., Fox et al., 2007). This underreporting might be due to the

fear of getting caught and being punished (R. M. Lee, 1993) or dueto a general reluctance to describe oneself in negative terms (e.g.,Heneman, Heneman, & Judge, 1997). Another concern with self-report CWB is common method bias (e.g., Campbell & Fiske,1959). When employees provide self-reports of CWB along withself-reports of other relevant variables, relationships betweenCWB and these other variables can be artificially inflated.

These sorts of disadvantages have recently caused a number ofresearchers to voice concerns over the use of single-source self-report measurement of CWB. For example, Barclay and Aquino(2010) listed self-report measurement as a methodological issuethat has posed an obstacle for CWB research. Fox et al. (2007)made the point that although using self-reports from employeesmay be the most viable methodology for early stages of CWBresearch, “the CWB community is now at the point that moreobjective, or nonincumbent, measures are needed to further ourunderstanding of the phenomenon” (p. 43). Other recent articleshave made similar points (e.g., Stewart et al., 2009). Implicit in allof these concerns is the idea that the present CWB knowledge basederived using multi-item self-report measures is deficient andperhaps even misleading and that the use of other sources of CWBinformation will tell us something new about explaining andpredicting CWB.

Thus, CWB researchers have been relying more and more onother-reports of employees’ CWB, despite the difficulty of col-lecting such other-reports relative to collecting self-reports. Other-reports of CWB mitigate some of the key concerns over self-reportCWB (e.g., common method bias, underreporting due to fear ofbeing caught). However, there are clear disadvantages to other-reports of CWB. Supervisors and coworkers may not have ade-quate opportunity to observe employees engaging in CWB. Addi-tionally, other-raters may fear retribution for reporting employees’CWB and therefore may be reluctant to provide accurate informa-tion (Fox et al., 2007). So, if researchers are to go to all of thetrouble of collecting other-ratings (or multisource ratings) ofCWB, then it is important to understand what additional, uniqueinformation is gained beyond what is presently known aboutself-report CWB.

1 This is not meant to imply that virtually all past CWB research hasrelied on self-report CWB. For example, literatures devoted to specificCWBs, such as absenteeism or theft, have often used organizational re-cords or other non-self-reports. Additionally, Ones, Viswesvaran, andSchmidt’s (1993) meta-analysis of the validity of integrity tests for pre-dicting CWB carried out moderator analyses wherein self-report CWBmeasures were considered separately from other-report CWB measures(although many of the CWB measures were narrow, individual CWBs suchas absenteeism and theft, in contrast to the broad, multi-item measuresincluding a wide range of CWBs that are the focus of the present study).More recently, studies by Dilchert et al. (2007) and Stewart et al. (2009),among others, have used broad, multi-item, nonself-report measurement ofCWB. We simply make the point in the present article that the vastmajority of primary studies in the past decade in which broad, multi-itemmeasures of CWB have been used have relied on self-report. Therefore, thecited meta-analyses that focused on prediction of broad, multi-item mea-sures of CWB also were forced to rely mostly on single-source self-reports.

614 BERRY, CARPENTER, AND BARRATT

Relationship Between Self- and Other-Report CWB

The correlation between self-reports and other-reports of CWBwas meta-analyzed in the present study. If the correlation is high,then this suggests that self- and other-ratings of CWB provide littleunique information. In this case, collecting CWB data from onesource would be essentially equivalent to collecting it from theother source or from multiple sources, and therefore the efficacy ofcollecting data from sources other than the self would be calledinto question. If the correlation between self- and other-ratings ofCWB is low, then this would suggest that the two sources providehighly unique information. To the degree that the unique informa-tion from each source accurately represents true levels of CWB,this means potential for incremental validity is high and thatmultisource ratings of CWB might increase knowledge and pre-diction of CWB. This would also suggest that the cumulativeliterature on CWB, which has relied mostly on self-report CWB(e.g., Berry, Ones, & Sackett, 2007; Dalal, 2005; Hershcovis et al.,2007), is not generalizable beyond self-report. This would meanthat new meta-analytic evidence regarding the nomological net ofother-report CWB would be needed.

There are reasons to expect that self-reports and other-reportsof CWB will each provide unique information but still overlapto some degree. Some reasons to expect unique information arethat other-raters may not have an opportunity to observe all ofthe CWBs that employees engage in and that self-raters maymisreport CWBs for fear of being caught or other impressionmanagement concerns. It is, however, unlikely that other-raterswill be completely unaware of employees’ engagement in CWB(especially relatively public CWBs such as harassment or gos-sip) and that the employees themselves will be completelyuntruthful about their engagement in CWB (if employees werecompletely untruthful and reported not engaging in any CWBs,it would be difficult to reconcile this with the cumulativeliterature showing that self-report CWB correlates with a num-ber of relevant variables); thus, it is likely there will be someoverlap in self- versus other-ratings. Therefore, similar to mul-tisource ratings in other domains such as overall job perfor-mance (Conway & Huffcutt, 1997) and personality (Mount,Barrick, & Strauss, 1994), it is hypothesized that there will bea positive correlation between self- and other-ratings of CWBbut that this correlation will not be perfect even when correctedfor error of measurement. Specifically:

Hypothesis 1: The correlation between self- and other-ratingsof CWB will be positive and greater than zero but will notreach unity.

There are reasons to expect certain variables to moderate therelationship between self- and other-reports of CWB. For one, thesource of the other-rating may moderate this relationship. Specif-ically, employees probably make greater efforts to hide their CWBfrom their supervisors, who can formally punish them, than theircoworkers. Additionally, in most settings, coworkers probablywork in closer proximity on a day-to-day basis with employeesthan do supervisors, and accordingly, coworkers should havegreater opportunity to observe and report employees engaging inCWB. Therefore:

Hypothesis 1a: The correlation between self-ratings and co-worker ratings will be greater than the correlation betweenself-ratings and supervisor ratings of CWB.

Another variable that is expected to moderate the relationshipbetween self- and other-ratings of CWB is whether the CWBsare public versus private. As described earlier, the more publicthe CWBs, the more opportunity other-raters should have toobserve those behaviors and make sound judgments of employ-ees’ levels of CWB, and the more self- and other-ratings ofCWB should converge. Interpersonal target CWBs (e.g., curs-ing or being rude to others) are generally more public behaviorsthan organizational target CWBs (e.g., theft of organizationalproperty). Therefore:

Hypothesis 1b: The correlation between self- and other-ratings will be higher for interpersonal target CWB than fororganizational target CWB.

A third variable that is expected to moderate the relationshipbetween self- and other-ratings of CWB is the degree to whichassurances of anonymity have been given to raters. Both self-and other-raters are more likely to provide accurate/truthfulratings of employees’ CWB when they feel more confident thattheir responses are anonymous and confidential, and thus, theirratings should show more convergence. Although most CWBstudies attempt at least some assurances of anonymity forparticipants, there are differences between studies in thestrength of these assurances. For instance, some CWB studiesmay simply tell participants their responses are anonymous,whereas other studies may take precautions such as havingparticipants mail their completed surveys directly to the re-searchers or assigning participants secret codes. The importantpoint is whether participants perceive there is greater anonym-ity of their responses, regardless of whether these sorts of stepstruly do assure greater anonymity. Therefore:

Hypothesis 1c: The correlation between self- and other-ratings of CWB will increase as researchers take more stepsto assure participants of the anonymity of their responses.

Relationships With Theoretically RelevantCommon Correlates

Hypothesis 1 suggests that self- and other-report CWB will eachreflect some unique variance. However, even if other-ratings cap-ture unique variance, if the unique variance is not valid—meaningthat the variance captured by other-rated CWB does not explainincrementally beyond that already explained by self-reportedCWB—it may not be worth going to the trouble to collect other-ratings of CWB. Thus, the second main purpose of the presentmeta-analysis was to advance understanding of the degree towhich other-ratings of CWB have different patterns or magnitudesof relationships with a common set of theoretically relevant vari-ables. If self- and other-ratings of CWB contain unique and validinformation, they should exhibit different patterns and magnitudesof correlations with these theoretically relevant variables. Theo-retical accounts have made the case that CWBs should be relatedto a number of important predictors and correlates, which hasprompted meta-analyses of the relationships between CWB and

615SELF- AND OTHER-REPORTS OF CWB

the Big Five personality traits (Berry, Ones, & Sackett, 2007),negative affectivity (Hershcovis et al., 2007), demographic vari-ables (Berry, Ones, & Sackett, 2007), justice perceptions (Berry,Ones, & Sackett, 2007), interpersonal conflict (Hershcovis et al.,2007), organizational constraints (Herschcovis et al., 2007), jobsatisfaction (Dalal, 2005), and organizational citizenship behaviors(Berry, Ones, & Sackett, 2007). Via correspondence with the leadauthors of each of the listed meta-analyses, it was determined thateach of these meta-analyses relied almost solely on self-reportmulti-item measures of CWB. Therefore, in the present study, eachof these relationships was meta-analyzed, but only including multi-item other-reports of CWB. A comparison of the meta-analyticestimates from the present meta-analysis versus the previous meta-analyses tests the degree to which self-ratings and other-ratings ofCWB have similar versus dissimilar patterns of relationships witha common set of theoretically relevant correlates, and therefore thedegree to which self- and other-ratings account for unique andvalid variance in true CWB.

However, if other-ratings have different patterns or magnitudesof relationships with common correlates, this could also be drivenby problems in ratings from at least one of the sources. Forinstance, if relationships differ in magnitude (e.g., self-ratings aremore strongly related to correlates than other-ratings), commonmethod variance could be inflating relationships between vari-ables. To the degree that this is true, the differing magnitudes ofrelationships that self- and other-report CWB have with theircommon correlates could signal a problem with self-report mea-sures, instead of being evidence that self-reports capture uniquevalid variance. Thus, before one can conclude that differing pat-terns or magnitudes of relationships with common correlates areevidence of unique valid variance, one must test other plausibleexplanations for the unique relationships (e.g., common methodvariance). To this end, we propose two hypotheses (2a and 2b) inthe present meta-analysis regarding reasonable alternative expla-nations for differing patterns/magnitude of relationships self- andother-reports have with their common correlates. First, if commonmethod variance is an issue, then self-reports of CWB will corre-late more strongly with the common correlates (or at least thosecommon correlates measured via self-report) than will other-reports of CWB. Therefore:

Hypothesis 2a: Self-ratings of CWB will correlate morestrongly than other-ratings of CWB with a common set oftheoretically relevant correlates.

Self- and other-report CWB can also differ in the pattern of rela-tionships they have with their common correlates (e.g., self- andother-report CWB may not be similarly related to the same corre-lates). A common concern is that many self-report measures areaffected by social desirability concerns (e.g., Spector, 2006). To thedegree that this concern is legitimate, self-report CWB measureswould be expected to correlate more strongly with additional self-reported variables that are similarly affected by social desirabilityconcerns. Generally, social desirability should be more of a concernfor variables in which employees are being asked to say positive ornegative things about themselves. In the present study, the commoncorrelates measured via self-report differ in the degree to which theyshould be affected by socially desirable responding. For instance,among the Big Five personality variables, Impression Management

scale scores are most strongly related to Conscientiousness, Agree-ableness, and Emotional Stability and less so to Extraversion andOpenness (Berry, Page, & Sackett, 2007; Li & Bagger, 2006; Paulhus& Johns, 1998). Self-reports of job satisfaction (Ones & Viswesvaran,1998) and negative affect (Chen, Dai, Spector, & Jex, 1997) have alsobeen linked to socially desirable responding. Furthermore, becausemost employees likely desire to see themselves as good organizationalcitizens, self-reports of organizational citizenship behaviors are alsosuspect in terms of socially desirable responding. In contrast, socialdesirability should be less of a concern for variables in which respon-dents are asked to report objective information about themselves (i.e.,demographics) or to provide ratings regarding aspects of their orga-nizational context (i.e., organizational justice, conflict in the work-place, or organizational constraints employees encounter). Therefore,to the degree that self-report CWB is contaminated by socially desir-able responding, self- and other-report CWB can be expected to haveless similar patterns of relationships with common correlates affectedby social desirability (i.e., Conscientiousness, Agreeableness, Emo-tional Stability, job satisfaction, negative affect, and organizationalcitizenship behavior), and more similar patterns of relationships withcommon correlates that are less affected by social desirability (i.e.,Openness, Extraversion, demographics, organizational justice, orga-nizational conflict, and organizational constraints; see Table 1 fordefinitions of each of these variables).

Hypothesis 2b: Self- and other-report CWB will have lesssimilar patterns of relationships with theoretically relevantcorrelates affected by socially desirable responding than withtheoretically relevant correlates relatively unaffected by so-cially desirable responding.

Do Self- and Other-Raters Report Similar Amounts ofCWB for Employees?

This represents another line of evidence for the usefulness ofother-ratings of CWB. As stated before, if self- and other-report CWBeach capture unique variance (Hypothesis 1), it is still only worthgoing to the trouble to collect other-ratings if the unique variancecaptured by other-ratings reflects useful, incremental information. Ifself- and other-report CWB reflect different information, then it isfurther expected that each source would report different meanamounts of CWB for employees. As was discussed above, there are anumber of reasons to expect self- and other-raters to provide dissim-ilar mean ratings, including rating source differences in the opportu-nity to observe CWB, other-raters’ over and/or underreporting of anemployee’s CWB, as well as self-raters’ underreporting of their ownCWB. These issues are likely to influence self- and other-ratingsdifferently, so it is not clear whether one source will systematicallyreport more or less CWB than the other. Therefore, we investigatedthe following research question:

Research Question 1: Is there a difference in the mean amountof CWB reported by self- and other-raters?

The same moderator analyses that were carried out for the corre-lation analyses were also carried out for the mean differences analy-ses. These mean difference moderator hypotheses were based on thesame logic as the correlation moderator hypotheses. Specifically:


Table 1Definitions for, and Measures of, Each Variable Meta-Analyzed

Variable Definition Measures included in the meta-analysis

Counterproductive work behavior (CWB)CWB Voluntary employee behaviors that are viewed by the

organization as contrary to its legitimate interests,violate significant organizational norms, andthreaten the well-being of the organization or itsmembers. Also includes interpersonal- (e.g.,violence, gossip) and organizational target(working slowly, damaging company property)dimensions.

Antisocial behavior (Robinson &O’Leary-Kelly, 1998);anticitizenship behavior (Ball et al.,1994); Counterproductive WorkBehavior Checklist (Fox et al., 2001;Spector et al., 2006); workplacedeviance (Aquino et al., 1999);Workplace Deviance Scale (Bennett& Robinson, 2000)

Big FiveEmotional Stability The personality trait reflecting the degree to which a

person is secure, is calm, has low anxiety, and haslow emotionality.

Big Five Inventory (John et al., 1991);International Personality Item Pool(Goldberg, 1999); PersonalCharacteristics Inventory (Barrick &Mount, 1999)

Extraversion The personality trait reflecting the degree to which aperson is sociable, assertive, talkative, ambitious,and energetic.

Measured using same inventories asEmotional Stability

Openness to Experience The personality trait reflecting the degree to which aperson is curious, intelligent, imaginative, andindependent.


Agreeableness The personality trait reflecting the degree to which aperson is likable, easy to get along with, andfriendly.


Conscientiousness The personality trait reflecting the degree to which aperson is hard working, dependable, and detailoriented.


Organizational citizenship behavior (OCB)OCB Work behaviors that support the broader

organizational, social, and psychologicalenvironment. Also includes interpersonal- andorganizational target dimensions.

Lee & Allen (2002); Podsakoff et al.(1990); L. J. Williams & Anderson(1991)

Organizational justiceDistributive justice The degree to which an employee feels the allocation

of outcomes was fair.Colquitt (2001); Price & Mueller

(1986)Interactional justice The degree to which an employee feels he or she has

been treated sensitively and is respected.Colquitt (2001); Niehoff & Moorman

(1993)Procedural justice Degree to which an employee feels the process by

which outcomes are distributed was fair.Colquitt (2001); Moorman (1991);

Niehoff & Moorman (1993)Affective variables

Job satisfaction Pleasurable emotional state resulting from theappraisal of one’s job as achieving or facilitatingthe achievement of one’s job values.

Brayfield & Rothe (1951); Cammannet al. (1979)

Negative affect Negative emotions or subjectively experiencednegative feelings.

Job-related Affective Well Being Scale(Van Katwyk et al., 2000); Positiveand Negative Affect Schedule(Watson et al., 1988)

Contextual variablesConflict Discrepant views or perceived incompatibilities

between individuals.Interpersonal Conflict at Work Scale

(Spector & Jex, 1998)Constraints Situations/things at work that prevent employees

from performing.Organizational Constraints Scale

(Spector & Jex, 1998)Demographics

Age Age (in years) of the participantGender 0 _ female, 1 _ maleTenure The number of years an employee has been

employed by an organization.


Hypothesis 3a: There will be a larger mean difference inCWB ratings between supervisors and employees than be-tween coworkers and employees.

Hypothesis 3b: There will be a larger mean difference be-tween self- and other-ratings of organizational target CWBthan interpersonal target CWB.

Hypothesis 3c: The mean difference between self- and other-ratings of CWB will decrease as the assurances of partici-pants’ anonymity increase.

Method

Search for Primary Data

A fivefold approach was used to identify studies containing usefulinformation for the present meta-analysis. First, keyword searches ofthe PsycINFO and ABI/Inform databases were performed using 29different keywords (see Table A1 in the Appendix for the searchterms). Second, the Web of Science was used to identify any articlesthat cited Bennett and Robinson (2000). Third, manual searches of the2005–2010 programs for Society for Industrial and OrganizationalPsychology and Academy of Management conferences were carriedout. Fourth, the reference sections of relevant meta-analyses (Berry,Ones, & Sackett, 2007; Dalal, 2005; Hershcovis et al., 2007; Lau, Au,& Ho, 2003; Salgado, 2002) were examined for useful citations. Fifth,prominent CWB researchers were contacted with requests for anyunpublished research.

Studies that contained enough information to extract either (a) acorrelation between self-report CWB and other-report CWB (i.e.,supervisor or coworker reports of CWB) or (b) a correlation betweenother-report CWB and some other variable were initially included.Only relationships based on at least three independent samples weremeta-analyzed (this was the same decision rule used in Berry, Ones,& Sackett, 2007, and Cohen-Charash & Spector, 2001). This resultedin a final database of 40 studies from which 50 independent samplescontaining 224 independent correlations were drawn (see Table A2 inthe Appendix for data regarding the primary studies). Twenty-onevariables were identified that met the inclusion criteria (see Table 1for a listing and definitions of each of these variables, along with themeasures of each variable that were included in the meta-analysis). Ofthese 21 variables, 18 were correlates of CWB that will hereafter becollectively referred to as the CWB correlates.

Coding of Study Characteristics

For each sample, the correlation between (a) self- and other-reports of CWB and/or (b) the correlation between other-reports ofCWB and CWB correlates was coded. In cases in which relation-ships with the overall CWB construct were of interest but multiplefacets of the overall construct were offered within the same sample(e.g., both interpersonal- and organizational target CWB measures,which are multiple facets of overall CWB, listed for the samesample), composite formulas (Ghiselli, Campbell, & Zedeck,1981, pp. 163–164) were used to estimate correlations with acomposite of the multiple measures. The mean and standard de-viation of CWB reported by self- and other-raters was also codedfor use in calculating standardized mean differences in CWB

between self- and other-ratings. It should be noted that an explicitsearch for studies reporting mean differences between self- andother-reported CWB was not carried out; instead, mean differenceswere coded in the correlational studies that provided relevantmeans and standard deviations.

Two potential moderating variables were coded. First, thesource of the other-rating (either supervisor or coworker) wascoded as a potential moderator variable. Second, the proceduresused by primary study researchers to assure participants of theiranonymity were coded. In no studies were participant responsescompletely anonymous; at the very least, there must be identifiersto link self-reports of CWB to other-reports of CWB. However,three procedures were commonly cited in primary studies as meth-ods for assuring participants of anonymity of responses: (a) Par-ticipants mailed completed paper-and-pencil surveys directly tothe researchers (as opposed to other practices with paper-and-pencil surveys such as having supervisors and employees simul-taneously fill out and return the surveys during the workday), (b)self-raters generated a secret code that was provided to the other-rater, and/or (c) ratings were provided online.2 Some primarystudies did not report using any of these “anonymity safeguards,”whereas some studies reported using as many as two. Thus, pri-mary studies were sorted into three different categories represent-ing the number of these three practices that were reported ashaving been used in the study (0, 1, or 2), based on the idea thatusing more of these practices would make participants feel theirresponses were more confidential. Efforts were also made to codefor a number of other potential moderators (e.g., opportunity ratershad to observe employees engaging in CWB, whether self- andother-raters were rating the same set of CWBs, whether Likert vs.frequency CWB scales were used, sample characteristics such asage and tenure), but primary studies either did not report relevantinformation often enough, or there was no variability across pri-mary studies on the moderator variable.

The second and third authors each independently coded five arti-cles containing 82 correlations and 740 pieces of coding informationand then met to quantify levels of intercoder agreement. The secondand third authors agreed 100% on all variables. After this point, thesecond and third authors divided the rest of the articles for coding.

Analyses

Hunter and Schmidt’s (2004) meta-analysis methods were usedfor meta-analyses of correlations and d values. Corrections for twostatistical artifacts were made. First, point-biserial correlationsinvolving the variable “gender” were individually corrected towhat they would be if each sample had a 50–50 gender split (seeTable A2 for gender splits for samples). Second, correlations and

2 Of course, providing ratings online does not assure participants’ ano-nymity. More important is whether completing measures online enhancesparticipants’ perceptions that their responses are anonymous, and there isevidence that this is the case. For instance, Richman, Kiesler, Weisband,and Drasgow’s (1999) meta-analysis demonstrated that social desirabilitydistortion was reduced when measures were completed using a computercompared with a paper-and-pencil format as well as to face-to-face inter-views. This suggests that completing surveys on a computer may leadrespondents to feel more comfortable providing information that may be ofa sensitive nature or socially undesirable.


d values were corrected for unreliability in both the predictor andthe criterion3 using the artifact distribution method (see Table 2 forpredictor and criterion reliability artifact distributions and theirsources).

For moderator analyses, studies were sorted into different catego-ries on the basis of study characteristics (e.g., whether other-ratingswere from supervisors vs. coworkers), and meta-analyses were carriedout within each moderator category. To determine whether correla-tions or d values (which were converted to point-biserial correlationsfor the significance analyses) differed significantly across moderatorcategories, formulas from Raju and Brand (2003; specifically, For-mulas 9 and 14) were used to test the significance of the differencebetween the corrected meta-analytic correlations.

For the meta-analyses involving correlations between other-reports of CWB and (a) organizational citizenship behavior, (b)organizational constraints, and (c) interpersonal conflict, it wascommon for the other-raters to rate both variables involved in thecorrelation (e.g., supervisor raters rating both employees’ CWBand organizational citizenship behavior [OCB]). Thus, commonmethod bias was a concern for such samples. Therefore, for theabove relationships, analyses were run two ways: (a) including thesamples in which other-raters rated both variables and (b) exclud-ing such common-source samples. Meta-analytic correlations weremuch higher when such common-source samples were included.So, throughout the rest of the present article, the focus is directedtoward analyses excluding such common-source samples; al-though, results with and without such samples are reported in thetables included in the present article.

Results

Relationship Between Self- and Other-Ratings of CWB

Hypothesis 1 suggested that self- and other-ratings of CWBwould be positively correlated, but less than unity. Table 3 listsresults for the relationship between self-ratings and other-ratingsof CWB. The average corrected self-other CWB correlation was.38. Both the confidence and credibility intervals were well abovezero and did not overlap with 1.0. This moderate positive corre-lation supports Hypothesis 1.

For the following moderator analyses (Hypotheses 1a–1c), thenumber of samples in some categories is relatively small, sosecond-order sampling error may affect results. Hypothesis 1asuggested that self-ratings of CWB would be more strongly cor-related with coworker ratings of CWB than supervisor ratings ofCWB. Table 3 also lists the corrected correlations between self-and supervisor ratings of CWB (.37) and between self- and co-worker ratings of CWB (.40). These correlations did not signifi-cantly differ (z � �.82), and therefore, Hypothesis 1a was notsupported.

Hypothesis 1b suggested that self- and other-ratings of interper-sonal target CWB (CWB-I) would correlate more strongly thanself- and other-ratings of organizational target CWB (CWB-O).Table 3 lists meta-analytic correlations between self- and other-ratings of CWB-I and CWB-O. The corrected correlation betweenself- and other-ratings of CWB-I was .51, whereas the correctedcorrelation between self- and other-ratings of CWB-O was .35.These correlations differed significantly (z � 4.35), providingsupport for Hypothesis 1b.4 Hypothesis 1c suggested that self- and

other-ratings of CWB would correlate more strongly as more stepswere taken by researchers to assure participants of their anonym-ity. All but one sample included at least one coded “anonymitysafeguard,” so analyses could only be carried out for samplesincluding one versus two anonymity safeguards. The self–otherCWB correlation was significantly greater in the two safeguardssamples (.44) than in the one safeguard samples (.28; z � �3.29).Thus, Hypothesis 1c was supported.

Comparison of the Relationships That Self-ReportCWB and Other-Report CWB Have With TheirCommon CWB Correlates

Table 4 lists full meta-analytic results for the correlations be-tween other-ratings of CWB and the set of CWB correlates.5

Table 5 provides a side-by-side comparison of just the correlations

3 The criterion in the present meta-analysis was other-ratings of CWB,which involved judgmental ratings about employees’ CWB from supervi-sors or coworkers. There has been debate about whether interrater reliabil-ity versus alpha reliability should be used for correcting correlations insuch instances (Schmidt, Viswesvaran, & Ones, 2000, argue that interraterreliability is most appropriate; other researchers have suggested that inter-rater reliabilities underestimate reliability; Murphy & DeShon, 2000).Therefore, the present meta-analysis offers separate estimates corrected forinterrater reliability versus alpha reliability in the dependent variable forcomparison purposes. One of the main purposes of the present meta-analysis was to compare the present study’s other-report CWB results withresults from previous meta-analyses using self-reports (which were cor-rected using alpha coefficients). So, for the sake of comparability with theprevious meta-analyses in which alpha coefficients were used in CWBcorrections, the results in our tables are based on the correlations correctedusing alpha coefficients, and we mostly focus our discussion on thoseestimates. However, within each table of results, there is a column listingthe correlation corrected for interrater reliability in the dependent variable.Comparing the alpha- versus interrater-corrected estimates shows that thepattern of results is exactly the same, but the magnitude of correlations isslightly higher for the interrater-corrected estimates.

4 We note that the confidence intervals and credibility intervals for theself–other CWB-I and CWB-O correlations overlap (this is also the casewith the other statistically significant moderator analyses). This suggeststhat, although the difference in the mean correlations is likely greater thanzero, there is still some noteworthy overlap of the sampling error distri-butions for each individual corrected mean correlation (as captured by theconfidence intervals) and the distributions of corrected mean correlations(as captured by the credibility intervals).

5 Attempts were also made to carry out moderator analyses testingwhether the relationships that other-report CWB had with the commoncorrelates differed depending on (a) whether other-ratings were fromsupervisors versus coworkers, (b) whether the other-ratings were of CWB-Iversus CWB-O, or (c) number of anonymity safeguards. In most cases,there were not enough samples in each moderator category for meaningfulmoderator analyses (i.e., there were fewer than three samples). There wereenough samples to carry out moderator analyses of the relationshipsother-report CWB had with job satisfaction, age, gender, and tenure;however, there was not enough variance in the results reported in Table 4for meaningful moderator analyses for job satisfaction and age. So, mod-erator analyses were carried out only for gender and tenure, but neithermoderating variable moderated the relationships between other-reportCWB and gender and tenure. Therefore, there is no evidence that therelationships between other-report CWB and the common correlates aremoderated by the variables included in the present study.


that self-report CWB and other-report CWB have with their com-mon CWB correlates. The other-report-corrected correlations areeach taken from Table 4 in the present meta-analysis; alpha-corrected correlations appear outside and interrater-corrected cor-relations appear inside parentheses. The self-report-corrected cor-relations are taken from three previous meta-analyses (Berry,Ones, & Sackett, 2007; Dalal, 2005; Hershcovis et al., 2007) inwhich the same relationships were examined, but relied almostexclusively on self-reports of CWB. There are two trends in Table5 that are of importance: (a) the degree to which the magnitude ofrelationships differs depending on the source of CWB ratings (thisdeals most directly with Hypothesis 2a) and (b) the degree towhich the pattern of relationships differs depending on the sourceof CWB ratings (most directly relevant to Hypothesis 2b).

Regarding the magnitude of relationships, self-report CWBtended to be slightly more related to CWB correlates than other-report CWB. The last column in Table 5 lists the differencebetween the self-report CWB and other-report CWB correlations(numbers outside parentheses used alpha-corrected other-reportCWB for this calculation, numbers inside parentheses usedinterrater-corrected other-report CWB), with positive numbersmeaning the self-report correlation is stronger. In 13 out of 18

relationships, the self-report correlation was stronger than theother-report correlation. In some instances, the correlation withself-report CWB was quite a bit stronger than the correlation withother-report CWB. For instance, organizational target OCB(OCB-O) was correlated �.44 and �.13 with self- and other-report CWB, respectively; job satisfaction was correlated �.37and �.21 with self- and other-report CWB, respectively. However,there were also instances in which the other-report correlation wasquite a bit stronger than the self-report correlation (e.g., interper-sonal target OCB [OCB-I], interactional justice). Furthermore, theoverall magnitude of relationships did not differ especiallystrongly between self-report and other-report CWB, with the av-erage difference across all common correlates being .07 (self-report stronger) when using the alpha-corrected other-report cor-relations. When the interrater-corrected correlations are used, thisdifference reduces to only .04. Thus, there is, at best, weak andsporadic support for Hypothesis 2a that self-ratings of CWB willcorrelate more strongly than other-ratings of CWB with a commonset of theoretically relevant correlates, suggesting common methodbias is not a large systematic concern.

Regarding the pattern of relationships, Hypothesis 2b suggestedthat self- and other-report CWB may show more similar relation-

Table 2Alpha and Interrater Reliability Artifact Distributions

Variable Type rxx SD N k Source

CWBCWB (self-rating) Alpha .77 — 16,721 49 Dalal (2005)CWB-I (self-rating) Alpha .84 0.07 6,878 26 Berry, Ones, & Sackett (2007)CWB-O (self-rating) Alpha .82 0.07 6,080 22 Berry, Ones, & Sackett (2007)CWB (other-rating) Alpha .91 0.05 3,411 21 Present studyCWB-I (other-rating) Alpha .88 0.07 4,594 26 Present studyCWB-O (other-rating) Alpha .89 0.06 4,958 28 Present studyCWB (other-rating) Interrater .665 — 1,125 13 Viswesvaran et al. (1996)CWB-I (other-rating) Interrater .665 — 1,125 13 Viswesvaran et al. (1996)CWB-O (other-rating) Interrater .665 — 1,125 13 Viswesvaran et al. (1996)

Big FiveEmotional Stability Alpha .78 0.11 — 370 Viswesvaran & Ones (2000)Extraversion Alpha .78 0.09 — 307 Viswesvaran & Ones (2000)Openness to Experience Alpha .73 0.12 — 251 Viswesvaran & Ones (2000)Agreeableness Alpha .75 0.11 — 123 Viswesvaran & Ones (2000)Conscientiousness Alpha .78 0.10 — 307 Viswesvaran & Ones (2000)

OCBOverall OCB Alpha .79 — 16,455 47 Dalal (2005)OCB-I Alpha .73 — 5,864 24 Dalal (2005)OCB-O Alpha .74 — 5,607 23 Dalal (2005)

Organizational justiceDistributive justice Alpha .91 — — 66 Hauenstein et al. (2001)Interactional justice Alpha .90 0.06 668 3 Berry, Ones, & Sackett (2007)Procedural justice Alpha .91 — — 66 Hauenstrein et al. (2001)

Affective variablesJob satisfaction Alpha .84 0.06 1,415 7 Connolly & Viswesvaran (2000)Negative affect Alpha .79 0.11 3,326 15 Connolly & Viswesvaran (2000)

Contextual variablesConflict Alpha .74 — 3,363 13 Spector & Jex (1998)Constraints Alpha .85 — 1,746 8 Spector & Jex (1998)

Note. The mean and standard deviation of interrater reliabilities drawn from Viswesvaran et al. (1996) are averages of Viswesvaran et al.’s supervisorand peer interrater reliabilities. Type � type of reliability coefficient used in the artifact distribution; rxx � reliability artifact distribution mean; SD �reliability artifact distribution standard deviation; N � reliability artifact distribution sample size; k � number of samples contributing to artifactdistributions; Source � study from which artifact distributions were drawn; CWB � counterproductive work behavior; CWB-I � interpersonal target CWB;CWB-O � organizational target CWB; OCB � organizational citizenship behavior; OCB-I � interpersonal target OCB; OCB-O � organizational targetOCB. Dashes represent instances in which that piece of information was unavailable.


ships with common correlates that are less affected by socialdesirability (i.e., Extraversion, Openness, demographics, organiza-tional justice, organizational conflict, and organizational con-straints). The average difference in the last column in Table 5(using alpha-corrected correlations for other-report CWB) for vari-ables relatively unaffected by social desirability was .02, whereasit was .13 for the variables that were relatively affected by socialdesirability (the average differences were �.01 and .10, respec-tively, when using interrater-corrected correlations). This providessome support for Hypothesis 2b. However, even differences incorrelations of around .10–.13 are not especially large (e.g., Berry,Ones, & Sackett, 2007). Furthermore, it is worth noting that theoverall pattern of relationships that self-report CWB and other-report CWB have with the CWB correlates is very similar. Forinstance, the vector of other-report CWB correlations was corre-lated .87 with the vector of self-report CWB correlations in Table5 (these correlations were .93 and .77 for the variables relativelyunaffected and affected by social desirability, respectively). Thesecorrelations are the same, regardless of whether alpha-corrected orinterrater-corrected correlations are used. Thus, although self- andother-reports of CWB do have less similar patterns of relationshipswith their common correlates when the common correlates arerelatively affected by social desirability, and the overall patterns ofrelationships are very similar for self- and other-report CWB.

Mean Differences Between Self- andOther-Ratings of CWB

Mean difference results are presented in Table 6. Positive dvalues in Table 6 indicate self-raters reported more CWB. Re-search Question 1 asked whether there was a difference in themean amount of CWB reported by self- and other-raters. Thecorrected mean difference was d � 0.35, and the confidenceinterval did not overlap with zero, suggesting that self-raters reportmore CWB than other-raters.

For the following moderator analyses (Hypotheses 1a–1c), thenumber of samples in some categories is relatively small, sosecond-order sampling error may affect results. Hypothesis 3asuggested there will be a larger mean difference in CWB ratingsbetween supervisors and employees than between coworkers andemployees. There was a significantly larger mean difference be-tween supervisors and employees (d � 0.44) than between co-workers and employees (d � 0.23; z � 2.17), providing support forHypothesis 3a.

Hypothesis 3b suggested there will be a larger mean differencebetween self- and other-ratings of CWB-O than of CWB-I. Therewas a significantly larger mean difference between self- and other-ratings of CWB-O (d � 0.46) than of CWB-I (d � 0.18; z ��2.80), providing support for Hypothesis 3b.

Hypothesis 3c suggested that the mean difference between self-and other-ratings of CWB will decrease as the assurances ofparticipants’ anonymity increase. The self–other CWB mean dif-ference was significantly larger in samples with one anonymitysafeguard (d � 0.49) than in samples with two anonymity safe-guards (d � 0.26; z � 2.08), providing support for Hypothesis 3c.

Is the Unique Variance Captured by Self- andOther-Report CWB Valid?

The moderate correlation between self- and other-reportCWB (.38) suggests that each source captures unique variance.What is not clear is whether the unique variance captured byself- and other-reports represents valid, unique variance (i.e.,true CWB variance). One relevant line of evidence is thecorrelations that self- and other-report CWB have with thecommon correlates once variance attributable to the othersource has been partialed out (e.g., what is the partial correla-tion between self-report CWB and a given common correlatewhen other-report variance has been partialed out of bothself-report CWB and the common correlate?). If other-report

Table 3Meta-Analysis Results for the Relationship Between Self-Report and Other-Report Counterproductive Work Behavior

Variable N k rm SDr r� SDr� r�,irr % var. CV10 CV90 CIL CIU z

Overall analysisSelf–other 3,503 21 .32 .16 .38 .18 .44 18.51 .16 .61 .30 .47

Moderator analysesSupervisor vs. coworker ratings

Self-supervisor 2,044 11 .31 .19 .37 .21 .43 12.61 .10 .64 .23 .50 �0.82Self-coworker 1,459 10 .33 .12 .40 .11 .47 40.74 .26 .54 .31 .49

Interpersonal vs. organizational target CWBSelf CWB-Other CWBI 1,500 9 .44 .14 .51 .14 .58 20.63 .32 .69 .40 .61 4.35�

Self CWBO-Other CWBO 1,500 9 .29 .17 .35 .18 .40 18.74 .12 .58 .22 .48Number of anonymity safeguards

One 1,128 6 .23 .08 .28 .06 .32 66.93 .20 .35 .20 .36 �3.29�

Two 868 7 .37 .09 .44 .04 .51 82.26 .38 .50 .36 .52

Note. rm � mean sample size-weighted correlation; SDr � sample size-weighted observed standard deviation of correlations; r� � mean samplesize-weighted correlation corrected for unreliability using alphas; SDr� � corrected standard deviation of corrected correlations; r�,irr � mean samplesize-weighted correlation corrected for unreliability using interrater reliabilities for other-rated variables and alphas for self-rated variables; % var. �percentage of variance attributable to statistical artifacts; CV10 and CV90 � 10% and 90% credibility values, respectively; CIL and CIU � lower and upperbounds, respectively, of the 95% confidence interval around the corrected mean correlation; z � the z statistic calculated using Formula 14 (and usingFormula 9 to calculate sampling variance) from Raju and Brand (2003) for assessing the significance of the difference between the corrected correlationswithin each moderator category (zs � �1.96 suggest a significant difference). Self CWBI-Other CWBI � correlation between self-ratings and other-ratingsof CWB-Interpersonal; Self CWBO-Other CWBO � correlation between self-ratings and other-ratings of CWB-Organizational.� p � .05.


CWB no longer correlates with its common correlates oncevariance from self-reports has been partialed out, then thiswould suggest that other-reports tell us nothing about employ-ees’ CWB that self-reports do not already tell us. The samewould be true of self-report CWB’s partial correlations.

Table 7 lists bivariate and partial correlations that self-reportand other-report CWB have with the common correlates. Variancebetween self- and other-report correlations was partialed out usingthe meta-analytic correlation of .38 between self- and other-reportCWB. Asterisks denote correlations with confidence intervals thatdo not overlap with zero. Fifteen of 18 partial correlations forself-report CWB and 10 of 18 partial correlations for other-reportCWB had confidence intervals that did not overlap with zero,suggesting that both self- and other-reports contain at least somevalid unique variance.

However, many of the other-report CWB partial correlations,despite having confidence intervals that do not overlap withzero, are relatively small (around .10 or lower). In order toassess the value that is added by measuring CWB via both self-and other-ratings (i.e., multisource ratings) of CWB instead ofjust via self-ratings of CWB, each of the common correlateswas hierarchically regressed on self-report CWB, and then onself-report and other-report CWB in a second step. The increase

in R2 from adding other-reports indicates the value of addingother-reports of CWB (see last column of Table 7). The in-creases in R2 were generally very small, only exceeding .02 inthree instances (this pattern was the same if R, instead of R2 isused; increases in R ranged from .00 to .17, with only threeincreases being greater than .03).6 The three instances in whichother-report CWB accounted for more than a .02 increase in R2

are noteworthy in that they involved the three most interper-sonal CWB correlates: OCB-I, interactional justice, and con-flict. In all, this suggests that the valid, unique variance addedby other-report CWB over self-report CWB is generally verysmall, except perhaps when the CWB correlate is especiallyinterpersonal in nature.

6 All of the analyses in Table 7 were also carried out (a) using observed,uncorrected correlations and (b) using interrater reliability-corrected cor-relations for any correlations involving other-report CWB. Results werevirtually identical and no conclusions changed, so the observed correlationand interrater reliability-corrected correlation results are not listed. Theseresults are available upon request from the first author.

Table 4Meta-Analysis Results for the Relationship Between Other-Report Counterproductive Work Behavior and the CWB Correlates

Variable N k rm SDr r� SDr� rirr,� % var. CV10 CV90 CIL CIU

Big FiveEmotional Stability 2,975 12 �.04 .09 �.05 .07 �.06 53.09 �.14 .04 �.11 .01Extraversion 1,066 7 .03 .13 .04 .13 .04 36.35 �.13 .20 �.08 .16Openness to Experience 890 6 �.10 .06 �.13 .00 �.15 100 �.13 �.13 �.19 �.07Agreeableness 2,246 9 �.18 .16 �.22 .18 �.25 14.94 �.44 .01 �.34 �.09Conscientiousness 3,332 13 �.15 .13 �.18 .14 �.21 22.03 �.36 �.01 �.27 �.10

OCBOverall OCB 1,509 11 �.47 .17 �.56 .19 �.65 15.41 �.80 �.32 �.68 �.43Overall OCB* 712 5 �.16 .09 �.19 .03 �.23 91.95 �.23 �.16 �.28 �.11OCB-I 2,016 9 �.33 .18 �.41 .21 �.48 10.98 �.68 �.14 �.56 �.27OCB-I* 612 4 �.27 .18 �.34 .20 �.39 17.97 �.59 �.08 �.55 �.12OCB-O 1,065 7 �.46 .13 �.56 .13 �.66 25.85 �.74 �.39 �.68 �.45OCB-O* 612 4 �.11 .09 �.13 .06 �.15 72.10 �.21 �.05 �.24 �.02

Organizational justiceDistributive justice 820 5 �.07 .20 �.07 .21 �.08 14.49 �.34 .19 �.27 .13Interactional justice 505 3 �.40 .21 �.45 .22 �.52 9.36 �.73 �.16 �.71 �.18Procedural justice 1,659 10 �.20 .18 �.22 .18 �.26 16.75 �.46 .01 �.35 �.10

Affective variablesJob satisfaction 2,231 13 �.19 .08 �.21 .04 �.25 82.63 �.26 �.16 �.26 �.16Negative affect 761 5 .16 .08 .19 .01 .23 99.28 .18 .20 .11 .28

Contextual variablesConflict 1,609 12 .39 .14 .48 .15 .56 26.99 .30 .67 .38 .58Conflict* 934 7 .31 .09 .38 .05 .44 75.86 .31 .44 .29 .46Constraints 1,262 9 .27 .13 .31 .12 .36 36.79 .16 .46 .21 .41Constraints* 838 6 .20 .09 .22 .04 .26 85.58 .17 .27 .14 .30

DemographicsAge 2,091 13 �.05 .07 �.06 .00 �.07 100 �.06 �.06 �.10 �.02Gender 2,080 13 �.07 .15 �.08 .13 �.09 27.21 �.25 .10 �.16 .01Tenure 1,552 10 �.06 .11 �.06 .07 �.07 55.63 �.16 .03 �.13 .01

Note. CWB � counterproductive work behavior; rm � mean sample size-weighted correlation; SDr � sample size-weighted observed standard deviationof correlations; r� � mean sample size-weighted correlation corrected for unreliability using alphas; SDr� � corrected standard deviation of correctedcorrelations; rirr,� � mean sample size-weighted correlation corrected for unreliability using interrater reliabilities for other-rated variables and alphas forself-rated variables; % var. � percentage of variance attributable to statistical artifacts; CV10 and CV90 � 10% and 90% credibility values, respectively;CIL and CIU � lower and upper bounds, respectively, of the 95% confidence interval around the corrected mean correlation. Asterisks denote meta-analyseswherein same-source samples were removed (e.g., supervisors rating both employees’ organizational citizenship behavior and CWB).


Discussion

Summary of Findings

The results of the present meta-analysis provided three lines ofevidence regarding the comparability of self- and other-reports ofCWB. First, self- and other-reports of CWB are moderately cor-related (corrected correlations of .38 or .44, depending on whetheralpha vs. interrater reliability is used for other-ratings). The cor-relation between self- and other-ratings becomes somewhat stron-ger (i.e., in the .40–.60 range) when the ratings are of CWB-I andwhen greater assurances of anonymity are provided. This meansthat self- and other-ratings of CWB intercorrelate more stronglythan ratings of job performance from multiple sources (Conway &Huffcutt, 1997). Thus, the general trend is for appreciable overlapbetween self- and other-ratings of CWB. However, the ratingsources clearly each capture a fair amount of unique variance.

This unique variance begs the question of whether self- andother-ratings of CWB have different patterns or magnitudes ofrelationships with their correlates. The second line of evidencepresented in this meta-analysis demonstrated that, with some ex-ceptions, self- and other-ratings of CWB have very similar rela-tionships with their common correlates. The patterns of relation-ships that self- and other-report CWB have with their commoncorrelates are almost identical, and self-reports of CWB correlatedwith their common correlates only slightly higher (between about

.04 and .07), on average, than other-reports. There are somenoteworthy exceptions, such as the stronger relationships thatself-report CWB has with Conscientiousness, Agreeableness, andEmotional Stability; the stronger relationships that other-reportCWB has with the interpersonal correlates of OCB-I, interactionaljustice, and conflict; and the pattern of greater differences inrelationships that self- and other-report CWB have with commoncorrelates more affected by social desirability. However, evenwhen both self-report CWB and the CWB correlates were likely tobe affected by social desirability, self-report CWB only correlatedwith the common correlates about .10–.13 higher than did other-report CWB, which is a relatively small difference (e.g., Berry,Ones, & Sackett, 2007; Cohen, 1992). Additionally, the correlationof .87 between the vectors of self- and other-report correlationswith the common correlates demonstrates that the general patternis for the nomological networks of the two sources to be verysimilar. Furthermore, the partial correlation analyses demonstratedthat for most of the CWB correlates, other-reports of CWB accountfor very little additional variance beyond self-reports of CWB.

The third line of evidence regarding the comparability of self-and other-reports of CWB presented in this meta-analysis was theaverage mean difference between the amounts of CWB reportedby the two sources. Across all conditions, self-raters reportedengaging in more CWB than other-raters reported them engagingin. Thus, the idea that self-raters will underreport their engagement

Table 5Side-by-Side Comparison of the Relationships That Other-Report CWB (Present Meta-Analysis) and Self-Report CWB(Previous Meta-Analyses) Have With a Common Set of CWB Correlates

VariableCorrelation with

other-report CWBCorrelation withself-report CWB

Source of self-report CWBcorrelation Differencea

Big FiveEmotional Stability �.05 (�.06)b �.23 Berry, Ones, & Sackett (2007) .18 (.17)Extraversion .04 (.04) �.03 Berry, Ones, & Sackett (2007) �.07 (�.07)Openness to Experience �.13 (�.15) �.06 Berry, Ones, & Sackett (2007) �.07 (�.09)Agreeableness �.22 (�.25) �.35 Berry, Ones, & Sackett (2007) .13 (.10)Conscientiousness �.18 (�.21) �.31 Berry, Ones, & Sackett (2007) .13 (.10)

OCBOverall OCB* �.19 (�.23) �.36 Berry, Ones, & Sackett (2007) .17 (.13)OCB-I* �.34 (�.39) �.21 Berry, Ones, & Sackett (2007) �.13 (�.18)OCB-O* �.13 (�.15) �.44 Berry, Ones, & Sackett (2007) .31 (.29)

Organizational justiceDistributive justice �.07 (�.08) �.14 Berry, Ones, & Sackett (2007) .07 (.06)Interactional justice �.45 (�.52) �.30 Berry, Ones, & Sackett (2007) �.15 (�.22)Procedural justice �.22 (�.26) �.27 Berry, Ones, & Sackett (2007) .05 (.01)

Affective variablesJob satisfaction �.21 (�.25) �.37 Dalal (2005) .16 (.12)Negative affect .19 (.23) .29 Hershcovis et al. (2007) .10 (.06)

Contextual variablesConflict* .38 (.44) .46 Hershcovis et al. (2007) .08 (.02)Constraints* .22 (.26) .33 Hershcovis et al. (2007) .11 (.07)

DemographicsAge �.06 (�.07) �.13 Berry, Ones, & Sackett (2007) .07 (.06)Gender �.08 (�.09) �.17 Berry, Ones, & Sackett (2007) .09 (.08)Tenure �.06 (�.07) �.05 Berry, Ones, & Sackett (2007) �.01 (�.02)

Note. CWB � counterproductive work behavior; OCB � organizational citizenship behavior; OCB-I � interpersonal target OCB; OCB-O �organizational target OCB. Asterisks denote common-source samples (e.g., supervisor rated both OCB and CWB) removed.a Positive numbers mean the correlation with self-report CWB was stronger than the correlation with other-report CWB. b Numbers outside parenthesesreflect correlations corrected using alpha coefficients for other-report CWB; numbers inside parentheses reflect correlations corrected using interraterreliabilities for other-report CWB.


in CWB was not supported in the present meta-analysis—at leastnot in comparison to other-raters. However, the common concernthat other-raters do not have adequate opportunity to observeemployees engaging in CWB (and especially CWB-O), and there-fore underreport employees’ CWB, was supported by the meandifference results of the present meta-analysis.

A number of hypotheses regarding moderating variables werealso tested. With the exception of Hypothesis 1a, each of thehypotheses was supported. The identified moderating variables donot change general conclusions of the meta-analysis (e.g., althoughthere was a larger mean difference between self- and supervisorratings than between self- and coworker ratings, this does not

Table 6Mean Differences Between Self- and Other-Ratings of Employees’ CWB

Variable N k dm SDd �� SD� ��,irr % var. CV10 CV90 CIL CIU z

Overall analysisSelf–oOther 2,574 17 .29 .25 .35 .22 .41 43.3 .07 .64 .21 .49

Moderator analysesSupervisor vs. coworker ratings

Self-supervisor 1,458 9 .37 .27 .44 .26 .52 35.5 .12 .77 .23 .65 2.17�

Self-coworker 1,116 8 .20 .18 .23 .07 .27 88.7 .14 .33 .08 .38Interpersonal vs. organizational target CWB

Self CWBI-Other CWBI 1,145 7 .15 .24 .18 .22 .24 41.1 �.10 .46 �.03 .38 �2.80�

Self CWBO-Other CWBO 1,145 7 .40 .45 .46 .49 .54 12.6 �.16 1.09 .08 .83Number of anonymity safeguards

One 1,128 6 .41 .25 .49 .24 .58 35.2 .19 .80 .26 .73 2.08�

Two 880 7 .22 .25 .26 .21 .30 50.7 �.01 .53 .04 .48

Note. Positive d values indicate self-report means were higher; dm � mean sample size-weighted d value; SDd � sample size-weighted observed standarddeviation of d values; �� mean sample size-weighted d value corrected for measurement error using alpha reliability; SD� � corrected standard deviationof alpha-corrected d values; ��,irr � mean sample size-weighted d value corrected for measurement error using alpha reliability for self-ratings and interraterreliability for other-ratings; % var. � percentage of variance attributable to artifacts; CV10 and CV90 � 10% and 90% credibility values, respectively; CIL

and CIU � lower and upper bounds, respectively, of the 95% confidence interval around the corrected mean d value; z � the z statistic calculated usingFormula 14 (and using Formula 9 to calculate sampling variance) from Raju and Brand (2003) for assessing the significance of the difference between thecorrected correlations (d values were converted to point-biserial correlations) within each moderator category (zs � �1.96 suggest a significant difference).Self CWBI-Other CWBI � mean difference between self-ratings and other-ratings of CWB-Interpersonal; Self CWBO-Other CWBO � mean differencebetween self-ratings and other-ratings of CWB-Organizational.� p � .05.

Table 7Bivariate and Partial Correlations Between Self- and Other-Report CWB and the CWB Correlates

CWB correlate Bivariate-other Partial-other Bivariate-self Partial-self �R2

Emotional Stability �.05 .04 �.23 �.25* .00*

Extraversion .04 .06 �.03 �.05 .00Openness to Experience �.13* �.13* �.06 �.01 .01*

Agreeableness �.22* �.10* �.35* �.31* .01*

Conscientiousness �.18* �.07* �.31* �.28* .01*

Overall OCB �.19* �.06 �.36* �.34* .00OCB-I �.34* �.30* �.21* �.09* .08*

OCB-O �.13* .04 �.44* �.46* .00Distributive justice �.07 �.02 �.14 �.13* .00Interactional justice �.45* �.39* �.30* �.15* .13*

Procedural justice �.22* �.14* �.27* �.22* .02*

Job satisfaction �.21* �.08* �.37* �.34* .01*

Negative affect .19* .09* .29* .26* .01*

Conflict .38* .24* .46* .37* .05*

Constraints .22* .11* .33* .29* .01*

Age �.06* �.01 �.13* �.13* .00Gender �.08 �.02 �.17 �.16* .00Tenure �.06 �.05 �.05 �.03 .00

Note. CWB � counterproductive work behavior; Bivariate-other � bivariate correlation between other-report CWB and the CWB correlate; Partial-other � correlation between other-report CWB and a given CWB correlate, with self-report CWB variance partialed out of both variables; Bivariate-self �bivariate correlation between self-report CWB and the CWB correlate; Partial-self � correlation between self-report CWB and a given CWB correlate, withother-report CWB variance partialed out of both variables; �R2 � the increase in the squared multiple correlation when regressing each common correlateon self- and other-ratings of CWB instead of just self-ratings of CWB. Asterisks mean that the confidence interval for the given correlation did not overlapwith zero (using .05 alpha level). The meta-analytic sample sizes on which the correlations were based were used for determining whether the confidenceinterval overlapped with zero.


change the conclusion that other-raters report less CWB thanself-raters). However, they do highlight some important points forfuture CWB research. First, supervisor raters report even lessCWB, relative to self-raters, than do coworker raters, suggestingsupervisors may have even less opportunity to observe employeesengaging in CWB than coworkers. Second, self- and other-ratersagree more about CWB-I than about CWB-O (i.e., ratings are moreintercorrelated, and mean differences are smaller). Third, self- andother-ratings of CWB agree more as the number of anonymitysafeguards increase, suggesting CWB researchers may get betterdata when participants perceive their responses as more anony-mous.

Is it Worth the Trouble of Collecting Other-Ratings orMultisource Ratings of CWB?

It is difficult to collect other-ratings or multisource ratings ofCWB in organizations, especially as compared with collectingself-ratings of CWB. Thus, one might ask whether it is worththe trouble of collecting other-ratings or multisource ratings.The results of the present meta-analysis suggest in many in-stances that there is little tangible value of other-ratings ofCWB (and by extension, multisource ratings) beyond self-ratings. Other-ratings and self-ratings have mostly similar pat-terns and magnitudes of relationships with other variables, andother-ratings often do not account for appreciable incrementalvariance over self-ratings. Furthermore, the mean differencesresults of the present meta-analysis suggest that coworker, andespecially supervisor, raters underreport employees’ CWB rel-ative to the employees themselves, which suggests that the useof other-report CWB entails a focus on a narrower subset ofemployees’ CWB. Especially if researchers take multiple care-ful steps to assure participants of the anonymity of their re-sponses, the value added by collecting other-reports of CWB isnot clear in many instances.

This is not to say that other-reports of CWB do not have value.For instance, the interpersonal CWB correlates OCB-I, interac-tional justice, and conflict were more strongly related to other-report CWB than self-report CWB. This may suggest that super-visor and coworker raters have an especially valid perspective onCWBs that are interpersonal in nature or that have interpersonalantecedents (e.g., breaches of interactional justice, strong interper-sonal conflict in the workplace). This conclusion must be temperedby concerns over second-order sampling error, as the meta-analyticrelationships between other-report CWB and OCB-I, interactionaljustice, and conflict were based on four, three, and seven samples,respectively. However, the stronger correlation between self- andother-reports of the relatively interpersonal CWB-Is also corrob-orates the idea that other-reports may be most useful in assessingrelatively public, interpersonal CWBs. Furthermore, the fact thatpatterns of results are so similar between self- and other-reportssuggests that for many purposes, one source is as good as the other.So, if other-reports are available, then their use should generally beappropriate. Additionally, in some instances, such as test valida-tion, organizations may view self-reports of CWB (or other crite-ria) skeptically. In such instances, other-reports may be a moreviable option. However, the results of the present meta-analysis donot generally support the idea that self-reports are always aninferior method of assessing CWB.

Implications for Theory and Practice

A practical implication of the results is that self- and other-ratings of CWB generally result in very similar patterns of find-ings. Therefore, organizational researchers should feel justified inusing either self- or other-ratings of CWB. Additionally, the find-ings of the present study indicate that supervisor and coworkerratings of CWB are relatively comparable (i.e., whether ratingscame from supervisors vs. coworkers only moderated self–othermean differences, and the mean difference effect was relativelysmall [d � 0.21], although statistically significant). This agreeswith the results of Viswesvaran, Schmidt, and Ones (2002), whofound that supervisor and coworker ratings of “compliance/acceptance of authority” (essentially avoidance of CWB) corre-lated .78. This is information future CWB researchers can draw onwhen deciding which rating sources are most appropriate, giventheir specific research question.

Another practical implication flows from the largely similarpatterns of relationships with correlates in nomological net-works that self- and other-ratings of CWB exhibited. The sci-entific knowledge base to date regarding the causes and conse-quences of CWB that practitioners can draw on has largelydepended on self-report measurement of CWB. The presentmeta-analysis suggests that use of other-report CWB does notdrastically change conclusions regarding the causes and conse-quences of CWB. Thus, practitioners do not presently havecause for concern over the relevance or legitimacy of theself-report research-based CWB interventions they recommendto organizations.

This also highlights perhaps the most important theoreticalimplication of the results of the present meta-analysis. Much ofthe recent empirical evidence supporting theories of CWB hasrelied on self-report measurement of CWB. In particular, muchof what we know about CWB’s nomological network (e.g.,Berry, Ones, & Sackett, 2007; Dalal, 2005; Hershcovis et al.,2007) has been based on self-report CWB measures. The pres-ent meta-analysis found that conclusions regarding the nomo-logical network of CWB remain largely the same when other-reports of CWB are used. This validates much of the cumulativeCWB literature and theory to date that has relied on multi-itemself-report CWB measurement.

Additional Issues, Limitations, and Directions forFuture Research

Other lines of evidence regarding the convergence of self-and other-report CWB that were not available in the presentmeta-analysis are relevant and present opportunities for futureresearch. For instance, an important line of evidence regardingthe similarity/dissimilarity of self- and other-report CWB is thedegree to which they exhibit similar factor structures (e.g.,Dilchert, Ones, Davis, & Rostow, 2007). Thus, measurementinvariance research would be useful. Additionally, the presentmeta-analysis was not able to directly test the degree to whichself- and other-ratings of CWB are accurate (i.e., reflect “trueCWB”) or the degree to which the shared variance between self-and other-ratings of CWB reflects true CWB variance versusshared, systematic error variance. Addressing this questionwould likely require experimental research, although the eco-


logical validity of such experimental CWB research might bedifficult to establish. Also, although the amount of uniquevariance captured by other-ratings was relatively small in mostinstances, it may be of interest to determine the cause of thisunique variance (e.g., what forms/types of CWB are other-raters more privy to observe their employees engaging in?).

It is worth noting that the methods used in the present study areapplicable not only to addressing issues regarding the comparabil-ity and nomological networks of self- and other-report CWB butalso to any constructs in the broader research literature that faceissues regarding the appropriateness of self- versus other-reportmeasurement. For instance, the appropriateness of self- versusother-report measurement is relevant for many other constructsranging from other performance behaviors (e.g., task performance,OCB, job/work withdrawal) to individual-differences measures(e.g., personality, adaptability) to contextual variables (e.g., orga-nizational conflict, organizational constraints, organizational sup-port) and beyond. We used a number of methods in the presentmeta-analysis to determine the relationship between self- andother-reports, the degree to which they have similar nomologicalnetworks, and the degree to which they account for unique versusoverlapping variance. These same methods can, and perhapsshould, be used for investigating similar issues with other variablessuch as those listed above.

References

References marked with an asterisk indicate studies included in themeta-analysis.

Aquino, K., Lewis, M. U., & Bradfield, M. (1999). Justice constructs,negative affectivity, and employee deviance: A proposed model andempirical test. Journal of Organizational Behavior, 20, 1073–1091.

Ball, G. A., Trevino, L. K., & Sims, H. P. (1994). Just and unjustpunishment: Influences on subordinate performance and citizenship.Academy of Management Journal, 37, 299–322.

Barclay, L. J., & Aquino, K. (2010). Workplace aggression and violence.In S. Zedeck (Ed.), APA handbook of industrial and organizationalpsychology (Vol. 3, pp. 615–640). Washington, DC: American Psycho-logical Association.

Barrick, M. R., & Mount, M. K. (1999). The Personal CharacteristicsInventory. Libertyville, IL: Wonderlic.

Bennett, R. J., & Robinson, S. L. (2000). Development of a measure ofworkplace deviance. Journal of Applied Psychology, 85, 349–360. doi:10.1037/0021-9010.85.3.349

Berry, C. M., Ones, D. S., & Sackett, P. R. (2007). Interpersonal deviance,organizational deviance, and their common correlates: A review andmeta-analysis. Journal of Applied Psychology, 92, 410 – 424. doi:10.1037/0021-9010.92.2.410

Berry, C. M., Page, R. C., & Sackett, P. R. (2007). Effects of self-deceptiveenhancement on personality-job performance relationships. Interna-tional Journal of Selection and Assessment, 15, 94–109. doi:10.1111/j.1468-2389.2007.00374.x

*Bing, M., Stewart, S. M., Davison, H. K., Green, P. D., McIntyre, M. D.,& James, L. R. (2007). An integrative typology of personality assess-ment for aggression: Implications for predicting counterproductiveworkplace behavior. Journal of Applied Psychology, 92, 722–744. doi:10.1037/0021-9010.92.3.722

*Bordia, P., Restubog, S. L. D., & Tang, R. L. (2008). When employeesstrike back: Investigating mediating mechanisms between psychologicalcontract breach and workplace deviance. Journal of Applied Psychology,93, 1104–1117. doi:10.1037/0021-9010.93.5.1104

Brayfield, A. H., & Rothe, H. F. (1951). An index of job satisfaction.Journal of Applied Psychology, 35, 307–311.

*Bruk-Lee, V., & Spector, P. E. (2006). The social stressors-counterproductive work behaviors link: Are conflicts with supervisorsand coworkers the same? Journal of Occupational Health Psychology,11, 145–156. doi:10.1037/1076-8998.11.2.145

Budd, J. W., Arvey, R. D., & Lawless, P. (1996). Correlates and conse-quences of workplace violence. Journal of Occupational Health Psy-chology, 1, 197–210. doi:10.1037/1076-8998.1.2.197

Camara, W. J., & Schneider, D. L. (1994). Integrity tests: Facts andunresolved issues. American Psychologist, 49, 112–119.

Cammann, C., Fichman, M., Jenkins, D., & Klesh, J. (1979). The MichiganOrganizational Assessment Questionnaire. Unpublished manuscript,University of Michigan, Ann Arbor.

Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminantvalidity by the multitrait-multimethod matrix. Psychological Bulletin,56, 81–105. doi:10.1037/h0046016

Chen, P. Y., Dai, T., Spector, P. E., & Jex, S. M. (1997). Relationbetween negative affectivity and positive affectivity: Effects ofjudged desirability of scale items and respondents’ social desirability.Journal of Personality Assessment, 69, 183–198. doi:10.1207/s15327752jpa6901_10

*Chiu, S., & Peng, J. (2008). The relationship between psychologicalcontract breach and employee deviance: The moderating role of hostileattributional style. Journal of Vocational Behavior, 73, 426–433. doi:10.1016/j.jvb.2008.08.006

Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159.doi:10.1037/0033-2909.112.1.155

Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in orga-nizations: A meta-analysis. Organizational Behavior and Human Deci-sion Processes, 86, 278–321. doi:10.1006/obhd.2001.2958

*Colbert, A. E., Mount, M. K., Harter, J. K., Witt, L. A., & Barrick, M. R.(2004). Interactive effects of personality and perceptions of the worksituation on workplace deviance. Journal of Applied Psychology, 89,599–609. doi:10.1037/0021-9010.89.4.599

*Colbert, A. E., Witt, L. A., & Mount, M. K. (2003, April). Interactiveeffects of organizational support and agreeableness on interpersonaldeviance. Poster presented at the meeting of Society for Industrial/Organizational Psychology, Orlando, FL.

Colquitt, J. A. (2001). On the dimensionality of organizational justice: Aconstruct validation of a measure. Journal of Applied Psychology, 86,386–400.

Connolly, J. J., & Viswesvaran, C. (2000). The role of affectivity in jobsatisfaction: A meta-analysis. Personality and Individual Differences,29, 265–281.

Conway, J. M., & Huffcutt, A. I. (1997). Psychometric properties ofmultisource performance ratings: A meta-analysis of subordinate, super-visor, peer, and self-ratings. Human Performance, 10, 331–360. doi:10.1207/s15327043hup1004_2

Dalal, R. S. (2005). A meta-analysis of the relationship between organi-zational citizenship behavior and counterproductive work behavior.Journal of Applied Psychology, 90, 1241–1255. doi:10.1037/0021-9010.90.6.1241

*de Jonge, J., & Peeters, M. C. W. (2009). Convergence of self-reportsand coworker reports of counterproductive work behavior: A cross-sectional multi-source survey among health care workers. Interna-tional Journal of Nursing Studies, 46, 699 –707. doi:10.1016/j.ijnurstu.2008.12.010

Dilchert, S., Ones, D. S., Davis, R. D., & Rostow, C. D. (2007). Cognitiveability predicts objectively measured counterproductive work behaviors.Journal of Applied Psychology, 92, 616 – 627. doi:10.1037/0021-9010.92.3.616

*Fallon, J. D., Avis, J. M., Kudisch, J. D., Gornet, T. P., & Frost, A. (2000).Conscientiousness as a predictor of productive and counterproductive


behaviors. Journal of Business Psychology, 15, 339–349. doi:10.1023/A:1007880203956

*Ferris, D. L. (2007). [Correlations between self- and coworker-reports ofcounterproductive work behavior]. Unpublished raw data.

*Ferris, D. L. (2008). [Correlations between self- and coworker-reports ofcounterproductive work behavior]. Unpublished raw data.

*Ferris, D. L., Brown, D. J., Berry, J. W., & Lian, H. (2008). Thedevelopment and validation of the Workplace Ostracism Scale. Journalof Applied Psychology, 93, 1348–1366. doi:10.1037/a0012743

Fox, S., & Spector, P. E. (1999). A model of work frustration-aggression.Journal of Organizational Behavior, 20, 915–931. doi:10.1002/(SICI)1099-1379(199911)20:6�915::AID-JOB918�3.0.CO;2-6

*Fox, S., Spector, P. E., Goh, A., & Bruursema, K. (2007). Does yourcoworker know what you’re doing? Convergence of self- and peer-reports of counterproductive work behavior. International Journal ofStress Management, 14, 41–60. doi:10.1037/1072-5245.14.1.41

Fox, S., Spector, P. E., & Miles, D. (2001). Counterproductive workbehavior (CWB) in response to job stressors and organizationaljustice: Some mediator and moderator tests for autonomy and emo-tions. Journal of Vocational Behavior, 59, 291–309. doi:10.1006/jvbe.2001.1803

Ghiselli, E. E., Campbell, J. P., & Zedeck, S. (1981). Measurementtheory for the behavioral sciences. San Francisco, CA: W. H. Free-man & Co.

Glomb, T. M. (2002). Workplace anger and aggression: Informingconceptual models with data from specific encounters. Journal ofOccupational Health Psychology, 7, 20 –36. doi:10.1037/1076-8998.7.1.20

*Goh, A. (2007). An attributional analysis of counterproductive workbehavior (CWB) in response to occupational stress (Unpublished doc-toral dissertation). University of South Florida.

Goldberg, L. R. (1999). A broad-bandwidth, public-domain, personalityinventory measuring the lower-level facets of several Five-Factormodels. In I. Mervielde, I. J. Deary, D. de Fruyt, & F. Ostendorf(Eds.), Personality psychology in Europe (pp. 7–28). Tilburg, theNetherlands.

*Greenbaum, R. L., Bardes, M., Mayer, D., & Priesemuth, M. (2009,August). Supervisors’ abuse of customers: The moderating role of moralobligation. Paper presented at the Annual Meeting of Academy ofManagement, Chicago, IL.

Hauenstein, N. M. A., McGonigle, T., & Flinder, S. W. (2001). A meta-analysis of the relationship between procedural justice and distributivejustice: Implications for justice research. Employee Responsibilities andRights Journal, 13, 39–56.

Heneman, H. G., Heneman, R. L., & Judge, T. A. (1997). Staffing orga-nizations. Madison, WI: Mendota House/Irwin.

Hershcovis, M. S., Turner, N., Barling, J., Arnold, K. A., Dupre, K. E.,Inness, M., . . . Sivanathan, N. (2007). Predicting workplace aggression:A meta-analysis. Journal of Applied Psychology, 92, 228–238. doi:10.1037/0021-9010.92.1.228

*Hung, T. K., Chi, N. W., & Lu, W. L. (2009). Exploring the relationshipsbetween perceived coworker loafing and counterproductive work behav-iors: The mediating role of a revenge motive. Journal of Business andPsychology, 24, 257–270. doi:10.1007/s10869-009-9104-6

Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Cor-recting error and bias in research findings (2nd ed.). New York, NY:Sage Publications.

John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The Big FiveInventory: Versions 4a and 54 [Technical report]. Berkeley: Instituteof Personality and Social Research, University of California, Berke-ley.

*Jones, D. A. (2004). Toward a better understanding of fairness in theworkplace: Attitude strength, predictive asymmetry, and the revenge

motive (Unpublished doctoral dissertation). University of Calgary, Al-berta, Canada.

*Judge, T. A., LePine, J. A., & Rich, B. L. (2006a). Loving yourselfabundantly: Relationship of the narcissistic personality to self- and otherperceptions of workplace deviance, leadership, and task and contextualperformance. Journal of Applied Psychology, 91, 762–776. doi:10.1037/0021-9010.91.4.762

*Judge, T., Scott, B. A., & Ilies, R. (2006b). Hostility, job attitudes, andworkplace deviance: Test of a multilevel model. Journal of AppliedPsychology, 91, 126–138. doi:10.1037/0021-9010.91.1.126

*Kessler, S. R., Singla, N., Galperin, B. L., & Spector, P. E. (2009,April). Can there be too much of a good thing? Self-esteem and itsimpact on counterproductive work behavior. Poster presented at themeeting of Society for Industrial/Organizational Psychology, NewOrleans, LA.

*Kickul, J. R., Neuman, G., Parker, C., & Finkl, J. (2001). Settling thescore: The role of organizational justice in the relationship betweenpsychological contract breach and anticitizenship behavior. EmployeeResponsibilities and Rights Journal, 13, 77–93. doi:10.1023/A:1014586225406

*Kidwell, R. E., & Bennett, N. (2001). Perceived work context andemployee job neglect. American Business Review, 19, 64–74.

*Kuenzi, M., & Carson, J. (2009, August). Ethical leadership and power:Combined effects on unit outcomes. Poster presented at the meeting ofthe Academy of Management, Chicago, IL.

Lau, V. C. S., Au, W. T., & Ho, J. M. C. (2003). A qualitative andquantitative review of antecedents of counterproductive behavior inorganizations. Journal of Business and Psychology, 18, 73–99.

*Lee, K. (2000). Job affect as a predictor of organizational citizenshipbehavior and workplace deviance (Unpublished doctoral dissertation).University of Western Ontario, Canada.

*Lee, K., & Allen, N. J. (2002). Organizational citizenship behavior andworkplace deviance: The role of affect and cognitions. Journal ofApplied Psychology, 87, 131–142. doi:10.1037/0021-9010.87.1.131

Lee, R. M. (1993). Doing research on sensitive topics. London, England:Sage.

Li, A., & Bagger, J. (2006). Using the BIDR to distinguish the effects ofimpression management and self-deception on the criterion validity ofpersonality measures: A meta-analysis. International Journal of Selec-tion and Assessment, 14, 131–141. doi:10.1111/j.1468-2389.2006.00339.x

Marcus, B., Schuler, H., Quell, P., & Humpfner, G. (2002). Measuringcounterproductivity: Development and initial validation of a Germanself-report questionnaire. International Journal of Selection and Assess-ment, 10, 18–35. doi:10.1111/1468-2389.00191

Moorman, R. H. (1991). Relationship between organizational justiceand organizational citizenship behaviors: Do fairness perceptionsinfluence employee citizenship? Journal of Applied Psychology, 76,845– 855.

Mount, M. K., Barrick, M. R., & Strauss, J. P. (1994). Validity of observerratings of the Big Five personality dimensions. Journal of AppliedPsychology, 79, 272–280. doi:10.1037/0021-9010.79.2.272

*Mount, M., Ilies, R., & Johnson, E. (2006). Relationship of personalitytraits and counterproductive work behaviors: The mediating effects ofjob satisfaction. Personnel Psychology, 59, 591– 622. doi:10.1111/j.1744-6570.2006.00048.x

Murphy, K. R. (1993). Honesty in the workplace. Pacific Grove, CA:Brooks/Cole.

Murphy, K. R., & DeShon, R. (2000). Interrater correlations do notestimate the reliability of job performance ratings. Personnel Psychol-ogy, 53, 873–900. doi:10.1111/j.1744-6570.2000.tb02421.x

Niehoff, B. P., & Moorman, R. H. (1993). Justice as a mediator of therelationship between methods of monitoring and organizational citizen-ship behavior. Academy of Management Journal, 36, 527–556.


*O’Brien, K. (2009). [Correlation between self- and peer-report counter-productive work behavior]. Unpublished raw data.

*O’Brien, K. E., & Allen, T. D. (2007). The relative importance ofcorrelates of organizational citizenship behavior and counterproductivework behavior using multiple sources of data. Human Performance, 21,62–88. doi:10.1080/08959280701522189

*O’Brien, K. E., & Loch, S. E. (2005, April). Development and correlatesof the Workplace Hostile Attribution Bias Survey. Paper presented at theannual meeting of the Society for Industrial/Organizational Psychology,Los Angeles, CA.

Ones, D. S., & Viswesvaran, C. (1998). The effects of social desirabilityand faking on personality and integrity assessment for personnel selec-tion. Human Performance, 11, 245–269.

Ones, D. S., Viswesvaran, C., & Schmidt, F. L. (1993). Comprehensivemeta-analysis of integrity test validities: Findings from implications forpersonnel selection and theories of job performance. Journal of AppliedPsychology, 78, 679–703.

Paulhus, D. L., & John, O. P. (1998). Egoistic and moralistic biases inself-perception: The interplay of self-deceptive styles with basic traitsand motives. Journal of Personality, 66, 1025–1060. doi:10.1111/1467-6494.00041

*Penney, L. M., & Spector, P. E. (2005). Job stress, incivility, and coun-terproductive work behavior (CWB): The moderating role of negativeaffectivity. Journal of Organizational Behavior, 26, 777–796. doi:10.1002/job.336

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003).Common method biases in behavioral research: A critical review of theliterature and recommended remedies. Journal of Applied Psychology,88, 879–903. doi:10.1037/0021-9010.88.5.879

Podsakoff, P. M., MacKenzie, S. B., Moorman, R. H., & Fetter, R. (1990).Transformational leader behaviors and their effects on followers’ trust inleader, satisfaction, and organizational citizenship behaviors. LeadershipQuarterly, 1, 107–142.

Price, J. L., & Mueller, C. W. (1986). Absenteeism and turnover amonghospital employees. Greenwich, CT: JAI Press.

*Puffer, S. M. (1987). Prosocial behavior, noncompliant behavior, andwork performance among commission salespeople. Journal of AppliedPsychology, 72, 615–621. doi:10.1037/0021-9010.72.4.615

Raju, N. S., & Brand, P. A. (2003). Determining the significance of correla-tions corrected for unreliability and range restriction. Applied PsychologicalMeasurement, 27, 52–71. doi:10.1177/0146621602239476

*Richards, D., & Schat, A. (2011). [Correlation between self- andsupervisor-ratings of counterproductive work behavior]. Unpublishedraw data.

Richman, W. L., Kiesler, S., Weisband, S., & Drasgow, F. (1999). Ameta-analytic study of social desirability distortion in computer-administered questionnaires, traditional questionnaires, and interviews.Journal of Applied Psychology, 84, 754 –775. doi:10.1037/0021-9010.84.5.754

Robinson, S. L., & O’Leary-Kelly, A. M. (1998). Monkey see, monkey do:The influence of work groups on the antisocial behavior of employees.Academy of Management Journal, 41, 658–672.

*Rotundo, M. (2006). [Intercorrelations between self-rated and supervisor-rated dimensions of performance]. Unpublished raw data.

Rotundo, M., & Spector, P. E. (2010). Counterproductive work behaviorand withdrawal. In J. L. Farr & N. T. Tippins (Eds.), Handbook ofemployee selection (pp. 489–511). New York, NY: Routledge.

Sackett, P. R. (2002). The structure of counterproductive work behaviors:Dimensionality and relationships with facets of job performance. Inter-national Journal of Selection and Assessment, 10, 5–11. doi:10.1111/1468-2389.00189

Sackett, P. R., & DeVore, C. J. (2001). Counterproductive behaviors atwork. In N. Anderson, D. S. Ones, H. K. Sinangil, & V. Viswesvaran

(Eds.), Handbook of industrial, work, and organizational psychology(Vol. 1, pp. 145–164). London, England: Sage.

*Sady, K., Spitzmuller, C., & Witt, L. A. (2008, August). Good employee,bad business: An interactionist approach to workplace deviance. Paperpresented at the annual meeting of the Academy of Management, Ana-heim, CA.

Salgado, J. F. (2002). The Big Five personality dimensions and counter-productive behaviors. International Journal of Selection and Assess-ment, 10, 117–125. doi:10.1111/1468-2389.00198

Schmidt, F. L., Viswesvaran, C., & Ones, D. S. (2000). Reliability is notvalidity and validity is not reliability. Personnel Psychology, 53, 901–912. doi:10.1111/j.1744-6570.2000.tb02422.x

*Skarlicki, D. P., & Folger, R. (1997). Retaliation in the workplace: Theroles of distributive, procedural, and interactional justice. Journal ofApplied Psychology, 82, 434–443. doi:10.1037/0021-9010.82.3.434

Spector, P. E. (2006). Method variance in organizational research: Truth orurban legend? Organizational Research Methods, 9, 221–232. doi:10.1177/1094428105284955

*Spector, P. E., Bauer, J., & Fox, S. (2010). Measurement artifacts in theassessment of counterproductive work behavior and organizational cit-izenship behavior: Do we know what we think we know? Journal ofApplied Psychology, 95, 781–790. doi:10.1037/a0019477

Spector, P. E., Fox, S., Penney, L. M., Bruursema, K., Goh, A., & Kessler,S. (2006). The dimensionality of counterproductivity: Are all counter-productive behaviors created equal? Journal of Vocational Behavior, 68,446–460.

Spector, P. E., & Jex, S. M. (1998). Development of four self-reportmeasures of job stressors and strain: Interpersonal Conflict at WorkScale, Organizational Constraints Scale, Quantitative Workload Inven-tory, and Physical Symptoms Inventory. Journal of Occupational HealthPsychology, 3, 356–367. doi:10.1037/1076-8998.3.4.356

*Stamper, C. L., & Masterson, S. S. (2002). Insider or outsider? Howemployee perceptions of insider status affect their work behavior.Journal of Organizational Behavior, 23, 875– 894. doi:10.1002/job.175

Stewart, S. M., Bing, M. N., Davison, H. K., Woehr, D. J., & McIntyre,M. D. (2009). In the eyes of the beholder: A non-self-report measure ofworkplace deviance. Journal of Applied Psychology, 94, 207–215. doi:10.1037/a0012605

*Thau, S., Crossley, C., Bennett, R. J., & Sczesny, S. (2007). The rela-tionship between trust, attachment, and antisocial work behaviors. Hu-man Relations, 60, 1155–1179. doi:10.1177/0018726707081658

*Tubre, T., Arthur, W., Jr., & Bennett, W., Jr. (2006). General models ofjob performance: Theory and practice. In W. Bennett, C. E. Lance, &D. J. Woehr (Eds.), Performance measurement: Current perspectivesand future challenges (pp. 175–203). Mahwah, NJ: Erlbaum.

Van Katwyk, P. T., Fox, S., Spector, P. E., & Kelloway, E. K. (2000).Using the Job-Related Affective Well-Being Scale (JAWS) to investi-gate affective responses to work stressors. Journal of OccupationalHealth Psychology, 5, 219–230.

Vardi, Y., & Weitz, E. (2004). Misbehavior in organizations: Theory,research, and management. Mahwah, NJ: Erlbaum.

Viswesvaran, C., & Ones, D. S. (2000). Measurement error in “Big Five Factors”personality assessment: Reliability generalization across studies and measures.Educational and Psychological Measurement, 60, 224–235.

Viswesvaran, C., Ones, D. S., & Schmidt, F. L. (1996). Comparativeanalysis of the reliability of job performance ratings. Journal of AppliedPsychology, 81, 557–574.

Viswesvaran, C., Schmidt, F. L., & Ones, D. S. (2002). The moderatinginfluence of job performance dimensions on convergence of supervisoryand peer ratings of job performance: Unconfounding construct-levelconvergences and rating difficulty. Journal of Applied Psychology, 87,345–354. doi:10.1037/0021-9010.87.2.345

Watson, D., Clark, L. A., & Tellegen, A. (1998). Development and validation of


brief measures of positive and negative affect: The PANAS scales. Journal ofPersonality and Social Psychology, 54, 1063–1070.

*White, L. A., Young, M. C., & Rumsey, M. G. (2001). Able implemen-tation issues and related research. In J. P. Campbell & D. J. Knapp(Eds.), Exploring the limits in personnel selection and classification (pp.525–558). Mahwah, NJ: Erlbaum.

Williams, L. J., & Anderson, S. E. (1991). Job satisfaction and organiza-tional commitment as predictors of organizational citizenship and in-rolebehaviors. Journal of Management, 17, 601–617.

*Williams, M. (1999). When is personality a predictor of performance?The moderating role of autonomy (Unpublished doctoral dissertation).Florida International University, Miami.

Appendix

Table A1List of Search Terms Used in the Literature Search for Primary Data

Search terms

Anticitizenship Poor work attendanceBullying Production devianceCounterproductive work behavior Property devianceCounterproductivity Property theftMisuse of resources SabotageMisuse of time Stealing propertyMisuse of workplace resources Stealing resourcesNoncompliant behavior TheftOrganization motivated aggression Undesirable work behaviorOrganizational aggression Workplace aggressionOrganizational deviance Workplace devianceOrganizational misbehavior Workplace deviant behaviorOrganizational retaliation behavior Workplace gossipPersonal aggression Workplace misconductPolitical deviance

(Appendices continue)


Tab

leA

2R

elev

ant

Pie

ces

ofIn

form

atio

nfo

rE

ach

Pri

mar

ySa

mpl

e

Ref

eren

ceSa

mpl

e#

NC

WB

-1C

WB

-2

Self

-re

port

-O

ther

-re

port

%fe

mal

e#

ofan

onym

itysa

fegu

ards

CW

Bco

rrel

ate

Oth

er-r

epor

t-C

WB

-co

rrel

ate

rr

d

Bin

get

al.

(200

7)3

176

Cow

orke

r—

——

——

OC

B-I

b�

.29

OC

B-O

b�

.56

Bor

dia

etal

.(2

008)

216

8Su

perv

isor

CW

B-O

——

—48

.2—

Gen

der

.06

Age

�.0

2T

enur

e�

.02

318

7Su

perv

isor

CW

B-I

——

—58

.0—

Gen

der

�.0

6A

ge�

.10

Ten

ure

�.0

4Su

perv

isor

CW

B-O

——

——

Gen

der

�.0

3A

ge.0

0T

enur

e.0

0B

ruk-

Lee

&Sp

ecto

r(2

006)

112

1Se

lfC

owor

ker

.30

�.1

176

.72

Cow

orke

r—

——

—G

ende

r�

.18

Age

�.1

8T

enur

e�

.03

Neg

ativ

eem

otio

n.2

5C

owor

ker

conf

lictc

.39

Supe

rvis

orco

nflic

tc.1

4C

owor

ker

conf

lictb

.50

Supe

rvis

orco

nflic

tb.3

6C

hiu

&Pe

ng(2

008)

123

3Su

perv

isor

CW

B-I

——

—70

.0—

Gen

der

�.0

9A

ge�

.07

Supe

rvis

orC

WB

-OG

ende

r�

.01

Age

�.1

1C

olbe

rtet

al.

(200

4)1

239

Supe

rvis

orW

E—

——

——

Em

otio

nal

Stab

ility

�.0

2Su

perv

isor

WE

Con

scie

ntio

usne

ss�

.23

231

9Su

perv

isor

CW

B-I

——

——

—E

mot

iona

lSt

abili

ty.0

1Su

perv

isor

CW

B-I

Con

scie

ntio

usne

ss�

.21

3a17

3—

——

——

Agr

eeab

lene

ss�

.50

4a12

2—

——

——

Agr

eeab

lene

ss�

.55

Col

bert

etal

.(2

003)

117

3Su

perv

isor

CW

B-I

——

—48

.0—

Gen

der

�.3

1A

ge�

.08

Em

otio

nal

Stab

ility

�.2

7E

xtra

vers

ion

�.2

0O

penn

ess

�.1

4C

onsc

ient

ious

ness

�.4

8

(App

endi

ces

cont

inue

)


Tab

leA

2(c

onti

nued

)

Ref

eren

ceSa

mpl

e#

NC

WB

-1C

WB

-2

Self

-re

port

-O

ther

-re

port

%fe

mal

e#

ofan

onym

itysa

fegu

ards

CW

Bco

rrel

ate

Oth

er-r

epor

t-C

WB

-co

rrel

ate

rr

d

222

9Su

perv

isor

CW

B-I

——

—42

.0—

Gen

der

�.0

5A

ge�

.09

Em

otio

nal

Stab

ility

�.0

8E

xtra

vers

ion

.08

Ope

nnes

s�

.08

Agr

eeab

lene

ss�

.37

Con

scie

ntio

usne

ss.0

13

122

Supe

rvis

orC

WB

-I—

——

68.0

—G

ende

r�

.34

Age

.16

Em

otio

nal

Stab

ility

.04

Ext

rave

rsio

n.1

8O

penn

ess

�.0

1C

onsc

ient

ious

ness

.06

deJo

nge

&Pe

eter

s(2

009)

154

Self

Cow

orke

r.4

6.2

8—

2Fa

llon

etal

.(2

000)

132

1Su

perv

isor

CW

B-O

——

——

—C

onsc

ient

ious

ness

.09

Ferr

is(2

007)

119

1Se

lfC

owor

ker

.42

Self

CW

B-I

Cow

orke

rC

WB

-I.4

5C

owor

ker

CW

B-O

Self

CW

B-O

.40

Ferr

is(2

008)

116

4Se

lfC

owor

ker

.55

Self

CW

B-I

Cow

orke

rC

WB

-I.5

8C

owor

ker

CW

B-O

Self

CW

B-O

.49

Ferr

iset

al.

(200

8)3

164

Peer

CW

B-I

——

—60

.0—

Gen

der

�.1

4A

ge�

.20

OC

B-I

�.1

7O

CB

-O�

.45

Peer

CW

B-O

——

—60

.0—

Gen

der

�.0

7A

ge.0

0O

CB

-I�

.23

OC

B-O

�.5

9Pe

er—

——

OC

Bco

wor

ker

�.4

6O

CB

-Ico

wor

ker

�.2

2O

CB

-Oco

wor

ker

�.5

7Fo

xet

al.

(200

7)1

132

Self

Cow

orke

r.3

0.2

4—

—Se

lfC

WB

-IC

owor

ker

CW

B-I

.47

Cow

orke

rC

WB

-OSe

lfC

WB

-O—

.13

(App

endi

ces

cont

inue

)


Tab

leA

2(c

onti

nued

)

Ref

eren

ceSa

mpl

e#

NC

WB

-1C

WB

-2

Self

-re

port

-O

ther

-re

port

%fe

mal

e#

ofan

onym

itysa

fegu

ards

CW

Bco

rrel

ate

Oth

er-r

epor

t-C

WB

-co

rrel

ate

rr

d

Cow

orke

rC

WB

-I—

——

—O

rgan

izat

iona

lco

nstr

aint

sc.1

2In

terp

erso

nal

conf

lictc

.25

Proc

edur

alju

stic

ec�

.29

Dis

trib

utiv

eju

stic

ec�

.06

—O

rgan

izat

iona

lco

nstr

aint

sb.3

6In

terp

erso

nal

conf

lictb

.61

Cow

orke

rC

WB

-O—

——

—O

rgan

izat

iona

lco

nstr

aint

sc.1

3In

terp

erso

nal

conf

lictc

.20

Proc

edur

alju

stic

ec�

.35

Dis

trib

utiv

eju

stic

ec�

.11

Org

aniz

atio

nal

cons

trai

ntsb

.40

Inte

rper

sona

lco

nflic

tb.5

2G

oh(2

007)

114

6Se

lfC

owor

ker

.42

.17

—2

Cow

orke

r—

——

—In

form

atio

nal

just

ice

�.1

7In

terp

erso

nal

just

ice

�.2

0In

terp

erso

nal

conf

lictc

.37

Inte

rper

sona

lco

nflic

tb.5

0In

terp

erso

nal

cons

trai

ntsc

.27

Inte

rper

sona

lco

nstr

aint

sb.3

9Jo

bco

ntex

tsco

nstr

aint

sc.2

5Jo

bco

ntex

tsco

nstr

aint

sb.3

8O

rgan

izat

iona

lco

nstr

aint

sc.2

8O

rgan

izat

iona

lco

nstr

aint

sb.4

1N

egat

ive

affe

ctiv

ity.2

4G

reen

baum

etal

.(2

009)

112

2Su

perv

isor

——

——

—O

CB

�.4

7H

ung

etal

.(2

009)

118

4Se

lfSu

perv

isor

.24

.00

—1

Self

CW

B-I

Supe

rvis

orC

WB

-I.2

7Su

perv

isor

CW

B-O

Self

CW

B-O

—.1

9Su

perv

isor

——

——

Neg

ativ

eaf

fect

ivity

.04

Jone

s(2

004)

117

3Se

lfC

owor

ker

.16

.45

—1

Cow

orke

rC

WB

-IC

owor

ker

CW

B-I

.23

Cow

orke

rC

WB

-OSe

lfC

WB

-O—

.17

(App

endi

ces

cont

inue

)


Tab

leA

2(c

onti

nued

)

Ref

eren

ceSa

mpl

e#

NC

WB

-1C

WB

-2

Self

-re

port

-O

ther

-re

port

%fe

mal

e#

ofan

onym

itysa

fegu

ards

CW

Bco

rrel

ate

Oth

er-r

epor

t-C

WB

-co

rrel

ate

rr

d

Cow

orke

r—

——

—Pr

oced

ural

just

ice–

Pay

�.1

2Pr

oced

ural

just

ice–

Dir

ect

�.0

9Pr

oced

ural

just

ice–

Sche

dulin

g�

.23

Dis

trib

utiv

eju

stic

e–Pa

yD

istr

ibut

ive

just

ice–

Sche

dulin

g�

.02

OC

Bb

.08

OC

B-O

b

—In

terp

erso

nal

just

ice

�.4

0In

form

atio

nal

just

ice

�.3

42

158

Cow

orke

r—

—82

.0—

Dir

ect

inte

ract

iona

lju

stic

e�

.18

OC

Bb

�.1

3G

ende

r�

.11

Age

�.4

0�

.02

�.0

5Ju

dge

etal

.(2

006a

)2

131

Self

Supe

rvis

or.2

0.9

4—

1N

euro

ticis

m.1

4Su

perv

isor

——

——

Ext

rave

rsio

n.1

2O

penn

ess

�.1

0A

gree

able

ness

�.0

5C

onsc

ient

ious

ness

�.0

7O

CB

-sel

f�

.14

Judg

eet

al.

(200

6b)

164

Supe

rvis

or—

——

——

Inte

rper

sona

lju

stic

e�

.24

Job

satis

fact

ion

�.1

4K

essl

eret

al.

(200

9)1a

96Se

lfSu

perv

isor

.72

.25

——

Job

satis

fact

ion

�.2

61b

85Su

perv

isor

——

——

Gen

der

�.2

4Su

perv

isor

——

—71

.8K

icku

let

al.

(200

1)1

165

Supe

rvis

or—

——

49.7

—G

ende

r�

.15

Ten

ure

�.2

6Pr

oced

ural

just

ice

�.4

9In

tera

ctio

nal

just

ice

�.5

7K

idw

ell

&B

enne

tt(2

001)

155

6Se

lfSu

perv

isor

.15

0O

CB

-I(s

uper

viso

r)�

.09

Supe

rvis

or—

—K

uenz

i&

Car

son

(200

9)1

311

Supe

rvis

orC

WB

-I—

——

——

OC

B-I

a�

.33

Lee

(200

0)1

146

Self

Cow

orke

r.3

3.3

2—

1O

CB

(com

posi

te)c

�.0

7C

owor

ker

——

—O

CB

-Oc

�.0

1O

CB

-Ic

�.1

1

(App

endi

ces

cont

inue

)


Tab

leA

2(c

onti

nued

)

Ref

eren

ceSa

mpl

e#

NC

WB

-1C

WB

-2

Self

-re

port

-O

ther

-re

port

%fe

mal

e#

ofan

onym

itysa

fegu

ards

CW

Bco

rrel

ate

Oth

er-r

epor

t-C

WB

-co

rrel

ate

rr

d

Lee

&A

llen

(200

2)1

155

Cow

orke

r—

——

—A

ge.0

0T

enur

e.1

0N

egat

ive

affe

ct.1

4Pr

oced

ural

just

ice

�.0

5M

ount

etal

.(2

006)

114

1Se

lfSu

perv

isor

.38

.02

——

Em

otio

nal

Stab

ility

�.2

2Se

lfC

WB

-ISu

perv

isor

CW

B-I

.48

—E

xtra

vers

ion

�.1

1Se

lfC

WB

-OSu

perv

isor

CW

B-O

.21

Ope

nnes

s�

.22

Supe

rvis

or—

—A

gree

able

ness

�.1

4C

onsc

ient

ious

ness

�.2

3Jo

bsa

tisfa

ctio

n�

.19

O’B

rien

(200

9)1

176

Cow

orke

r—

——

——

Ext

rave

rsio

n�

.04

Con

scie

ntio

usne

ss�

.18

Em

otio

nal

Stab

ility

�.0

2A

gree

able

ness

�.0

9O

’Bri

en&

Alle

n(2

007)

120

7Se

lfSu

perv

isor

.36

.41

—1

OC

B-I

a�

.48

Self

CW

B-I

Supe

rvis

orC

WB

-I.4

8—

OC

B-O

a�

.54

Self

CW

B-O

Supe

rvis

orC

WB

-O.1

9O

CB

-Ic

�.1

6Su

perv

isor

——

OC

B-O

c�

.14

Org

.ju

stic

e�

.15

Job

satis

fact

ion

�.1

0C

onsc

ient

ious

ness

�.1

6O

’Bri

en&

Loc

h(2

005)

117

6Se

lfC

owor

ker

.22

.25

——

OC

Bc

�.1

217

6C

owor

ker

——

—Jo

bsa

tisfa

ctio

n�

.12

175

Dis

trib

utiv

eju

stic

e.1

317

2Pr

oced

ural

just

ice

�.0

811

1O

CB

b�

.48

Ten

ure

�.0

4Pe

nney

&Sp

ecto

r(2

005)

115

5Se

lfC

owor

ker

.24

�.0

5—

2N

egat

ive

affe

ct.1

9Se

lfC

WB

-IC

owor

ker

CW

B-I

.30

—C

onfl

ictc

.22

Self

CW

B-O

Cow

orke

rC

WB

-O.2

2C

onfl

ictb

.50

Cow

orke

r—

—C

onst

rain

ts.0

7Jo

bsa

tisfa

ctio

n�

.25

Puff

er(1

987)

114

1Su

perv

isor

——

——

—O

CB

a�

.74

Job

satis

fact

ion

�.1

6

(App

endi

ces

cont

inue

)


Tab

leA

2(c

onti

nued

)

Ref

eren

ceSa

mpl

e#

NC

WB

-1C

WB

-2

Self

-re

port

-O

ther

-re

port

%fe

mal

e#

ofan

onym

itysa

fegu

ards

CW

Bco

rrel

ate

Oth

er-r

epor

t-C

WB

-co

rrel

ate

rr

d

Ric

hard

s&

Scha

t(2

011)

115

3Se

lfSu

perv

isor

.72

.19

153

Self

CW

B-I

Supe

rvis

orC

WB

-I.7

0Se

lfC

WB

-OSu

perv

isor

CW

B-O

.67

Rot

undo

(200

6)1

30Se

lfSu

perv

isor

.17

——

Sady

etal

.(2

008)

117

1C

owor

ker

——

——

—E

mot

iona

lSt

abili

ty�

.11

Satis

fact

ion

with

wor

k�

.19

Satis

fact

ion

with

cow

orke

rs�

.17

Skar

licki

&Fo

lger

(199

7)1

167

Cow

orke

r—

——

——

Dis

trib

utiv

eju

stic

e�

.44

Proc

edur

alju

stic

e�

.53

Inte

ract

iona

lju

stic

e�

.54

Spec

tor

etal

.(2

010)

111

9Se

lfSu

perv

isor

.48

.48

—2

OC

Bc

�.3

32

140

Self

Supe

rvis

or.4

2.6

1—

2O

CB

-Ico

mpo

site

c�

.32

111

9Su

perv

isor

——

—O

CB

-Oco

mpo

site

c�

.26

214

0Su

perv

isor

——

—In

terp

erso

nal

conf

lict

atw

ork

.39

OC

Bc

�.2

0O

CB

-Ico

mpo

site

c�

.57

OC

B-O

com

posi

tec

�.0

2In

terp

erso

nal

conf

lict

atw

ork

.37

Stam

per

&M

aste

rson

(200

2)1

257

Supe

rvis

or—

——

75.0

—T

enur

e�

.16

Gen

der

.21

Age

�.0

6O

CB

-Ia

�.4

5T

hau

etal

.(2

007)

287

Supe

rvis

or–

——

53.0

—G

ende

r�

.04

310

6Su

perv

isor

——

——

—A

ge.0

0T

enur

e.0

8A

ge�

.03

Ten

ure

.07

Tub

reet

al.

(200

6)1

287

Self

Supe

rvis

or.1

4.4

7—

1W

hite

etal

.(2

001)

159

0Su

perv

isor

——

——

—E

mot

iona

lSt

abili

ty.0

3A

gree

able

ness

.07

Con

scie

ntio

usne

ss.2

0

(App

endi

ces

cont

inue

)


Tab

leA

2(c

onti

nued

)

Ref

eren

ceSa

mpl

e#

NC

WB

-1C

WB

-2

Self

-re

port

-O

ther

-re

port

%fe

mal

e#

ofan

onym

itysa

fegu

ards

CW

Bco

rrel

ate

Oth

er-r

epor

t-C

WB

-co

rrel

ate

rr

d

M.

Will

iam

s(1

999)

194

Supe

rvis

or—

——

——

OC

Ba

�.3

5C

onsc

ient

ious

ness

�.0

6N

euro

ticis

m�

.12

Ext

rave

rsio

n.2

1O

penn

ess

�.0

5A

gree

able

ness

�.0

4Jo

bsa

tisfa

ctio

n�

.05

Not

e.C

WB

-1�

firs

tC

WB

vari

able

(e.g

.,se

lf-r

atin

gsof

CW

B);

CW

B-2

�se

cond

CW

Bva

riab

le(e

.g.,

supe

rvis

orra

tings

ofC

WB

);Se

lf-r

epor

t-O

ther

-rep

ort

r�

corr

elat

ion

betw

een

self

-an

dot

her-

repo

rtC

WB

;Sel

f-re

port

-Oth

er-r

epor

td�

stan

dard

ized

mea

ndi

ffer

ence

betw

een

self

-an

dot

her-

ratin

gsof

CW

B(p

ositi

veds

mea

nse

lf-r

epor

twas

high

er);

%fe

mal

e�

perc

enta

geof

the

sam

ple

that

was

fem

ale;

#of

anon

ymity

safe

guar

ds�

the

num

ber

ofas

sura

nces

ofan

onym

itypr

ovid

edto

part

icip

ants

(0,

1,or

2);

CW

Bco

rrel

ate

�va

riab

leco

rrel

ated

with

othe

r-re

port

CW

B;

Oth

er-r

epor

t-C

WB

-cor

rela

ter

�co

rrel

atio

nbe

twee

not

her-

repo

rtC

WB

and

the

CW

Bco

rrel

ate;

Self

�se

lf-r

epor

t;Su

perv

isor

�su

perv

isor

repo

rt;C

owor

ker

�co

wor

ker

repo

rt;O

CB

-Ian

dO

CB

-O�

Inte

rper

sona

l-an

dor

gani

zatio

nal

targ

etor

gani

zatio

nal

citiz

ensh

ipbe

havi

ors,

resp

ectiv

ely;

CW

B-I

and

CW

B-O

�in

terp

erso

nal-

and

orga

niza

tiona

lta

rget

coun

terp

rodu

ctiv

ew

ork

beha

vior

s,re

spec

tivel

y;W

E�

with

hold

ing

effo

rt.

Das

hes

resp

rese

ntin

stan

ces

inw

hich

that

piec

eof

info

rmat

ion

was

not

avai

labl

e.a

Supe

rvis

or-r

atin

gs.

bC

owor

ker

ratin

gs.

cSe

lf;

all

othe

rno

nbeh

avio

ral

corr

elat

esw

ere

self

-rep

orte

d.

Rec

eive

dJu

ne9,

2010

Rev

isio

nre

ceiv

edN

ovem

ber

11,

2011

Acc

epte

dN

ovem

ber

28,

2011

�


Do Other-Reports of Counterproductive Work Behavior Provide an

Documents

Transcript of Do Other-Reports of Counterproductive Work Behavior Provide an