the mixed blessing of leader sense of humor:...

r Academy of Management Journal2018, Vol. 61, No. 1, 348–369.https://doi.org/10.5465/amj.2015.1088

THE MIXED BLESSING OF LEADER SENSE OF HUMOR:EXAMINING COSTS AND BENEFITS

KAI CHI YAMNational University of Singapore

MICHAEL S. CHRISTIANUniversity of North Carolina Chapel Hill

WU WEIWuhan University

ZHENYU LIAOWashington University in St. Louis

JARED NAISingapore Management University

Workplace humor is ubiquitous, yet scholars know little about how it affects employees’behaviors in organizations. We draw on an emerging psychological theory of humor—benign violation theory—to suggest that a leader’s sense of humor often conveys counter-normative social information in organizations. We integrate this theory with socialinformation processing theory to develop hypotheses about the effects of a leader’s senseof humor on follower behavior. We suggest that although a leader’s sense of humor ispositively associated with leader2member exchange and ultimately work engagement,it can also signal to followers the acceptability of norm violation at work. These per-ceptions in turn are positively associated with followers’ deviance. Furthermore, wepropose that these indirect effects are moderated by leader aggressive humor. Data fromtwo three-wave field studies in China and the United States provide support for ourhypotheses. Taken together, our results suggest that a leader’s sense of humor can bea mixed blessing and elicit unforeseen negative behaviors from their followers.

A sense of humor1 is often said to be a criticalcomponent of successful leadership. Broadly de-fined, a sense of humor refers to a trait-like individ-ual tendency to use or display behaviors, attitudes,

and abilities relating to amusement during socialinteractions (Martin, 2001). Successful leaders an-ecdotally often use their sense of humor to garnersupport, motivate their followers, and even createlasting memories for the world. For example, formerCalifornia governor Arnold Schwarzenegger was hitby an egg in a rally by an angry protester. In his re-sponse to multiple media outlets, he said “this guyowes me bacon now.” His humor eased tension andeven gathered support from protesters. In the cor-porate world, many successful CEOs are likewisepraised for their sense of humor. When asked toadvise junior analysts in his organization, WarrenBuffet said “I try to buy stock in businesses that are sowonderful that an idiot can run them. Becausesooner or later, one will.” Practitioners thus oftencredit increased employee satisfaction and perfor-mance to leaders’ senses of humor (Katz, 1996).Given the perceived effectiveness of humor, some

The last three authors contributed equally and thus arelisted in an arbitrary order.We thankAssociate EditorAmyColbert and three anonymous reviewers for their con-structive feedback throughout the review process. We alsothank Scott J. Reynolds, Christopher Barnes, and RyanFehr for helpful feedback on an earlier draft of this manu-script. This research was supported by NUS start up grantWBS R-317-000-117-133, the National Natural ScienceFoundation ofChinaunderGrant 71272231, and theSocialScience Foundation of the Ministry of Education of Chinaunder Grant 17YJA630105.

1 Following past research on humor (Mesmer-Magnus,Glew, & Viswesvaran, 2012), in this paper we consideredthe terms humor and sense of humor to be synonymous.

348

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s expresswritten permission. Users may print, download, or email articles for individual use only.

https://doi.org/10.5465/amj.2015.1088

political and business leaders have even hired hu-mor coaches to improve their leadership effective-ness (Dampier & Walton, 2013).

We draw on benign violation theory (BVT;McGraw & Warren, 2010) to explain when and whya leader’s sense of humor influences the workplacein positive and negative ways. Briefly stated, BVTsuggests that the display of humor often necessitatesa benign norm violation. We integrate BVT withsocial information processing theory (Salancik &Pfeffer, 1978) to propose that a leader’s sense of hu-mor signals the acceptability of norm violations inthe organization, leading followers to engage in in-creased workplace deviance, or “voluntary behaviorthat violates significant organizational norms, and inso doing, threatens thewell-being of the organizationand/or its members” (Robinson & Bennett, 1995:556). Moreover, our framework also extends pastresearch onpositive outcomes of humor (e.g.,Mesmer-Magnus, Glew, & Viswesvaran, 2012). We argue thatby violating benign norms in their interactions withfollowers, leader humor creates interpersonal rap-port by signaling relational openness andpermissivedispositional qualities, thus positively relating to thequality of leader2member exchanges (LMX) andfollower work engagement. In other words, we sug-gest that leader sense of humor can have indirecteffects on followers’ deviance and work engagementvia perceived acceptability of norm violation andLMX respectively.

Furthermore, we provide a more complete un-derstanding of the “mixed blessing” of a leader’ssense of humor by examining the moderating role ofthe style of humor a leader tends to use. We con-ceptualize sense of humor as a “broad bandwidth”trait construct, and humor styles as more specific

“narrow facets” that represent within-person con-sistency in tendencies to enact a sense of humorwithcertain patterns of behavior (see Moon, 2001). Whilehaving a sense of humor is valence free, a leadermay have a specific tendency to express his or hersense of humor in interpersonally negative or posi-tiveways.Wehelp advance the literature by focusingon the tendency to use an aggressive humor style(e.g., Martin, Puhlik-Doris, Larsen, Gray, & Weir,2003; Mesmer-Magnus et al., 2012), thus providinga more nuanced and precise view of the effects ofa leader’s sense of humor on followers. Specifically,we theorize that the indirect relation of leader humorwith follower deviance will be strengthened, andthe indirect relation of leader humor with followerwork engagement will be weakened, when a leadertends to use aggressive humor, a negatively valencedstyle of humor that is carried out “at the expenseand detriment of one’s relationships with others”(Martin et al., 2003: 52); often referred to as teasingwith a humorous undertone (Mesmer-Magnus et al.,2012).

We theorize that the effects of a leader’s sense ofhumor on followers’ perceptions will be moderatedby aggressive humor style, and ultimately be asso-ciated with follower outcomes. The more aggressivea leader’s style, themore a sense of humorwill signalacceptability of norm violations, which will be pos-itively associated with deviance; in contrast, themore aggressive a leader’s style, the less a sense ofhumor will signal high quality LMX, which willbe negatively associated with work engagement(Figure 1). We test our hypotheses with two three-wave field studies of full-time employees in Chinaand the United States. In the first study, we examineour basic proposition of the indirect effect of leader

FIGURE 1Theoretical Model of the Current Research

Leader sense ofhumor

Perceivedacceptability ofnorm violation

Leader aggressivehumor

Followerdeviance

Leader memberexchange

Workengagement

2018 349Yam, Christian, Wei, Liao, and Nai

sense of humor on follower deviance through signalsof norm violation acceptability. In the second study,we extend these results by testing our full theoreticalframework with a moderated mediation model andboth positive and negative follower outcomes.

Our research makes a number of theoretical andpractical contributions. First, humor research inorganizations has been characterized as “sporadic”(Duncan, Smeltzer, & Leap, 1990), partly becausethe literature is fragmented and lacks an overarch-ing theoretical framework. Although scholars havewritten about leader humor, most research is theo-retical in nature (Cooper, 2005, 2008; Duncan, 1982;Malone, 1980). As a result, there is very little em-pirical research examining the effects of a leader’ssense of humor. As Crawford (1994: 54) noted “ofall the communicative strategies that leaders utilize,the use of humor is most promising, but least un-derstood.” Leadership scholars consequently lacka complete understanding of a prevalent phenome-non in organizations, highlighting the need fornew theoretical and empirical investigations. Byintroducing BVT (McGraw & Warren, 2010), whichhas been largely supported by multiple empiricalstudies in social psychology (e.g.,McGraw&Warner,2014; McGraw & Warren, 2010; McGraw, Williams,& Warren, 2014; McGraw, Warren, Williams, &Leonard, 2012), our research moves the organiza-tional literature forward by providing a frameworkfor understanding humor in organizations. Second,in contrastwith past research,we argue that humor isnot a panacea, suggesting that although a leader’ssense of humor may increase work engagementamong followers, it may also increase follower de-viance. Third, we further contribute to the literatureof leader humor by demonstrating why and whenthese effects occur, both critical components forbuilding and testing theory (Colquitt & Zapata-Phelan, 2007). In terms of why, we hypothesizethat follower perceived acceptability of norm viola-tions is the mechanism underlying the negative as-sociation of leader sense of humor with deviancewhereas heightened LMX underlies the positive as-sociation of leader sense of humor with work en-gagement. In terms of when, we identify the style ofhumor used—humor that is aggressive in nature—for which these effects will be the strongest or willdissipate. Finally, our findings shed light on the prosand cons of leader humor in the workplace, pro-viding practitioners a more nuanced understandingof this mixed blessing and a guide to leverage thepositive outcomes of leader humor while minimiz-ing its negative impacts.

THEORY AND HYPOTHESIS DEVELOPMENT

Benign Violation Theory

McGraw and colleagues (McGraw&Warner, 2014;McGraw & Warren, 2010; McGraw et al., 2012) pro-posed BVT to explain what is perceived to be hu-morous. BVT suggests that humor is comprised ofthree interrelated components. First, a norm viola-tion must occur. The norm violation can come informs of physical or symbolic violations, and canrange from social to moral norm violations. Forexample, we often laugh when being tickled (i.e.,a violation of a physical norm), or when hearingstories or jokes that violate an expected social norm(e.g., “what do dinosaurs and decent lawyers havein common? They are both extinct”). Second, thenorm violation must be perceived as benign—ornonthreatening—in nature. For example, peoplegenerally laugh when tickled by loved ones(i.e., perceived as benign), but not when tickled bystrangers (i.e., perceived as threatening and hencenot funny). In addition, the lawyer joke above wouldlikely not be perceived as humorous if it was actu-ally accompanied by images of dead lawyers. Inother words, excessive norm violations that offend orthreaten the perceiver can dampen humor. Third,humor requires the first two conditions to be inter-preted simultaneously (McGraw & Warren, 2010).Like all other theories, a critical boundary conditionof BVT is that it does not attempt to explain all typesof humor generation. Rather, BVT argues that thebroadest domain of humor is produced by leveragingbenign violations.

Empirical research has largely supported the pre-dictions made by BVT. For example, when partici-pants are temporally, socially, hypothetically, orspatially distant from the violations (i.e., high psy-chological distance), they tend to view those viola-tions as benign in nature and hence funny (McGrawet al., 2012). In a longitudinal study, for example,participants found jokes regarding the destruction ofHurricane Sandy to be not funny during the crisis(i.e., when psychological distance is close and hencemalign violations), but progressively funnier as thepassageof time assuages the threateningnature of thedisaster by expanding the psychological distancebetween the participants and the event (McGrawet al., 2014). Neuroscientific studies have also pro-vided some support for BVT. For example, Goeland Dolan (2001) found that the juxtapositions oftwo mental sets (i.e., norm violation and benignity)lead to neurological activity in the ventral medialprefrontal cortex (a region that is associated with

350 FebruaryAcademy of Management Journal

affective responses) and the experience of humor.Beyond empirical studies, BVT has also been sup-ported by numerous anecdotes. For example, co-medians David Letterman and John Stewart wereheavily criticized for joking too soon after the 9/11terrorist attack.

Given that humor often requires norm viola-tions, displays of humor will likely signal to othersthat norm violations are socially acceptable in in-terpersonal interactions. This is particularly rele-vant in organizational settings, which are highlysocial environments where norms are constantlycommunicated, signaled, and learned (Bettenhausen& Murnighan, 1985). In the following sections, weintegrate BVT with social information processingtheory (Salancik & Pfeffer, 1978) to suggest thata leader’s sense of humor sends two important im-plicit messages to followers. First, it signals theacceptability of norm violations, which will bepositively associated with follower deviance. Sec-ond, it signals relational openness and permissive-ness, which will be positively associated with highquality LMX that drives work engagement. Finally,we suggest that these mediated relationships aremoderated by a particular humor style—aggressivehumor.

Integrating Benign Violation and SocialInformation Processing Theories

Benign violation theory suggests that humor in-volves violations of norms. However, in order tounderstand the implications and outcomes of humorin organizational settings, BVT must be integratedwith theories that are specific to the dynamicsinherent in the workplace (Heath & Sitkin, 2001).According to social information processing theory(Salancik & Pfeffer, 1978), employees do not operatein a vacuum in the workplace. Rather, employeesactively seek to understand and behave in congru-ence with the norms and expectations within theirorganizationsbyprocessing surrounding social cues.From a social information processing perspective,leaders are seen as role models, or symbols of the“way things are done,” and guide the ways followersorganize and make sense of their environment(Shamir, House, & Arthur, 1993). Thus, leaders’ ac-tions and interpersonal styles can send powerfulmessages and social cues to followers. Throughprocessing these cues, followers learnwhat behavioris expected, rewarded, and punished in the organi-zation (Hogg, 2010). The processing of this in-formation also helps followers reduceuncertainty by

enabling them to make sense of the normative envi-ronment in the workplace (Van den Bos, 2001).

Social information processing suggests that fol-lowers create cognitive representations of leaderbehavior in specific interactions as indicators of theexpectations and values of the organization. Thesecognitive representations—referred to as scripts(Gioia &Manz, 1985), schemas (James& Jones, 1976),or molar perceptions (Schneider, 1990)—are sym-bolic and general, rather than specific. As followersattempt to make meaning of their environment, theywill look to leaders’ behaviors in specific situationsas gestalt representations of basic principles to beapplied across multiple situations (James, Hater,Gent, & Bruni, 1978). Thus, leader behavior sym-bolizes “the way things are around here” (Zohar &Luria, 2004: 322; see also Schneider & Reichers,1983, Zohar, & Tenne-Gazit, 2008).

Thus, when integrated with benign violation the-ory, social information processing theory suggeststhat when leaders display a sense of humor in in-teractions with followers, the implications are be-yond simple mimicry (i.e., followers understandthat humor is expected and rewarded and will be-have accordingly); a meaningful message is con-veyed about the values of the organization: behavingcounter-normatively “is the way things are done.”We argue that leaders who display a sense of humorduring interpersonal interactions with followerscommunicate two important signals. The first, thatviolating norms is acceptable, can have deleteriousconsequences in terms of follower deviance. Thesecond, that the leader and followers are in a per-missive exchange relationship, can have positiveeffects on follower work engagement. We begin bydiscussing the path to deviance.

Implications for Norm Violation Acceptability andWorkplace Deviance

In organizations, norms can be institutionalized asformal rules (e.g., organizational codes of conduct) oras informal perceived descriptive norms (e.g., benice to your coworkers; Morris, Hong, Chiu, & Liu,2015; Yam, Chen, & Reynolds, 2014). When leadersdisplay humor and, as a result, violate norms, fol-lowers will likely perceive that it is socially accept-able to violate norms in the organization for tworeasons. First, leaders’ formal position makes themstrong sourcesof normative expectations. Leaders, asrole models, are more likely to be observed by fol-lowers who are scanning the environment for in-formation on how to behave in the work context


(Robinson & O’Leary-Kelly, 1998). In other words,leaders who make light of norm violations in orderto produce humor are likely to imply to followersthat mild norm violations in the organizations aregenerally acceptable. Second, when a leader acts ina humorous manner, others will likely react withlaughter and amusement, an implicit signal of ap-proval (Billig, 2005). Followers will be likely to in-terpret this social information as signaling theacceptability of norm violations. When a norm vio-lation is enacted—and interpreted by others—ina playful, humorous way, it also signals to followersthat violations need not be taken seriously or scru-tinized (Ford, 2000).

Extant empirical research provides some supportfor this hypothesis. For example, Ford,Wentzel, andLorion (2001) found that men who were previouslyexposed to sexist humor later believe that there is anincreased tolerance of violating sexism norms com-pared to men who were not exposed to such humor.In another study, participants who were exposed tojokes targeted at outgroups were more likely to en-dorse outgroup prejudice compared to participantsin the control condition (Hodson, Rush, &MacInnis,2010). Interestingly, both studies found that sexist orracist comments not conveyed in a humorous man-ner did not have the same effects, suggesting thathumor expands the bound of appropriate behaviorsby creating new norms that tolerate behaviors typi-cally viewed as deviant. We thus hypothesize:

Hypothesis 1. Leader sense of humor is posi-tively associated with followers’ perceived ac-ceptability of norm violations

Through its effects on perceived norm violationacceptability, leader humor is likely to increase fol-lower deviance. Workplace deviance, by definition,involves the violation of organizational norms thatthreaten the well-being of an organization and/or itsmembers (Bennett & Robinson, 2000; Robinson &Bennett, 1995). Deviance can generally be classifiedalong two dimensions—interpersonal and organi-zational deviance. Examples of interpersonal de-viance include making fun of someone at work andacting rudely toward coworkers, whereas examplesof organizational deviance include taking propertyfrom work and coming in late to work without per-mission (Bennett & Robinson, 2000). Although thetwo dimensions of deviance focus on differenttargets, in this research we focus on overall work-place deviance. This approach is consistent withprior research (e.g., Christian & Ellis, 2011; Judge,Scott, & Ilies, 2006; Lee & Allen, 2002) as the two

dimensions are often highly correlated and haveoverlapping antecedents (Berry, Ones, & Sackett,2007; Hershcovis et al., 2007).

As followers learn that it is acceptable to violatenorms in an organization as a result of leader’s senseof humor, they are more likely to engage in deviantbehavior. The key to this is the socially constructedperception that norm violations are acceptable. Forexample, although bribery violates the law in someAsian countries, it is a socially acceptedwayof doingbusiness and people engage in this practice becausetheybelieve that suchbehavior iswidely tolerated bytheir peers, leaders, and organizations (Buffalo &Rodgers, 1971). Also consistent with this idea is thatrole models who display antisocial behavior havea stronger influence on antisocial behavior amongother individuals in work groups than group mem-berswhoarenot considered rolemodels (Robinson&O’Leary-Kelly, 1998). Leaders are role models in or-ganizations given their formal position and status,and are thus likely to influence not only the per-ceptions (i.e., acceptability of norm violation), butalso the behaviors (i.e., deviance) of their followers.In addition, to the extent that violating norms isperceived as socially acceptable, followers are morelikely to engage in these behaviors because they be-lieve that thesebehaviorswill gounpunishedandaresupported by their leaders. For example, employeesare more likely to engage in unethical behaviors inorganizations that have poorly conceived rules andcommunity codes (Peterson, 2002; Yam, Reynolds,& Hirsh, 2014). We thus hypothesize:

Hypothesis 2. The perceived acceptability ofnorm violations mediates the relationship be-tween leader sense of humor and follower work-place deviance.

STUDY 1

Participants and Procedure

We contacted 340 part-time Master of BusinessAdministration (MBA) students to participate in thisresearch. All of our participants worked full-timeandwere enrolled in aMBAprogrampart-time in theevenings or weekends at a large university in CentralChina. In our initial contactwith the participants, weprovided a general overview of the research (e.g., athree-wave study, organizational behavior research)but didnot disclose any specific researchhypothesesto participants. A total of 215 participants (Mage 530.75, Mtenure with leader 5 3.44 years, 56.9% male)completed all three waves of the surveys, yielding


a response rate of 63.24%. This relatively high re-sponse rate was due to consistent communicationbetween the author team and the participants, aswell as monetary compensation. Our participantscame from a variety of industries, most notably fromthe banking, finance, manufacture, and governmentsectors.

Each of the three waves was separated by roughlytwo weeks. At Time 1, participants completed a mea-sure of leader sense of humor. At Time 2, participantscompleted a measure of perceived acceptability ofnorm violation. At Time 3, participants self-reportedtheir own interpersonal and organizational devi-ance. All survey items were translated from English toMandarin, and then back translated following estab-lished best practices for survey translation (Brislin,1970). Each participant was compensated with 20RMB (3 USD) for their time per survey.

Measures

Leader sense of humor. We measured leadersense of humorwith a seven-item scale adapted fromThorson and Powell (1993). Since the original scalewas designed formeasuring recognition of oneself asa humorous person, we reworded the items to reflectperceptions of others’ sense of humor for the purposeof this study. A sample item is “my leader uses hu-mor to entertain coworkers.”All of the items for thisscale are listed in the Appendix (1 5 strongly dis-agree to 7 5 strongly agree; a 5 .96).2

Perceived acceptability of norm violations. Wemeasured perceived acceptability of norm viola-tions with a five-item scale developed by Van Kleef,Homan, Finkenauer, Gundemir, and Stamkou(2011). Participants were asked to what extent theythought it acceptable for a person in the organiza-tions to be “asocial,” “immoral,” “improper,”“rude,” and “well-mannered (reverse-coded)” (1 5not at all acceptable to 75highly acceptable;a5 .77).

The five adjectives were chosen because they repre-sent broad social norms observed in daily lives andacross many different organizations.

Deviance. We measured deviance with a well-established 19-item scale (Bennett & Robinson,2000). We used this particular scale to measure de-viance because it has been used in prior researchwith Chinese participants (Fehr, Yam, He, Chiang, &Wei, 2017), ensuring high content validity. Partici-pants self-reported on both interpersonal (e.g., “madefun of someone at work”) and organization deviance(e.g.,“takenproperty fromworkwithoutpermission”)that theyhad committed on a frequency scale (15 notat all to 55 quite often,a5 .96).Weused a self-reportrather than an other-report measure of deviance be-cause many workplace deviant behaviors are oftendone in private without the knowledge of coworkersor leaders. In addition, deviance is commonly self-reported in prior research (for a review, see Berryet al., 2007).3

RESULTS

Preliminary Analyses

Descriptive statistics for Study 1 are presented inTable 1.

Prior to hypothesis testing, we conducted a seriesof confirmatory factor analyses (CFAs) to ensure thedistinct factor structure of our three key variables.The hypothesized three-factor model composed ofleader sense of humor, perceived acceptability ofnorm violations, and deviance demonstrated a goodfit to the data, x2 (116)5 310.85, RMSEA5 .08, CFI5.94. This three-factor model was also superior to al-ternative models, including a two-factor model inwhich leader sense of humor and deviance were setto load on a single factor (Δx2(2) 5 765.78, p , .01,RMSEA 5 .19, CFI 5 .71); a two-factor model inwhich perceived acceptability of norm violationsand deviance were set to load on a single factor(Δx2(2) 5 674.34, p , .01, RMSEA 5 .19, CFI 5 .74);a two-factor model in which leader sense of humorand perceived acceptability of norm violations wereset to load on a single factor (Δx2(2)5 654.01, p, .01,RMSEA 5 .18, CFI 5 .75).

2 To provide empirical evidence for the valence-freenature of our leader sense of humor scale,we conducted anMTurk study with 100 participants. Specifically, we pre-sented all seven items to them in random order. Afterreading each item, participants were asked “Please recallthe most recent incident in which someone does this atyour workplace. To what extent would you consider thetype of humor he/she used as: –2 5 negative, 0 5 neutral,2 5 positive.” A one-sample t-test (comparing the mean tozero) suggested that these behaviors are neither negative norpositive (M5 .07, t [99]5 1.32, p5 .19). These results sug-gest that our leader sense of humor scale is valence-free.

3 We tested ourhypotheses by treating interpersonal andorganizational deviance as a single construct for pre-sentational parsimony in both Studies 1 and 2. Resultsremained identical when the two types of deviance weretreated as two separated dependent variables.


FollowingBecker (2005),we ran all of our analyseswith and without demographic controls and the re-sults were essentially identical with the inclusion ofthese variables. Moreover, past humor research hasfailed to find meaningful differences in terms ofa sense of humor across gender or age (Moran, Rain,Page-Gould, & Mar, 2014; Thorson & Powell, 1993).For the purpose of presentational parsimony, wethus present the results without controls but providetheir bivariate correlations with the study variablesin Table 1.

Tests of Hypotheses

We used ordinary least squares regression to testHypothesis 1. Leader sense of humor at Time 1 waspositively associatedwith perceived acceptability ofnormviolations at Time 2 (adjustedR25 .05,b5 .23,p, .01, Table 2). To test Hypothesis 2,we conducteda bootstrapping-based mediation test using thePROCESSmacro (Hayes, 2013). This procedure is anextension of the Sobel test (Sobel, 1982) and is rec-ommended over alternative procedures (e.g., Baron& Kenny, 1986) because it does not assume a normalsampling distribution of indirect effects, and simu-lation studies have shown that it is more valid and

statistically powerful than traditional methods(MacKinnon, Lockwood, & Williams, 2004). AsPreacher and Hayes (2008) recommended, we esti-mated the indirect effect of leader sense of humor onfollower deviance via perceived acceptability ofnorm violation using unstandardized coefficientsand a bootstrapping procedure with 5,000 resam-ples to produce a 95% confidence interval aroundthe estimated indirect effects. The bootstrappedindirect effect is significant if the bias-corrected95% confidence interval (CI) excludes zero. Resultsrevealed that leader sense of humor was associatedwith increased follower deviance, mediated byperceived acceptability of norm violation in theorganization (indirect effect5 .04, SE5 .02, 95%CI5 .01 to .09; direct effect 5 .08, SE 5 .03, 95% CI 5.01 to .15; total effect5 .12, SE5 .03, 95%CI5 .05 to.19). Together, these results provide support forHypotheses 1 and 2.

Study 1 Discussion

In Study 1,we found support for Hypotheses 1 and2 by conducting a three-wave field study in China.Leader sense of humor, perceived acceptability ofnorm violations, and deviance were all temporally

TABLE 1Correlation Table for Study 1

Variables Means (SD) 1 2 3 4 5

1. Leader sense of humor (T1) 4.42 (1.57) .962. Acceptability of norm violations (T2) 1.89 (1.18) .23** .773. Deviance (T3) 1.57 (.84) .22** .33** .964. Follower age (T1) 30.75 (5.54) .01 .05 –.05 (–)5. Follower gendera (T1) 1.57 (.50) .10 .15* .13 .16* (–)

*p , .05**p , .01

a 15 Female, 2 5MaleNote: Alphas are presented on the diagonal.

TABLE 2Regression Analyses for Study 1

DV5 Perceived acceptability ofnorm violation DV5 Deviance

Variables B SE B b B SE B b

Leader sense of humor .17 .05 .23** .08 .04 .15*Perceived acceptability of norm violation .21 .05 .30**Adjusted R2 .05 .12

*p , .05**p , .01


separated in order to reduce common methodeffects (Podsakoff, MacKenzie, Lee, & Podsakoff,2003). Nevertheless, Study 1 is incomplete in that itonly tested the negative path of leader sense of hu-mor to follower outcomes. In the following sections,we argue that humor can be a “mixed blessing ,”suggesting that although leader sense of humormight be positively associated with follower de-viance, it can also lead to positive follower out-comes such as LMX and work engagement. Wefurther argue that a leader’s humor style moderateseach path. We then conducted a second study toexamine the full model.

A Mixed Blessing: Implications for LMX andWork Engagement

Social information processing theory and BVTalso suggest a positive path for leader humor forthree reasons. First, humor reduces social distancebetween a leader and followers (Graham, 1995;Mesmer-Magnus et al., 2012). Leaders who use hu-mor implicitly send a signal that they deemphasizethe hierarchical difference between themselvesand followers because they are willing to violatethe normal hierarchical system in organizations(Cooper, 2008). Second, when leaders interact ina manner that violates norms, this signals that theleader ispermissive—acceptingof counter-normativebehavior—with their followers, thus signaling thattheir relationship with particular followers is un-complicated, playful and open. Indeed, research inbehavioral ethics suggests that leaderswho aremoral,thus less likely to violate ethical norms, are perceivedto be less warm and permissive (Wellman, Mayer,Ong, & DeRue, 2016). Third, a leader who displayshumor is demonstrating a willingness to be vul-nerable because they are openly violating norms,and thus likely seem less guarded and more openduring social exchanges with followers. By model-ing vulnerability, a leader implies that the norma-tive interaction between the leader and follower isone that is amicable and safe. Followers who seetheir leaders as de-emphasizing hierarchy, permis-sive, and willing to display vulnerability may per-ceive the leader to be more relationship-oriented(Cooper, 2008; Decker & Rotondo, 2001), leading tothe sense that the leader2follower relationship is ofhigh quality. As such, leader humor is perceived tohave positive relational intent and serves to main-tain positive work relationships with followers(Gkorezis, Petridou, & Xanthiakos, 2014; Pundt &Venz, 2017).

More generally, extant research suggests thatpeople who exhibit a sense of humor are more en-joyable to be around as they increase the positiveaffect of people around them. For example, humor-ous people enjoy more successes in forming roman-tic relationships (Bressler & Balshine, 2006) andfriendships (Kalbfleischl, 2013). Specific to theworkplace, humor has been theorized to be an ef-fective ingratiation tactic for gaining social capitaland establishingpositive interpersonal relationships(Cooper, 2005). As a result, we suggest that leaderswho display a sense of humor around their followerswill create higher levels of LMX (Robert, Dunne, &Iun, 2016), which is an indication of the relationshipquality between a leader and follower (Graen &Scandura, 1987; Graen & Uhl-Bien, 1995).

Hypothesis 3. Leader sense of humor is posi-tively associated with LMX.

Through its effects on LMX,we suggest that leaderhumor is likely to increase follower work engage-ment. Work engagement is defined as a relativelyenduring state of mind referring to the simultaneousinvestment of personal energies in work perfor-mance (Christian, Garza, & Slaughter, 2011). En-gagement is a conceptually and empirically uniquemotivational construct that is evoked when an em-ployee is able to develop a feeling of self-investmentof their personal resources at work (Kahn, 1990).Work engagement thus involves a holistic use ofcognitive, emotional, and physical energy (Rich,Lepine, & Crawford, 2010), and empirical studiessuggest that follower work engagement can be cul-tivated by leader behavior, including LMX (Matta,Scott, Koopman, & Conlon, 2015; for a meta-analyticreview, see Christian et al., 2011).

In order to personally invest their full energy atwork, an employeemust feel comfortable and safe toexpress themselves (Kahn, 1990; May, Gilson, &Harter, 2004). A high quality of LMX as a result ofa leader’s sense of humor will ensure employees feelsafe to be their true self, which in turn enables themto fully invest their personal energy in their workroles. Furthermore, when a leader displays humor,the resultant high quality LMX provides social sup-port and reduces the negative effects of stress. In-deed, humor is associated with an optimisticreinterpretation of negative events and can assist incoping via increased social support (Ziv, 1981). Al-though no studies have examined the role of leaderhumor on follower work engagement, there is ev-idence that humor might reduce burnout (Abel,2002)—a close antipode to work engagement


(e.g., Schaufeli, Salanova, Gonzalez-Roma, & Bakker,2002). In sum, we propose that leader humor will in-crease LMX quality and, through heightened LMX, fol-lowers would feel more engaged and invest theirpersonal energy in their work.

Hypothesis 4. Leader sense of humor is posi-tively associated with follower work engage-ment, mediated by increased LMX.

The Moderating Role of Leader Aggressive Humor

Thus far, Hypotheses 124 provide an account ofthe effects of leader’s sense of humor on followerdeviance and work engagement. Nonetheless, thereare reasons to believe that these effects will partlydepend on the specific style of humor that a leaderdisplays. Integrating BVT and social informationprocessing theorywith research ondifferent styles ofhumor (Chen & Martin, 2007; Martin et al., 2003;Vernon, Martin, Schermer, & Mackie, 2008), we ar-gue that leader aggressive humor will exacerbatethe negative effects of a leader’s sense of humor onfollower deviance and attenuate the positive ef-fects of a leader’s sense of humor on follower workengagement.

Whereas a sense of humor represents a generaltendency to display any style of humor (Martin,2001), traditional conceptualizations of aggres-sive humor refer to a more specific style of hu-mor that is aimed at teasing or ridiculing (Martinet al., 2003; also known as disparagement humor;Zillmann, 1983). In organizations, it is relativelyrare to observe excessively aggressive humoraimed solely at ridiculing others (e.g., racist jokesaimed at a follower; Romero & Cruthirds, 2006).Rather, leaders often usemild aggressive humor inthe workplace, which includes sarcasm, satire, orteasing to humorously convey disapproving in-formation to their followers (e.g., “John, you seemso busy” to sarcastically convey a message thatJohn is not putting up enough effort). Similar toa leader’s sense of humor, in this research we donot focus on whether a leader uses aggressivehumor toward a specific follower or not, butrather a leader’s general tendency in using thisstyle of humor toward all followers of his/herworkgroup.

We suggest that a leader’s sense of humor signalsmore severe norm violations when the leader’s hu-mor style is aggressive compared to other styles ofhumor. Displaying a sense of humor coupledwith anaggressive humor style effectively violates normsin two ways: it not only involves using humor—a

benign violation—but it also signals violation of so-cial norms of civility. Aggressive humor signals tofollowers that the accepted social norm of being re-spectful toward others is not important. Thus,a leader who has this kind of sense of humor andtends to be aggressive communicates (a) that violat-ing organizational norms is acceptable for reasonsthat we have outlined above, but added to this is (b)the signal that violating norms of “human decency”is acceptable. Theoretically, compared to otherforms of humor, Martin et al. (2003) categorized ag-gressive humor as a form of hostile behavior. Forexample, sarcasm, a form of aggressive humor, isoften associated with increased interpersonal con-flict in a work group because it conveys disapproval,contempt, and scorn humorously (Huang, Gino, &Galinsky, 2015; Zhang & Liao, 2015). Given its hos-tile interpersonal nature, we expect that the malignnormviolation, signaled byaggressivehumor, buildson the benign violation of the nature of a leader’ssense of humor, exacerbating the perception of theseverity of acceptability of violations of norms. Thus,we argue that a leader who has a strong sense of hu-mor and tends to use aggressive humor would signalto followers an even stronger perceived acceptabilityof norm violation in the workplace, which ulti-mately leads to more behavior that is deviant. Wehypothesize:

Hypothesis 5. The indirect effect of leader senseof humor on follower deviance, via perceivedacceptability of norm violation, is moderated byleader aggressive humor such that the indirecteffect is stronger when leader aggressive humoris high, but weakens when leader aggressivehumor is low.

Similarly,we suggest that a leaderwhohas a strongsense of humor and often uses aggressive humor isless likely to build effective leader2follower re-lationships, ultimately attenuating thepositive effectof a leader’s sense of humor on follower work en-gagement. Aggressive humor, even in its mild form,is often carried out at the expense of the leader’s re-lationship with his/her followers. At the dyadiclevel, even if an aggressive joke is perceived as hu-morous, it remains harmful to interpersonal re-lationships (Toplak & Katz, 2000). Research suggeststhat well-meaning remarks delivered in an aggres-sive form (e.g., saying “your work ethic is obviouslypoor” sarcastically to someone who clearly workedoptimally) is often less well received compared tothe same intent, but delivered with more sincerecontent (e.g., “yourwork ethic is obviously excellent”;


Pexman & Olineck, 2002). More generally, aggressivehumor is perceived as more contemptuous (Dews &Winner, 1995) and is more likely to instigate conflict(Huang et al., 2015) than other forms of humor. Asa result, the effects of leader sense of humor on LMXwill be attenuated when a leader uses aggressivehumor.

The attenuating effects of aggressive humor styleson the relationship between leader humor and fol-lower LMX may also reach beyond the targeted fol-lower, providing a social interpretation that affectsbehavior (Salancik & Pfeffer, 1978). Research hassuggested that followers often react to their leadersnegatively when leaders treat their peers poorly(Mitchell, Vogel, & Folger, 2015), even when thefollower is well treated (Christian, Christian, Garza,&Ellis, 2012;He, Fehr,Yam,Long,&Hao, 2017). Thisstreamof research suggests that the positive effects ofleader sense of humor on interpersonal outcomeswill be significantly reduced even by merely ob-serving leaders who use aggressive humor stylestoward other followers, rather than experiencingleader aggressive humor first-hand. We suggestwhen a leader has a strong sense of humor, strongLMX is less likely to be formed if the leader tends touse aggressive humor. We thus hypothesize:

Hypothesis 6. The indirect effect of leader senseof humor on follower work engagement, viaLMX, is moderated by leader aggressive humorsuch that the indirect effect is stronger whenleader aggressive humor is low, but weakenswhen leader aggressive humor is high.

STUDY 2

Participants and Procedure

We contacted 700 full-time employees to partici-pate in this research through Qualtrics, a third-partyonline survey administration company in theUnitedStates (for recent examples of data collection usingQualtrics, see DeCelles, DeRue, Margolis, & Ceranic,2012; Long, Bendersky, & Morrill, 2011). In our ini-tial contact with the participants, we provideda general overview of the research (e.g., a three-wavestudy, organizational behavior research) but did notdisclose any specific research hypotheses to partic-ipants as in Study 1. We also informed participantsthat they must have daily interaction with theirleaders in order to participate. A total of 288 partic-ipants agreed to participate in the study and 200participants (Mage 5 43.26, Mtenure with leader 5 5.93years, 40% male; 72.5% Caucasian, 10% Asian

American, 8.5% Hispanic American) completed allthree waves of surveys, yielding a response rate of28.57%. Most of these participants worked in sales(24.5%), banking and financial services (17.5%), andengineering (15%).

Each of the three waves was separated by roughly10days.AtTime 1, participants completedmeasuresof leader sense of humor and leader aggressive hu-mor. At Time 2, participants completed measures ofperceived acceptability of norm violation and LMX.At Time 3, participants self-reported their own de-viance and work engagement. Each participant wascompensated with $5 for their time per survey andreceived a $10 bonus if they completed all threesurveys.

Measures

Leader sense of humor. We measured leadersense of humor as in Study 1 (15 strongly disagree to7 5 strongly agree; a 5 .97).

Leader aggressive humor. We measured leaderaggressive humor with an eight-item scale adaptedfromMartin et al. (2003). Since the original scalewasdesigned for self-report rather than other-report, wereworded the items to reflect perceptions of theleaders’ aggressive humor for the purpose of thisstudy. A sample item is “my leader’s sense of humoroften offends others.” All of the items for this scaleare listed in the Appendix (1 5 strongly disagree to7 5 strongly agree; a 5 .76).

Perceived acceptability of norm violation. Wemeasured perceived acceptability of norm violationas in Study 1 (1 5 strongly disagree to 7 5 stronglyagree; a 5 .80).

Leader member exchange. We measured LMXwith the well-established eight-item measure de-veloped by Bauer and Green (1996). Sample itemsinclude “I usually know how satisfied my leader iswithme” and “I knowwhere I standwithmy leader”(15 strongly disagree to 75 strongly agree; a5 .96).

Deviance. We measured deviance as in Study 1(15 strongly disagree to 75 strongly agree; a5 .91).

Work engagement. We used Rich et al. (2010)’s18-item scale to measure work engagement. Workengagement is theorized to have three sub-dimensions—physical (e.g., “I work with intensityonmy job”), emotional (e.g., “I am enthusiastic inmyjob”), and cognitive engagement (e.g., “Atwork, I paya lot of attention to my job”). As our theorizing didnot distinguish the three different types of engage-ment and because they were highly correlated (rsranged from .72 to .87, ps, .01), we used the average


score to form an overall composite of work engage-ment (1 5 strongly disagree to 7 5 stronglyagree; a 5 .91).

RESULTS

Preliminary Analyses

Descriptive statistics for study variables and con-trols for Study 2 are presented in Table 3.

We again conducted a series of CFAs to ensure thedistinct factor structure of our key variables. Thehypothesized six-factor model (i.e., leader sense ofhumor, leader aggressive humor, perceived accept-ability of norm violation, LMX, deviance, and workengagement) demonstrated a good fit to the data,x2(309) 5 699.95, RMSEA 5 .08, CFI 5 .91, and wasalso superior to three alternative five-factor models,including a five-factormodel inwhich leader sense ofhumor and leader aggressive humor were set to loadon a single factor (Δx2(5)5 202.21, p, .01, RMSEA5.10, CFI5 .87); a five-factormodel inwhich LMX andperceived acceptability of norm violation were set toload on a single factor (Δx2(5) 5 763.03, p , .01,RMSEA 5 .14, CFI 5 .74); and a five-factor model inwhich deviance and work engagement were set toload on a single factor (Δx2(5) 5 1353.64, p , .01,RMSEA 5 .17, CFI5 .61). Finally, to ensure that ourtemporal separationdidnot affect the factor structure,we further compared our six-factor model to a three-factor model in which the variables were grouped bytime (i.e., leader sense of humor and aggressive hu-mor; LMX and perceived acceptability of norm vio-lation; work engagement and deviance). Results

suggested that the six-factor model demonstrated su-perior fit to this model (Δx2(12) 5 2311.12, p , .01,RMSEA5 .21, CFI5 .40)

As with Study 1, we ran all of our analyses withand without demographic controls and the resultswere essentially identical with the inclusion of thesevariables (Becker, 2005). For the purpose of pre-sentational parsimony, we present the results with-out any controls.

Test of Hypotheses

As in Study 1, we conducted ordinary leastsquares regression to test Hypothesis 1. Leader senseof humor at Time 1 was positively associated withperceived acceptability of norm violations in theorganization at Time 2 (b 5 .27, p , .01). To testHypothesis 2, we conducted a bootstrapping-basedmediation test using the PROCESS macro (Hayes,2013). Results revealed that leader sense of humorwas associated with increased follower deviance,mediated by perceived acceptability of norm viola-tion in the organization (indirect effect 5 .02, SE 5.01, 95%CI5 .002 to .05;direct effect5 .03,SE5 .02,95% CI 5 2.02 to .08; total effect 5 .05, SE 5 .02,95% CI 5 .01 to .10). These results provide addi-tional support for Hypotheses 1 and 2.

To examine the positive pathwayof leader sense ofhumor and Hypotheses 324, we again conducteda bootstrapping-based mediation test using thePROCESS macro (Hayes, 2013). First, we found thatleader sense of humor at Time 1 was positively as-sociated with LMX at Time 2 (b 5 .16, p , .05). Re-sults further revealed that leader sense of humorwas

TABLE 3Correlation Table for Study 2

Variables Means (SD) 1 2 3 4 5 6 7 8 9

1. Leader sense of humor (T1) 3.88 (1.60) .972. Leader aggressive humor (T1) 3.40 (.98) .06 .763. Accept. of norm violations (T2) 2.22 (1.04) .28** .15* .804. LMX (T2) 5.14 (1.41) .15* –.12† –.04 .965. Deviance (T3) 1.48 (.55) .15* .05 .21** –.131 .916. Work engagement (T3) 5.49 (1.19) .23** –.04 .01 .19** –.14* .987. Follower age (T1) 43.26 (13.13) –.33** –.13† –.13† –.07 –.07 –.17* (–)8. Follower gendera (T1) 1.60 (.49) –.03 .04 .12† –.11 –.12† –.04 –.07 (–)9. Follower raceb (T1) 1.28 (.45) .07 .09 –.09 .00 –.09 .11 –.08 –.09 (–)

†p , .10*p , .05

**p , .01a 15 Male, 25 Femaleb 15 White, 25 Non-white

Note: Alphas are presented on the diagonal.


associated with increased work engagement, medi-ated by increased LMX (indirect effect 5 .02, SE 5.02, 95%CI5 .003 to .06;direct effect5 .15,SE5 .05,95% CI5 .05 to .26; total effect5 .17, SE5 .06, 95%CI 5 .07 to .28), providing support for Hypotheses 3and 4.

To test Hypothesis 5, we began by examining theinteractive effect of leader aggressive humor andleader sense of humor on perceived acceptability ofnormviolation. In Step 1, both leader sense of humor(b 5 .27, p , .01) and leader aggressive humor (b 5.14, p , .05) were positively associated with per-ceived acceptability of norm violation (Table 4). InStep 2, results suggested that after the inclusion ofthe interaction term the model explained signifi-cantly more variance (adjusted R2 5 .13; ΔR2 5 .04,p, .01) and that the interaction termwas significant(b 5 .20, p , .01). To aid interpretation, we plottedthe interaction effect in Figure 2 and it was in theexpected direction. We then utilized the methods ofHayes (2013) to test Hypothesis 5 in an integrativefashion at one standard deviation above and belowthe mean of the moderator (i.e., leader aggressivehumor).When leader aggressivehumorwashigh, themediated model was significant (conditional in-direct effect 5 .03, SE 5 .02, 95% CI 5 .002 to .07).When leader aggressive humor was low, however,the mediated model was not significant (conditionalindirect effect 5 .01, SE 5 .01, 95% CI 5 2.003 to.03). The index ofmoderatedmediationwas likewisesignificant (Index 5 .02, SE 5 .01, 95% CI 5 .001 to.04), providing full support forHypothesis 5. In otherwords, although leader sense of humor is generallyassociated with increased deviance, this effect isstronger when leader aggressive humor is high butdissipates when leader aggressive humor is low.

We followed the same procedure in testing Hy-pothesis 6. In Step 1, leader sense of humor (b5 .16,p , .05) was positively associated with LMX,whereas leader aggressive humor (b52.13, p, .10)was negatively (albeit marginally) associated withLMX (Table 5). In Step 2, results suggested that afterthe inclusion of the interaction term the modelexplained significantlymore variance (adjustedR25.05; ΔR2 5 .02, p, .05) and that the interaction termwas significant (b 5 .14, p , .05). To aid in-terpretation, we plotted the interaction effect inFigure 3 and it was in the expected direction. Wethen utilized the methods of Hayes (2013) to testHypothesis 6 in an integrative fashion at one stan-dard deviation above and below the mean of themoderator (i.e., leader aggressive humor). Whenleader aggressive humor was high, the mediated

model was not significant (conditional indirect ef-fect52.00, SE5 .02, 95% CI52.03 to .04). Whenleader aggressive humor was low, however, themediated model was significant (conditional in-direct effect 5 .04, SE 5 .02, 95% CI 5 .02 to .09).The index of moderated mediation was likewise sig-nificant (Index 5 2.02, SE 5 .01, 95% CI 5 2.06to2.003), providing full support forHypothesis 6 (seeTable 6). In other words, although leader sense ofhumor is generally associated with increased workengagement, this effect is stronger when leader ag-gressive humor is low but dissipates when leaderaggressive humor is high. Together, our results sug-gested that leader sense of humor is indeed a mixedblessing and revealed the important moderating roleof leader aggressive humor.4 For purposes of pre-sentational parsimony, we have also presented allresults in a path model (see Figure 4).

GENERAL DISCUSSION

Across two three-wave field studies in China andthe United States, we found consistent support forour hypotheses that leader sense of humor is amixedblessing. On the one hand, it signals to followers theacceptability of normviolations in the organizations,which in turn is positively associated with followerdeviance. On the other hand, it increases LMX,which in turn is positively associated with followerwork engagement. We further demonstrated thatthese mediated effects were moderated by leaders’humor styles. Specifically, humorous leaders whotend to use aggressive humor were most likely topromote follower deviance and least likely to en-courage follower work engagement. Below, we dis-cuss the theoretical and practical contributions ofour work and suggest future research directions.

Theoretical Implications

Our research makes a number of important theo-retical contributions to the literature on leadershipandhumor.Humor is pervasive in theworkplace, yetresearch on humor in organizations has almost beennon-existent for several decades. In fact, the samecould be said in other disciplines such as psychology

4 We controlled for LMX and work engagement whentesting Hypothesis 5 and controlled for perceived accept-ability of norm violation and deviance when testingHypothesis 6. Adding these controls to the model didnot affect the findings and hence we presented theuncontrolled analyses for presentational parsimony.


and sociology. For example, sociologist MurrayDavis (1995) once characterized the study of humoras a “stillborn field.”We suspect that these sporadicresearch attemptswere partly due to a lack of a coherenttheoretical framework. By introducing benign viola-tion theory (McGraw & Warren, 2010) to the organiza-tional literature, we believe that future researchers cansystematically examine the effects of humor in worksettings. Perhaps more importantly, whereas benignviolation theory was originally developed to explain“what makes things funny,” we integrate this theorywith social information processing theory and thus ex-tend BVT by applying it to understand the negative consequences of observing and processing leader humor.

Our study also contributes to research on therelational nature of leadership. In particularly, our“mixed blessing” approach demonstrated that theeffects of leader humor are more nuanced thanpreviously assumed. By taking an initial step to-ward examining the negative outcomes of leaderhumor to their followers, we found that leaderhumor is associated with greater follower de-viance. By drawing from past work on the re-lational benefits of humor (e.g., Romero &Cruthirds, 2006), we found that leader humor isassociated with increased follower work en-gagement. Thus, our results indicate that al-though leaders may improve the quality of their

TABLE 4Regression Analyses for the Negative Path of Humor (Study 2)

DV5 Perceived acceptability of norm violation

Model 1 Model 2


LMX –.05 .05 –.07 –.03 .05 –.04Leader sense of humor .18 .04 .27** .19 .04 .29**Aggressive humor .14 .07 .14* .13 .07 .12†

Leader sense of humor .13 .04 .20**X Aggressive humor adjusted R2 .09 13DR2 .04**

†p , .10*p , .05

**p , .01

FIGURE 2The Interactive Effect of Leader Sense of Humor and Aggressive Humor on Acceptability

of Norm Violation (Study 2)

1

2

3

4

5

6

7

Acc

epta

bili

ty o

f N

orm

Vio

lati

on

LowAggressiveHumor

HighAggressiveHumor

High Leader Sense of HumorLow Leader Sense of Humor


relationships with followers and followers’ workengagement using humor, they may also un-intentionally increase perceptions of the accept-ability of norm violations and deviance. Ourresearch thus provides a more complete un-derstanding of the effects of leader humor andhighlights the need for a dialectical perspective onthe consequences of humor usage in organizations.

Third, our study extends the extant literature onworkplace humor by examining style. Whereas pastresearch on leader humor often aggregated varioustypes of humor (Avolio, Howell, & Sosik, 1999), wesuggest that sense of humor is too broad to capturethe nuances of outcomes, and that style of humor is

a more specific factor that should be considered inunderstanding the consequences of leader humor.As our results suggest, whereas sense of humor wasassociated positively with deviance and work en-gagement (i.e., the two mediated effects), aggressivehumoramplifies ornullifies sucheffects. In addition,past research on specific humor styles tends to focuson positive forms of humor (for a review, seeMesmer-Magnus et al., 2012) and our work is one ofthe few that establishes the importance of a negativeform of humor, aggressive humor, as a moderatorbetween the links of leader sense of humor and fol-lower outcomes. Therefore, our work contributes tothehumor literaturebroadlybyexploring tendencies

TABLE 5Regression Analyses for the Positive Path of Humor (Study 2)

DV 5 LMX

Model 1 Model 2


Perceived acceptability of norm violation –.10 .10 –.07 –.05 .10 –.04Leader sense of humor .14 .06 .16* .13 .06 .15*Aggressive humor –.18 .10 –.13† –.17 .10 –.12†

Leader sense of humor .14 .06 .16*X Aggressive humor adjusted R2 .03 .05DR2 .02*

†p , .10*p , .05

FIGURE 3The Interactive Effect of Leader Sense of Humor and Aggressive Humor on LMX (Study 2)

1

2

3

4

5

6

7

Low Leader Sense of Humor

LM

X LowAggressiveHumor

HighAggressiveHumor

High Leader Sense of Humor


to use humor in tandem with aggressive humor. Webelieve this approach can best inform scholars andpractitioners when leader humor is most and leastbeneficial to their followers.

Finally, our study contributes to work on organi-zationalnorms (Morris et al., 2015) by suggesting thatnorm violations may be seen as domain general andtherefore have far-reaching effects on behavior.In both of our studies, we found that leader humorwas associated with a range of different deviant

behaviors that might not conceptually map on to thespecific forms of leader humor. Although past workon humor and normative beliefs suggests that hu-morous norm violations may transfer within do-main, such that sexist jokes lead others to perceivethat sexism is acceptable (e.g., Ford et al., 2001), ourstudy extends this idea by suggesting that norm vi-olations may be interpreted more broadly than thespecific norm being violated by a joke or a funnystory. Leader humor may lead employees to form

TABLE 6Summary of Indirect Effects and Conditional Indirect Effects

Paths & effects Estimates SE 95% confidence intervals

Leaders’displays of humor→Perceivedacceptabilityof norm violation→ Follower deviance

Indirect effects .02 .01 [.002, .051]Moderated mediationHigh aggressive humor .03 .02 [.003, .082]Low aggressive humor .01 .01 [–.003, .035]

Indirect difference .03 .02 [.002, .082]Leaders’ displays of humor→ LMX→ Work

engagementIndirect effects .02 .01 [.001, .060]Moderated mediationHigh aggressive humor –.00 .02 [–.032, .032]Low aggressive humor .04 .02 [.006, .103]

Indirect difference –.04 .03 [–.121,2.004]

FIGURE 4Path Model of the Results (Study 2)

Leader sense ofhumor

Perceivedacceptability ofnorm violation

Leader aggressive humor

Followerdeviance

Leader memberexchange

Workengagement .15*

.10†

.03 .01

.01 .19**

.13*

.13*

–.09

.10

.18**

–.14*

Unstandardized path estimates are reported. Solid lines depict the hypothesized relationships and dashed lines indicate relationships thatare not hypothesized.

† p , .10*p , .05

**p , .01


a gestalt perception of norm violation acceptabilitythat overarches the normative landscape of theorganization.

Practical Implications

Although successful leaders are often able to usehumor to motivate their followers to achieve greaterperformance (e.g., Avolio et al., 1999; Vecchio,Justin, & Pearce, 2009), our findings call attentionto a potential risk that humorous leaders should beaware of. To be clear, we are not suggesting thatleaders should stop displaying humor in the work-place. Instead, as our mixed blessing model dem-onstrates, leaders must strive to embody the valuesthey possess while simultaneously creating an en-vironment in which norm adherence is encouraged.We suggest that one caveat lies in the type of humorbeing displayed—aggressive humor harms relation-ship quality and elicits more norm violating behav-iors from followers. Therefore, we encourage leadersto continue expressing humor, but to minimize theusage of aggressive humor as much as possible. Ad-mittedly, it can be difficult to dictate what kind ofhumor a leader should display. Organizations caneducate leaders on the potentially negative conse-quences of aggressive humor and encourage theusage of more positive forms of humor throughtraining (Prerost, 1993). Through appropriate train-ing, we believe that humorous leaders can reap thebenefits (i.e., more engaged followers) while mini-mizing the negative consequences (i.e., deviantfollowers). Indeed, compared to other structuralpolicies (e.g., employee retreats designed to increaseengagement), leader sense of humor could be a cost-less strategy in creating a more relaxed and engagedworkplace.

Although leaders can be trained to reduce theirusage of aggressive humor (Prerost, 1993) and hencereduce follower deviance, our research revealedboth main and indirect effects of leader sense of hu-mor on follower deviance even when aggressivehumor is not taken into account. Therefore, we sug-gest that it is equally important to socialize em-ployees to espouse the normative values in anorganization and refrain from engaging in deviantbehavior. One way to achieve this is by having a for-mal code of conduct for workplace interactionsamong colleagues. Another way to achieve this is byreinforcing identification with the organization.When employees identify themselves as an integralpart of the organization, they will be less likely toengage in behaviors that would potentially harm the

organization or its members (Rousseau, 1998). Withstrong socialization programs for newcomers anda culture that promotes identification with the or-ganization, the effects of leaders’ displays of humoron follower deviance may be mitigated.

The broader implication of our research is thatleaders need to be mindful of their status as rolemodels. Leaders need to be aware that their actionsserve as social cues for their followers and can causeboth positive and negative consequences. Therefore,leaders should be careful in how they portraythemselves to their followers in their interactions.Leaders should seek to increase their self-monitoringskills, such that they will be more aware of whattypes of humor are appropriate in differentsituations.

Limitations and Future Research Directions

Although there are several strengths of the currentwork, such as our effort to replicate the findingsacross two distinct cultures using multi-wave de-signs, several limitations warrant further discussionand future research. First and foremost, although inboth studies we adapted a validated measure to as-sess leader sense of humor (Thorson&Powell, 1993),thismeasuremight not havebeenvalence-free in thata few items appear to capture not just a leader’spredisposition of using humor, but also her/his pre-disposition of using positive humor. Although thenull correlation between leader sense of humor andaggressive humor (r5 .06,p5 .39) in Study 2 andoursupplemental study suggest that our measure ofleader sense of humor indeed is valence-free, westrongly recommend future research to develop new,rigorous humor assessments, specifically in thecontext of organizational studies.

Second, we have only examined one boundarycondition of the link between leader sense of humorand follower outcomes, namely leader aggressivehumor. Future research can examine whether otherstyles of leader humor might moderate the effects ofleader sense of humor on follower outcomes. Forexample, self-deprecating humor might be the beststyle of humor indevelopinghigh-qualityLMXgivenits non-threatening nature (Martin et al., 2003).Given that a complete view of the leadership processinvolves not just leader characteristics, but alsofollower characteristics (Howell & Shamir, 2005;Kellerman, 2008), we encourage future research toexamine additional follower characteristics as po-tential moderators. For example, followers who arehigh in perspective taking (Galinsky & Moskowitz,


2000) might understand the higher-order messagesthat humorous leaders are signaling (e.g., a relaxingworkplace rather than mere norm violations) andhence be less likely to engage in deviance.

Third, humor research in organizational behavioris rather sporadic (Duncan et al., 1990). Althoughweexamined follower deviance and work engagement,we encourage future research to continue to explorethe additional consequences of leader humor andtheir underlying mechanisms. On the positive side,given the norm violating nature of most humor,a leader’s sense of humormight facilitate followers tothink “outside of the box” and achieve greater crea-tivity (Cooper, 2008). On the negative side, followersmightmimic leaders’displays of aggressivehumor totheir peers, thus leading to increased group re-lational conflict and decreased group cohesion. Forthese negative behavioral consequences, it wouldbe useful to examine both other- and self-reportedsources because self-reported negative behaviors(e.g., deviance) might suffer from biases associatedwith social desirability (Berry et al., 2007) or moti-vated forgetting (Kouchaki & Gino, 2016), a limita-tion of the current set of studies.

Fourth, given the correlational nature of our fieldstudies, we were not able to definitively establishcausal inferences. To partially offset concerns of re-verse causality,we followed recentwork (e.g.,Matta,Scott, Colquitt, Koopman, & Passantino, 2016) tocompare the values of Akaike’s Information Crite-rion (AIC) and Bayesian Information Criterion (BIC)for our hypothesized model and alternative, reservecausal models. According to Kline, the model withthe smallest AIC and BIC fits the data best and is “theone most likely to replicate” (Kline, 2011: 220). Re-sults suggested that thehypothesizedmodel inStudy1 (Humor → Perceived acceptability of norm viola-tion → Deviance) had lower AIC and BIC values(AIC 5 1129.83, BIC 5 1153.43) compared to thereverse causal model (Deviance → Perceived ac-ceptability of norm violation → Humor: AIC 51402.28, BIC 5 1425.86). Similarly, the two hy-pothesized models (positive and negative paths ofleader sense of humor) in Study 2 had lower AIC andBIC values compared to the reverse causal models(Hypothesized positive path: AIC 5 1336.16, BIC 51359.25 vs. reverse causal positive path: AIC 51452.85, BIC 5 1475.91; Hypothesized negativepath: AIC 5 894.78, BIC 5 917.87 vs. reverse causalnegative path: AIC 5 1323.02, BIC 5 1346.11). Al-though these results reinforce the validity in the di-rectionality of our findings, we recommend futureresearch to conduct experimental investigations of

our proposed mechanisms. For example, futureresearch could create simulated leader2followerinteractions and manipulate, rather than measure,leader humor.

Relatedly, becausewewere not able to definitivelyestablish causal inferences, there might be other al-ternative mediating factors that could explain ourfindings. As Zhao, Lynch, and Chen (2010) sug-gested, complementary mediation is likely whenboth the direct and indirect effects are statisticallysignificant and in the same direction (i.e., the pathfrom leader sense of humor to deviance in Study 1and the path from leader sense of humor to workengagement in Study 2). We therefore suggest thatthere might be other complementary mediatingmechanisms derived from other theories that aredriving our results and recommend future researchto examine these mechanisms more closely.

Fifth, our measure of deviance in both studies isself-reported in nature. While many deviant behav-iors are done in private and hence an other-reportformat would not have been applicable, we suggestfuture research control for factors such as social de-sirability biases that might likely create a floor effectfor this measure.

Finally, although norm violation and deviance areoften perceived as negative in many organizations,some organizations may explicitly encourage suchbehaviors. In industries such as entertainment ortourism, employees are often required to displayhumor. In such industries, a leader’s sense of humormay not signal norm violation, but rather norm ad-herence. In other words, organizational or industrialand societal norms sometimes might be in conflict,highlighting the important moderating role of in-dustry. Future research should examine the impli-cations for leader’s sense of humor in industrieswhere humor is explicitly encouraged.

CONCLUSION

In this research, we integrated benign violationand social information processing theories andfound that leader’s sense of humor is a mixed bless-ing. It leads to increased follower deviance and in-creased work engagement. Although the currentresearch sheds some light on the link betweenleader’s sense of humor and follower outcomes, werecognize that we have only taken one step towarda more complete understanding of the effects ofleader humor in organizations; many questions re-main for future studies to tackle. Nonetheless, wehope the introduction of benign violation theory and


the current work can spark additional research onhumor in organizational behavior.

REFERENCES

Abel, M. H. 2002. Humor, stress, and coping strategies.Humor: International Journal of Humor Research,15: 365–381.

Avolio, B. J., Howell, J.M., & Sosik, J. J. 1999. A funny thinghappened on the way to the bottom line: Humor asa moderator of leadership style effects. Academy ofManagement Journal, 42: 219–227.

Baron, R., & Kenny, D. 1986. The moderator–mediatorvariable distinction in social psychological research:Conceptual, strategic, and statistical considerations.Journal of Personality and Social Psychology, 51:1173–1182.

Bauer, T. N., & Green, S. G. 1996. Development of leader-member exchange: A longitudinal test. Academy ofManagement Journal, 39: 1538–1567.

Becker, T. E. 2005. Potential problems in the statisticalcontrol of variables in organizational research: Aqualitative analysis with recommendations. Organi-zational Research Methods, 8: 274–289.

Bennett, R. J., & Robinson, S. L. 2000. Development ofa measure of workplace deviance. The Journal ofApplied Psychology, 85: 349–360.

Berry, C. M., Ones, D. S., & Sackett, P. R. 2007. Interpersonal deviance, organizational deviance, and theircommon correlates: A review and meta-analysis. TheJournal of Applied Psychology, 92: 410–424.

Bettenhausen, K., & Murnighan, J. 1985. The emergenceof norms in competitive decision making groups.Administrative Science Quarterly, 30: 350–372.

Billig, M. 2005. Laughter and ridicule: Towards a socialcritique of humor. London: Sage Publications.

Bressler, E. R., & Balshine, S. 2006. The influence of humoron desirability. Evolution and Human Behavior, 27:29–39.

Brislin, R. W. 1970. Back-translation for cross-cultural re-search. Journal of Cross-Cultural Psychology, 1:185–216.

Buffalo, M., & Rodgers, J. W. 1971. Behavioral norms,moral norms, and attachment: Problems of devianceand conformity. Social Problems, 19: 101–113.

Chen, G. H., & Martin, R. A. 2007. A comparison of humorstyles, coping humor, and mental health betweenChinese and Canadian university students. Humor:International Journal of Humor Research, 20:215–234.

Christian, J. S., Christian, M. S., Garza, A. S., & Ellis, A. P.2012. Examining retaliatory responses to justice

violations and recovery attempts in teams. The Jour-nal of Applied Psychology, 97: 1218–1232.

Christian,M. S., &Ellis,A. P. 2011. Examining the effects ofsleep deprivation on workplace deviance: A self-regulatory perspective. Academy of ManagementJournal, 54: 913–934.

Christian, M. S., Garza, A. S., & Slaughter, J. E. 2011. Workengagement: A quantitative review and test of its re-lations with task and contextual performance. Per-sonnel Psychology, 64: 89–136.

Colquitt, J. A., & Zapata-Phelan, C. P. 2007. Trends in the-ory building and theory testing: A five-decade study ofthe Academy of Management Journal. Academy ofManagement Journal, 50: 1281–1303.

Cooper, C. 2008. Elucidating the bonds of workplace hu-mor: A relational process model. Human Relations,61: 1087–1115.

Cooper, C. D. 2005. Just joking around? Employee humorexpression as an ingratiatory behavior. Academy ofManagement Review, 30: 765–776.

Crawford, C. B. 1994. Theory and implications regardingthe utilization of strategic humor by leaders. TheJournal of Leadership Studies, 1: 53–67.

Dampier, P., &Walton,A. 2013.WhiteHousewit,wisdom,and wisecracks. London: Barzipan Publishing.

Davis, M. S. 1995. The sociology of humor: A stillbornfield? Sociological Forum, 10: 327–339.

Decker, W. H., & Rotondo, D. M. 2001. Relationshipsamong gender, type of humor, and perceived leadereffectiveness. Journal of Managerial Issues, 13:450–465.

DeCelles, K. A., DeRue, D. S., Margolis, J. D., & Ceranic,T. L. 2012. Does power corrupt or enable? When andwhy power facilitates self-interested behavior. TheJournal of Applied Psychology, 97: 681–689.

Dews, S., &Winner, E. 1995. Muting themeaning: A socialfunction of irony.Metaphor and Symbol, 10: 3–19.

Duncan, W. J. 1982. Humor in management: Prospects foradministrative practice and research. Academy ofManagement Review, 7: 136–142.

Duncan, W. J., Smeltzer, L. R., & Leap, T. L. 1990. Humorand work: Applications of joking behavior to man-agement. Journal of Management, 16: 255–278.

Fehr, R., Yam, K. C., He, W., Chiang, J., & Wei, W. 2017.Polluted work: A self-control perspective on air pol-lution appraisals, organizational citizenship, andcounterproductive work behavior. OrganizationalBehavior andHumanDecision Processes. Publishedonline ahead of print.

Ford, T. E. 2000. Effects of sexist humor on tolerance ofsexist events. Personality and Social PsychologyBulletin, 26: 1094–1107.


Ford, T. E., Wentzel, E. R., & Lorion, J. 2001. Effects ofexposure to sexist humor on perceptions of normativetolerance of sexism. European Journal of SocialPsychology, 31: 677–691.

Galinsky, A. D., & Moskowitz, G. B. 2000. Perspective-taking: Decreasing stereotype expression, stereo-type accessibility, and in-group favoritism. Journalof Personality and Social Psychology, 78:708–724.

Gioia, D. A., & Manz, C. C. 1985. Linking cognition andbehavior: A script processing interpretation of vicari-ous learning. Academy of Management Review, 10:527–539.

Gkorezis, P., Petridou, E., & Xanthiakos, P. 2014. Leaderpositive humor and organizational cynicism: LMX asa mediator. Leadership and Organization Develop-ment Journal, 35: 305–315.

Goel, V., & Dolan, R. J. 2001. The functional anatomy ofhumor: Segregating cognitive and affective compo-nents. Nature Neuroscience, 4: 237–238.

Graen, G. B., & Scandura, T. A. 1987. Toward a psychologyof dyadic organizing. Research in –OrganizationalBehavior, 9: 175–208.

Graen, G. B., & Uhl-Bien, M. 1995. Relationship-basedapproach to leadership: Development of leader-member exchange (LMX) theory of leadership over25 years: Applying a multi-level multi-domain per-spective. The Leadership Quarterly, 6: 219–247.

Graham, E. E. 1995. The involvement of sense of humor inthe development of social relationships. Communi-cation Reports, 8: 158–169.

Hayes,A. F. 2013. Introduction tomediation,moderation,and conditional process analysis: A regression-based approach. New York: Guilford Press.

He, W., Fehr, R., Yam, K. C., Long, L. R., & Hao, P. 2017.Interactional justice, leader2member exchange, andemployee performance: Examining the moderatingrole of justice differentiation. Journal of Organiza-tional Behavior, 38:537–557.Published online aheadof print.

Heath, C., & Sitkin, S. B. 2001. Big‐B versus Big‐O:What isorganizational about organizational behavior? Jour-nal of Organizational Behavior, 22: 43–58.

Hershcovis, M. S., Turner, N., Barling, J., Arnold, K. A.,Dupre, K. E., Inness,M., LeBlanc,M.M.,&Sivanathan,N. 2007. Predicting workplace aggression: A meta-analysis. The Journal of Applied Psychology, 92:228–238.

Hodson, G., Rush, J., & MacInnis, C. C. 2010. A joke is justa joke (except when it isn’t): Cavalier humor beliefsfacilitate the expression of group dominance motives.Journal of Personality and Social Psychology, 99:660–682.

Hogg, M. A. 2010. Influence and leadership. In S. T. Fiske,D. T. Gilbert, & G. Lindzey (Eds.),Handbook of socialpsychology: 1166–1207. Hoboken, New Jersey: JohnWiley & Sons.

Howell, J. M., & Shamir, B. 2005. The role of followers inthe charismatic leadership process: Relationships andtheir consequences. Academy of Management Re-view, 30: 96–112.

Huang, L., Gino, F., & Galinsky, A. D. 2015. The highestform of intelligence: Sarcasm increases creativity forboth expressers and recipients. Organizational Be-havior and Human Decision Processes, 131: 162–177.

James, L. R., Hater, J. J., Gent, M. J., & Bruni, J. R. 1978.Psychological climate: Implications from cognitivesocial learning theory and interactional psychology.Personnel Psychology, 31: 783–813.

James, L. R., & Jones, A. P. 1976. Organizational structure:A review of structural dimensions and their con-ceptual relationships with individual attitudes andbehavior. Organizational Behavior and HumanPerformance, 16: 74–113.

Judge, T. A., Scott, B. A., & Ilies, R. 2006. Hostility, jobattitudes, and workplace deviance: Test of a multi-level model. The Journal of Applied Psychology, 91:126–138.

Kahn, W. A. 1990. Psychological conditions of personalengagement and disengagement at work. Academyof Management Journal, 33: 692–724.

Kalbfleischl, P. I. 2013. Interpersonal communication:Evolving interpersonal relationships. Oxford: Psy-chology Press.

Katz, J. 1996. Families and funny mirrors: A study of thesocial construction and personal embodiment ofhumor. American Journal of Sociology, 101:1194–1237.

Kellerman, B. 2008. Followership: How followers arecreating change and changing leaders. Boston, MA:Harvard Business School Press.

Kline, R. B. 2011. Principles and practice of structuralequation modeling. New York: Guilford.

Kouchaki, M., & Gino, F. 2016. Memories of unethical ac-tions become obfuscated over time. Proceedings ofthe National Academy of Sciences of the UnitedStates of America, 113: 6166–6171.

Lee, K., & Allen, N. J. 2002. Organizational citizenshipbehavior and workplace deviance: The role of affectand cognitions. The Journal of Applied Psychology,87: 131–142.

Long, C., Bendersky, C., & Morrill, C. 2011. Fairnessmonitoring: Linking managerial controls and fairnessjudgments in organizations. Academy of Manage-ment Journal, 54: 1045–1068.


MacKinnon, D. P., Lockwood, C. M., & Williams, J. 2004.Confidence limits for the indirect effect: distributionof the product and resampling methods.MultivariateBehavioral Research, 39: 99–128.

Matta, F. K., Scott, B., Colquitt, J., Koopman, J., & Passan-tino, L. 2016. Is consistently unfair better than spo-radically fair? An investigation of justice variabilityand stress. Academy of Management Journal. Pub-lished online ahead of print.

Matta, F. K., Scott, B.A., Koopman, J., &Conlon, D. E. 2015.Does seeing “eye to eye” affect work engagement andorganizational citizenship behavior? A role theoryperspective on LMX agreement. Academy of Man-agement Journal, 58: 1686–1708.

Malone, P.B. 1980.Humor:Adouble-edged tool for today’smanagers? Academy of Management Review, 5:357–360.

Martin, R. A. 2001. Humor, laughter, and physical health:Methodological issues and research findings. Psy-chological Bulletin, 127: 504–519.

Martin, R. A., Puhlik-Doris, P., Larsen, G., Gray, J., & Weir,K. 2003. Individual differences in uses of humor andtheir relation to psychological well-being: Develop-ment of the Humor Styles Questionnaire. Journal ofResearch in Personality, 37: 48–75.

May, D. R., Gilson, R. L., & Harter, L. M. 2004. The psy-chological conditions of meaningfulness, safety andavailability and the engagement of the human spirit atwork. Journal of Occupational and OrganizationalPsychology, 77: 11–37.

McGraw, A. P., & Warren, C. 2010. Benign violationsmaking immoral behavior funny. Psychological Sci-ence, 21: 1141–1149.

McGraw, P., &Warner, J. 2014. The humor code: A globalsearch for what makes things funny. New York:Simon & Schuster.

McGraw, A. P., Warren, C., Williams, L. E., & Leonard, B.2012. Too close for comfort, or too far to care? Findinghumor in distant tragedies and close mishaps. Psy-chological Science, 23: 1215–1223.

McGraw, A. P.,Williams, L. E., &Warren, C. 2014. The riseand fall of humor: Psychological distance modulateshumorous responses to tragedy.Social Psychological& Personality Science, 5: 566–572.

Mesmer-Magnus, J., Glew, D. J., & Viswesvaran, C. 2012.A meta-analysis of positive humor in the workplace.Journal of Managerial Psychology, 27: 155–190.

Mitchell, M., Vogel, R., & Folger, R. 2015. Third parties’reactions to abusive supervision of coworkers. TheJournal of Applied Psychology, 100: 1040–1055.

Moran, J. M., Rain, M., Page-Gould, E., & Mar, R. A. 2014.Do I amuse you? Asymmetric predictors for humor

appreciation and humor production. Journal ofResearch in Personality, 49: 8–13.

Morris, M. W., Hong, Y., Chiu, C., & Liu, Z. 2015. Nor-mology: Integrating insights about social norms tounderstand cultural dynamics. OrganizationalBehavior and Human Decision Processes, 129:1–13.

Moon, H. 2001. The two faces of conscientiousness: Dutyandachievement striving in escalationof commitmentdilemmas. The Journal of Applied Psychology, 86:533–540.

Peterson, D. K. 2002. The relationship between unethicalbehavior and the dimensions of the ethical climatequestionnaire. Journal of Business Ethics, 41: 313–326.

Pexman, P. M., & Olineck, K. M. 2002. Understandingirony:Howdo stereotypes cue speaker intent? Journalof Language and Social Psychology, 21: 245–274.

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff,N. P. 2003. Common method biases in behavioralresearch: A critical review of the literature and rec-ommended remedies. The Journal of Applied Psy-chology, 88: 879–903.

Preacher, K. J., & Hayes, A. F. 2008. Asymptotic andresampling strategies for assessing and comparingindirect effects in multiple mediator models. Behav-ior Research Methods, 40: 879–891.

Prerost, F. J. 1993.A strategy to enhance humor productionamong elderly persons: Assisting in the managementof stress. Activities, Adaptation and Aging, 17:17–24.

Pundt, A., & Venz, L. 2017. Personal need for structure asa boundary condition for humor in leadership. Jour-nal of Organizational Behavior, 38: 872107.

Rich, B. L., Lepine, J. A., & Crawford, E. R. 2010. Job en-gagement: Antecedents and effects on job perfor-mance. Academy of Management Journal, 53:617–635.

Robert, C., Dunne, T., & Iun, J. 2016. The impact of leaderhumor on subordinate job satisfaction: The crucialrole of leader2subordinate relationship quality.Group & Organization Management, 41: 375–406.

Robinson, S. L., & Bennett, R. J. 1995. A typology of deviantworkplace behaviors: A multidimensional scalingstudy. Academy of Management Journal, 38: 555–572.

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

Romero, E. J., & Cruthirds, K.W. 2006. The use of humor inthe workplace. The Academy of Management Per-spectives, 20: 58–69.


Rousseau, D. M. 1998. Why workers still identify with or-ganizations. Journal of Organizational Behavior, 19:217–233.

Salancik, G. R., & Pfeffer, J. 1978. A social informationprocessing approach to job attitudes and task design.Administrative Science Quarterly, 23: 224–253.

Schaufeli, W. B., Salanova, M., Gonzalez-Roma, V., &Bakker, A. B. 2002. The measurement of engagementand burnout: A two sample confirmatory factor ana-lytic approach. Journal of Happiness Studies, 3:71–92.

Schneider, B. 1990. Organizational climate and culture.San Francisco, CA: Jossey-Bass.

Schneider, B., & Reichers, A. E. 1983. On the etiology ofclimates. Personnel Psychology, 36: 19–39.

Shamir, B., House, R. J., & Arthur, M. B. 1993. The moti-vational effects of charismatic leadership: A self-concept based theory. Organization Science, 4:577–594.

Sobel, M. E. 1982. Asymptotic confidence intervals forindirect effects in structural equation models. Socio-logical Methodology, 13: 290–312.

Thorson, J. A., & Powell, F. C. 1993. Relationships of deathanxiety and sense of humor. Psychological Reports,72: 1364–1366.

Toplak, Maggie, & Katz, Albert N. 2000. On the uses ofsarcastic irony. Journal of Pragmatics, 32: 1467–1488.

Van den Bos, K. 2001. Uncertainty management: The in-fluence of uncertainty salience on reactions to per-ceived procedural fairness. Journal of Personalityand Social Psychology, 80: 931–941.

Van Kleef, G. A., Homan, A. C., Finkenauer, C., Gundemir,S., & Stamkou, E. 2011. Breaking the rules to rise topower: How norm violators gain power in the eyes ofothers. Social Psychological & Personality Science,2: 500–507.

Vecchio, R. P., Justin, J. E., & Pearce, C. L. 2009. The in-fluence of leader humor on relationships betweenleader behavior and follower outcomes. Journal ofManagerial Issues, 21: 171–194.

Vernon, P. A., Martin, R. A., Schermer, J. A., & Mackie, A.2008.A behavioral genetic investigation of humor stylesand their correlations with the big-5 personality di-mensions. Personality and Individual. Differences: AJournal of Feminist Cultural Studies, 44: 1116–1125.

Wellman, N., Mayer, D. M., Ong, M., & DeRue, D. S. 2016.When are do-gooders treated badly? Legitimatepower, role expectations, and reactions to moral ob-jection in organizations. The Journal of AppliedPsychology, 101: 7832814.

Yam, K. C., Chen, X. P., & Reynolds, S. J. 2014. Ego de-pletion and its paradoxical effects on ethical decision

making. Organizational Behavior and Human De-cision Processes, 124: 204–214.

Yam, K. C., Reynolds, S. J., & Hirsh, J. B. 2014. The hungrythief: Physiological deprivation and its effect on un-ethical conduct. Organizational Behavior and Hu-man Decision Processes, 125: 123–133.

Zhang, Y., & Liao, Z. 2015. Consequences of abusive su-pervision: A meta-analytic review. Asia PacificJournal of Management, 32: 959–987.

Zhao,X., Lynch, J.G., &Chen,Q.2010.ReconsideringBaronandKenny:Myths and truths aboutmediation analysis.The Journal of Consumer Research, 37: 197–206.

Zillmann, D. 1983. Disparagement humor. In P. E.McGhee& J. H. Goldstein (Eds.), Handbook of humor re-search: 85–107. New York: Springer.

Ziv, A. 1981. Psychology of humor (Hebrew). Tel-Aviv:Yahdav.

Zohar, D., & Luria, G. 2004. Climate as a social-cognitiveconstruction of supervisory safety practices: Scripts asproxy of behavior patterns. The Journal of AppliedPsychology, 89: 322–333.

Zohar, D., & Tenne-Gazit, O. 2008. Transformationalleadership and group interaction as climate anteced-ents: A social network analysis. The Journal of Ap-plied Psychology, 93: 744–757.

KaiChiYam ([email protected]) is an assistant professorof management at the National University of Singapore.He received his PhD from the University of Washington.His research focuses on behavioral ethics, leadership, andhumor.

Michael S. Christian ([email protected]) is theSarah Graham Kenan Associate Professor of Organiza-tional Behavior at the Kenan–Flagler Business School atthe University of North Carolina at Chapel Hill. He com-pleted his PhD in management at the University of Ari-zona. Dr. Christian’s current research is focused on humanenergy, engagement, and self-regulation as well as on themotivational drivers of deviant and unethical behavior.

WuWei ([email protected]) is a professor at the Schoolof Economics andManagement atWuhanUniversity,China.He received his PhD from the School of Management atHuazhong University of Science and Technology. His re-search interests include team leadership, employee deviantbehavior, impression management, and nonmarket action.

ZhenyuLiao ([email protected]) is a post-doctoral researchfellow in the Olin Business School at the WashingtonUniversity in St. Louis. His research interests includeleadership complexities and dynamics, episodic leader–member interactions, and behavioral ethics.


mailto:[email protected]




Jared Nai ([email protected]) is an Assistant Professorof Organizational Behavior and Human Resources at theLee Kong Chian School of Business at Singapore Manage-ment University. He completed his PhD in Managementand Organization at the National University of Singapore.His current research interests include prosocial behavior,deviance, and advice seeking.

APPENDIX

Leader sense of humor (Studies 1 and 2)

1. Say things in such away as tomakepeople laugh2. Is regarded as someone of a wit by coworkers3. Uses humor to entertain coworkers4. Cracks people up with the things he/she says5. Can ease a tense situation by saying something

funny6. Can exert control over a group by uses of humor7. Say clever things that amuse others

Leader aggressive humor (Study 2)

1. If someone makes a mistake, my leader willoften tease them about it.

2. My leader’s sense of humor often offends others.3. When telling jokes or saying funny things, my

leader usually is not very concerned about howother people are taking it.

4. My leader does not like it when people usehumor as a way of criticizing or putting some-one down (reverse-coded).

5. Sometimes my leader thinks of something thatis so funny that he/she can’t stop himself/herself from saying it, even if it is not appro-priate for the situation.

6. My leader never participates in laughing atothers even if all others are doing it (reverse-coded).

7. If my leader doesn’t like someone, he/she oftenuses humor or teasing to put them down.

8. Even if something is really funny to my leader,he/shewill not laugh or joke about it if someonewill be offended (reverse-coded).



Copyright of Academy of Management Journal is the property of Academy of Managementand its content may not be copied or emailed to multiple sites or posted to a listserv withoutthe copyright holder's express written permission. However, users may print, download, oremail articles for individual use.

the mixed blessing of leader sense of humor:...

Documents