HUMOR AS CATALYST AND NEUTRALIZER OF LEADERSHIP EFFECTIVENESS

This study examines the effects of (in)consistent leadership behaviors in promoting (or suppressing) relevant work outcomes for temporary employees such as interns. Specifically, to better understand the drivers of internship effectiveness, we hypothesized that supervisor humor interacts with leadership style, sending implicit messages about the organizational and supervisory relationship, thus shaping interns’ attitudes and behaviors. Using a sample of 164 interns, we empirically examined the moderating effect of humor (affiliative and aggressive) on the relationship between leadership styles (transformatio-nal and laissez-faire), attitudes (satisfaction and stress), and behaviors (negligence and job acceptance intentions) using a two-wave research design. Our findings were consistent with the hypotheses, sug-gesting that humor needs to be tailored to leadership styles to predict interns’ attitudinal and behavioral responses, with different types of humor interacting differently across leadership styles. Implications for further research are discussed.


INTRODUCTION
Organizations face a challenging business environment, marked by diversity, complexity, and constant disruptions (Siggelkow & Rivkin, 2005). In this context, leaders must make decisions knowing that uncertainty and ambiguity are the new normal.
However, research shows that, sometimes, leaders' use of humor is inconsistent with their leadership style, that is, their adopted behavioral patterns to influence and motivate followers' behavior (Tremblay & Gibson, 2016;Vecchio, Justin, & Pearce, 2009). Such inconsistencies have important consequences on followers' attitudes and behaviors, since they expect leaders to act predictably and consistently (De Cremer, 2003;Mullen, Kelloway, & Teed, 2011). Thus, both leadership styles and humor are key to establishing supervisor-subordinate bonds in the workplace (Pundt & Herrmann, 2015), and when inconsistent with each other, they can trigger vexing mixed feelings and thoughts among followers that characterize ambivalence (Ashforth, Rogers, Pratt, & Pradies, 2014;Methot, Melwani, & Rothman, 2017).
Such ambivalent feelings are especially stronger for nonregular and precarious employees, such as interns. Interns are different from regular employees: they possess less selfefficacy and autonomy due to limited work experience (Wendlandt & Rochlen, 2008) and have ambiguous and often precarious career status (Rose, 2017). Consequently, interns are particularly susceptible to immediate supervisors directly responsible for introducing, accompanying, and guiding them throughout the internship period (Beenen, 2014;Kenny et al., 2015;Liu, Ferris, Xu, Weitz, & Perrewé, 2014;McHugh, 2017;Rose, Teo, & Connell, 2014). Research shows that leadership behaviors are instrumental in helping interns make sense of their new reality (D'Abate, Youndt, & Wenzel, 2009;Zhao & Liden, 2011) and clarify role expectations (Liu, Xu, & Weitz, 2011). In ambivalent situations, interns feel pulled in opposite directions, causing discomfort, disorientation, and apprehension. By impeding interns from making sense of the appropriate norms and building clear expectations, ambivalence may affect their propensity to commit to or disengage from the organization (Ashforth et al., 2014;Methot et al., 2017

LEADERSHIP AND THE USE OF HUMOR
Due to humor's relational and informational roles in interpersonal communication (Cooper et al., 2018;Pundt & Herrmann, 2015), humor is considered a crucial tool to help managers achieve their objectives (Cooper, 2008;Pundt, 2015;Tremblay & Gibson, 2016). In a recent empirical study with Brazilian executives, optimism and sense of humor were found to be two of the most relevant managerial competences of modern leaders (Sant'Anna, Campos, & Lótfi, 2012). Leaders seek humor, both contrived and spontaneous, to make sense of incongruities that inevitably exist in business settings (Hatch, 1997;Huang & Kuo, 2011). Further, humor is considered especially useful during organizational entry, when ideas, values, and tasks are shared massively and relationships are formed (Heiss & Carmack, 2012;Lynch, 2002;Sobral & Islam, 2015). Given humor's ability to stimulate rapid affective bonds (Robert & Wilbanks, 2012), leaders' use of humor may be particularly effective at early stages of employment or traineeship, condensing drawn-out relational processes between leaders and followers into rich momentary exchanges.
We define humor as a multifaceted construct reflecting a behavioral condition (i.e., expression of humor vs. a "sense" of humor) involving an intention to be amusing, either by verbal (jokes and witticisms) or non-verbal (visual and gestural) means (Cooper, 2005). Humor is thus fundamentally interpersonal, although not necessarily positive. Martin et al. (2003) (Martin et al., 2003).
In addition, affiliative humor provides important information that may be unstated in formal rules or difficult to express directly (Adelswärd & Oberg, 1998). Some information may be difficult or tacit but can be disseminated imperceptibly and effectively through humor (Gruner, 1999), leading to, among other things, increased acceptance of leader messages (Greatbatch & Clark, 2003;Zepeda, Franco, & Preciado, 2014). This informational function of humor is critical for newcomers, as they use humor to interpret and assimilate job expectations and organizational culture and develop new affiliations (Heiss & Carmack, 2012).
In contrast to affiliative humor, aggressive humor is maladaptive, hurtful to others, and counterproductive for building and maintaining high-quality leader-follower exchange relationships (Pundt & Herrmann, 2015). This is because it increases perceived social distance between leaders and followers (Kim, Lee, & Wong, 2016). While affiliative humor sends positive messages about an organization, aggressive humor can be used to sabotage or undermine organizational objectives (Fleming & Spicer, 2002). Such humor, far from signaling a safe atmosphere, acts to vent dissatisfaction (Sturdy & Fineman, 2001), sending the opposite message to newcomers and increasing organizational anxiety and stress (Huo, Lam & Chen, 2012). Aggressive humor can also hurt unintentionally, as what supervisors find "funny" may be unintentionally hurtful. This indirectness differentiates negative humor from related yet distinct concepts such as abusive supervision (Tepper, 2000).

INCONSISTENT LEADERS AND AMBIVALENT INTERNS
Humor can be an important tool for leadership; however, research shows that sometimes using humor can be inconsistent with the leadership style, that is, the leaders' set of behaviors to provide direction and motivate followers (Avolio et al., 1999;Tremblay & Gibson, 2016 The polarized nature of leadership and humor styles, roughly categorized as constructive or offensive, makes inconsistencies between these behaviors more salient (Tremblay & Gibson, 2016;Vecchio et al., 2009).
Leaders' inconsistency is found to be harmful to both followers and organizations (Tremblay & Gibson, 2016). However, while desirable, consistency is somewhat challenging for leaders constantly confronted by multifaceted goals and required to balance contradictory demands and play multiple roles . When manifested by leaders, inconsistent behaviors are sources of subordinates' vexing mixed feelings that characterize ambivalence (Ashforth et al., 2014). Ambivalence refers to holding opposing affective/cognitive orientations toward another, such as loving and hating the same person, simultaneously. The experience of ambivalence tends to be aversive and dysfunctional (Ashforth et al., 2014;Methot et al., 2017), even if non-conscious, driving responses aimed at warding off these undesirable feelings.
The (in)consistency of leaders' behaviors is particularly important for predictions regarding leadership effectiveness, as ambivalence experienced by interns likely triggers attitudinal and behavioral reactions. That is, humor buffers or amplifies the effects of leadership styles, and these moderating effects differ according to the form of humor (Robert & Wilbanks, 2012). Similar to the literature noting positive-negative differences in affect-related phenomena (Watson, Clark & Tellegen, 1984), affiliative humor should act primarily on positive attitudes, while aggressive humor should be associated with strain-related and negative attitudes. Specifically, we hypothesize that affiliative humor works mainly by interacting with leadership styles to augment positive intern attitudes (i.e., internship satisfaction), while offensive and aggressive humor interacts with leadership styles to increase anxiety-related negative attitudes (i.e., stress).

Transformational leadership and humor
Transformational leaders act as mentors and role models for employees, encouraging subordinates to transcend individual aspirations for the organization (Judge & Piccolo, 2004).
Transformational leadership helps interns experience lower  (Tremblay & Gibson, 2016). Accordingly, a mixture of constructive leadership and offensive humor inculcates in interns a sense of being simultaneously attracted to and repulsed by supervisors, thus fostering ambivalent feelings toward them (Ashforth et al., 2014).
This ambivalent condition is stressful and aversive for employees (Uchino et al., 2012); it is more likely to affect vulnerable groups such as interns, thus leading to negative outcomes.
H1a. Consistency between transformational leadership and affiliative humor augments transformational leaders' positive effects on satisfaction, leading to more positive outcomes (lower neglect and higher job acceptance intentions).
H1b. Inconsistency between transformational leadership and aggressive humor neutralizes transformational leaders' stress reduction effects, leading to more negative outcomes (greater neglect and lower job acceptance intentions).

Laissez-faire leadership and humor
While transformational leaders provide a vision for followers, laissez-faire leadership can be a detached leadership style where leaders evade responsibilities, fail to assist followers, and hesitate to take positions on important issues (Bass, 1985).
Affiliative humor by a laissez-faire supervisor reflects inconsistency, since positive affective signals in the negative leadership relationship context seem contradictory and conflicting (Tremblay & Gibson, 2016). Therefore, by mixing an unsupportive leadership style with affiliative humor, supervisors contribute to emergence of interns' ambivalent feelings and, consequently, reduction of satisfaction.
However, use of aggressive humor by laissez-faire supervisors is likely to be seen as consistent and predictable by interns (Uchino et al., 2012). Both the laissez-faire style and aggressive humor indicate disinterest in subordinates (Cooper, 2008) and likely reinforce each other's effects on the supervisorintern relationship. That is, aggressive humor likely diminishes the leaders' credibility further. In this scenario, such humor may be especially bad, promoting decidedly negative attitudes in interns and affecting their subsequent commitment to the organization.
H2a. Inconsistency between laissez-faire leadership and affiliative humor reinforces laissez-faire leaders' negative effects on satisfaction, leading to more negative outcomes (greater neglect and lower job acceptance intentions).
H2b. Consistency between laissez-faire leadership and aggressive humor reinforces laissez-faire leaders' stress generation effects, leading to more negative outcomes (greater neglect and lower job acceptance intentions).

Sample and procedure
Participants were recruited from career centers and internship programs of a major Brazilian private university in Rio de Janeiro, Brazil. Brazil has the advantage of an extensive history of internships in higher education, typically involving students on sponsored projects during their final years before graduation. In
We selected four items to represent transformational leadership subscales' core dimensions: idealized influence or charisma

Humor
Supervisors' humor was measured using subscales for affiliative and aggressive humor from the Humor Styles Questionnaire (Martin et al., 2003). Respondents rated the frequency of a set of leader's behaviors on a five-point scale, ranging from 1 (never or very seldom) to 5 (very often). A sample item for the affiliative humor scale is, "My supervisor enjoys making people laugh," while that for aggressive humor scale is "If someone makes a mistake, my supervisor will often tease them about it." The affiliative and aggressive humor scales' reliability was 0.83 and 0.91, respectively.

Internship satisfaction
Intern

Job stress
Intern stress was assessed with a five-item scale adapted from Ivancevich and Matteson's (1980) Stress Diagnostic Survey using a five-point Likert-type scale, ranging from 1 (strongly disagree) to 5 (strongly agree). We selected items to represent the scale's core dimensions: role ambiguity, role conflict, quantitative and qualitative role overload, and career development concerns. Some examples of the items are "My job duties and work objectives are unclear to me" and "I work on unnecessary tasks and projects." The scale's reliability was 0.85.

Job acceptance intentions and negligent behavior
Intention to be permanently hired by the host company and negligent behaviors were both measured with three-item scales adapted from Leck and Saunders (1992

RESULTS
To verify the discriminant validity of our measures, we conducted a series of confirmatory factor analyses. First, we estimated fit indices of two baseline measurement models: (1) a four-factor model for the antecedents (variables measured at T1) and (2) a four-factor model for mediators and outcomes (variables measured at T2). Results showed a good model fit for both models: (1) χ2(160) = 260.3, p < 0.001; comparative fit index (CFI) = 0.94; root mean square error of approximation (RMSEA) = 0.062; and (2) χ2(129) = 221.1, p < 0.001; CFI = 0.94; RMSEA = 0.066. Second, we contrasted these two models' fit against that of alternative models. Baseline models showed a significantly better fit than alternative measurement models did, thus confirming that our measures are adequate to capture the intended constructs.  Step 1 Step 2 Step 1 Step 2

Controls
Leader prototypical behavior Note: Slopes at one standard deviation above the mean and one standard deviation below the mean.

Main predictors
Results also show the negative interaction of aggressive humor by laissez-faire leadership (b = -0.13, p< 0.05), which is corroborated by the relationship slopes between laissez-faire leadership and stress for high (simple slope =0.12, p > 0.10) and low (simple slope = 0.39, p < 0.01) aggressive humor, as shown in Graph 3. Contrary to our prediction, this finding suggests that supervisors' use of aggressive humor may neutralize, and not amplify as predicted, the negative impact of laissez-faire leadership on interns' stress. Note: Slopes at one standard deviation above the mean and one standard deviation below the mean.
Furthermore, we predicted that supervisors' humor (affiliative and aggressive) moderates (indirectly, through stress and satisfaction) the impact of leadership styles on interns' behaviors (negligence and job acceptance intentions)-this has been termed the conditional indirect effect or moderated mediation (Preacher, Rucker, & Hayes, 2007

LIMITATIONS AND FUTURE RESEARCH
Some limitations to this study must be acknowledged.
Besides the two-wave design to reduce the likelihood of common method bias, we controlled for prototypical leadership behavior to exclude the chance that the differential effects of leadership styles could be attributed solely to supervisors' prototypical characteristics. Moreover, since the focus of our study was interaction effects and not direct effects, this common method concern is less problematic. In fact, the presence of common methods bias is shown to reduce interaction estimates, leading to more conservative effects (Siemsen, Roth, & Oliveira, 2010).