Supervisory consequences of abusive supervision: An investigation of sense of power, managerial self-efficacy, and task-oriented leadership behavior

Supervisory consequences of abusive supervision: An investigation of sense of power, managerial self-efficacy, and task-oriented leadership behavior

Organizational Behavior and Human Decision Processes 154 (2019) 80–95 Contents lists available at ScienceDirect Organizational Behavior and Human De...

564KB Sizes 0 Downloads 19 Views

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

Contents lists available at ScienceDirect

Organizational Behavior and Human Decision Processes journal homepage: www.elsevier.com/locate/obhdp

Supervisory consequences of abusive supervision: An investigation of sense of power, managerial self-efficacy, and task-oriented leadership behavior

T



Dong Jua,1, Mingpeng Huangb,1, Dong Liuc, Xin Qind, , Qiongjing Hue, Chen Chend a

Business School, Beijing Normal University, China Business School, University of International Business and Economics, China c Scheller College of Business, Georgia Institute of Technology, USA d Sun Yat-sen Business School, Sun Yat-sen University, China e School of Management, Zhejiang University, China b

A R T I C LE I N FO

A B S T R A C T

Keywords: Abusive supervision Sense of power Managerialself-efficacy Diary method Experiment

While a large number of studies have shown the detrimental effects of abusive supervision on subordinates’ work attitudes and outcomes, little is known about how abusive supervision impacts supervisors themselves. Drawing upon self-perception theory and power-dependence theory, we take a unique actor-focused approach to examine how and when engaging in abusive supervisory behavior may benefit actors (i.e., supervisors). Specifically, we propose that abusive supervisory behavior is positively related to supervisors’ state sense of power, which in turn positively relates to their managerial self-efficacy and task-oriented leadership behavior. Furthermore, the relationship between abusive supervisory behavior and state sense of power and the positive indirect effect of abusive supervisory behavior on managerial self-efficacy via state sense of power are stronger for supervisors with low, rather than high, levels of chronic sense of power. Our hypotheses are substantially supported by a multi-wave field diary study (Study 1) conducted across 10 consecutive workdays and three experiments (Studies 2a, 2b, and 3). Moreover, supplementary analyses showed that abusive supervisory behavior was positively related to sense of power and managerial self-efficacy only in the short term (i.e., these relationships turned negative after one week). Our findings contribute to the abusive supervision literature by delineating a nuanced view of the supervisory outcomes of abusive supervision.

1. Introduction “Having a basis for power is not enough. The individual must act…” – Mintzberg (1983, p. 25) Abusive supervision—defined as hostile and non-physical behavior directed against subordinates (Tepper, 2000)—has attracted much research attention over the past two decades (Martinko, Harvey, Brees, & Mackey, 2013). A large number of studies have revealed that abusive supervision has negative impacts on subordinates’ attitudinal, affective, and behavioral outcomes, including reduced job satisfaction, organizational commitment, and in-role and extra-role performance, and increased psychological distress and deviant behaviors (for reviews, see Martinko et al., 2013; Tepper, 2007). However, the literature has largely neglected the effects of abusive supervision on supervisors themselves. Addressing this research topic is crucial for developing a more

complete understanding of how abusive supervision functions as doing so can provide theoretical insights into and empirical evidence for how supervisors may gain or suffer from abusing subordinates (Foulk, Lanaj, Tu, Erez, & Archambeau, 2018; Qin, Huang, Johnson, Hu, & Ju, 2018). In addition, both management scholars and practitioners have constantly emphasized the importance of leadership development (Day, Harrison, & Halpin, 2009; Lord & Hall, 2005). Knowledge of the potential consequences of abusive supervisory behavior for actors is helpful for designing leadership training programs to reduce abusive supervision in the workplace. For example, knowledge of the potential benefits supervisors derive from abusive supervisory behavior can be leveraged to explore other less detrimental ways to help supervisors gain the same benefits. Therefore, it is meaningful to examine whether, how, and when engaging in abusive supervisory behavior affects supervisors themselves. Taking an actor-centric (i.e., supervisor-centric) perspective, in this



Corresponding author at: Sun Yat-sen Business School, Sun Yat-sen University, No.135, Xinggang West Road, Guangzhou, Guangdong 510275, China. E-mail addresses: [email protected] (D. Ju), [email protected] (M. Huang), [email protected] (D. Liu), [email protected] (X. Qin), [email protected] (Q. Hu), [email protected] (C. Chen). 1 The first two authors contributed equally to this paper. https://doi.org/10.1016/j.obhdp.2019.09.003 Received 26 March 2018; Received in revised form 30 August 2019; Accepted 10 September 2019 Available online 17 September 2019 0749-5978/ © 2019 Elsevier Inc. All rights reserved.

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

test our theorized model (see Fig. 1), we conducted a field diary study (Study 1) and three experiments (Studies 2a, 2b, and 3). This research makes several major contributions to the abusive supervision and power literatures. First, our research contributes to the abusive supervision literature by uncovering an important potential benefit of abusive supervisory behavior for actors in terms of how they perceive themselves as managers. Previous studies on the consequences of abusive supervision have almost exclusively focused on the subordinate side (for exceptions, see Foulk et al., 2018; Qin et al., 2018) and have concluded that abusive supervision is almost universally detrimental (Mackey, Frieder, Brees, & Martinko, 2017). By linking abusive supervision to supervisors’ state sense of power and managerial self-efficacy, this research challenges this conclusion and suggests that in the short term, abusive supervision may actually have some benefits for supervisors as it is positively related to their positive self-perceptions as managers (i.e., state sense of power and managerial self-efficacy) and subsequent positive behaviors (i.e., task-oriented leadership behavior). Second, our research further extends the understanding of the consequences of abusive supervision for the supervisors by exploring when such behavior provides greater or fewer benefits for supervisors. Specifically, we argue that the positive indirect relationship between abusive supervisory behavior and managerial self-efficacy via state sense of power is stronger for supervisors with low (versus high) levels of chronic sense of power. In other words, from the power perspective, supervisors with low chronic sense of power are more likely to benefit from abusive supervisory behavior than those with high chronic sense of power. Research on the strategic use of abusive supervision (Khan, Moss, Quratulain, & Hameed, 2018; Tepper, Duffy, & Breaux-Soignet, 2012) has suggested that supervisors may abuse their subordinates with the intent to establish power over their teams. Our research extends these studies by revealing that abusive supervision may indeed allow supervisors with low chronic sense of power to experience stronger state sense of power. Third, we also contribute to the power literature by integrating selfperception theory and power-dependence theory to explore behavioral factors (i.e., abusive supervisory behavior) that affect state sense of power. Power-dependence theory posits that individuals’ power originates from other individuals’ dependence on them for valued resources (Emerson, 1962; Molm, 1991; Wee et al., 2017). Studies drawing upon this theory have often focused on objective control over resources that others value (e.g., Tepper et al., 2009). Our research extends the literature by shifting this focus to supervisors’ subjective perceptions of control over valued resources. Drawing upon self-perception theory, we propose that individuals’ sense of how much others depend on them is affected by behavioral factors (i.e., abusive supervisory behavior). In addition, we also extend previous research on the antecedents of sense of power (e.g., organizational positions; Anderson et al., 2012) by revealing that managers’ sense of power is affected by the way they treat their subordinates.

research, we draw upon self-perception theory (Bem, 1982; Casciaro, Gino, & Kouchaki, 2014) and power-dependence theory (Emerson, 1962; Molm, 1991; Wee, Liao, Liu, & Liu, 2017) to illuminate the process through which abusive supervision triggers supervisory outcomes. The core tenet of self-perception theory is that individuals’ perceptions of themselves are affected by their observations of their own overt behaviors (Bem, 1982). We argue that abusive supervisory behavior affects actors’ (i.e., supervisors’) perceptions of themselves as managers. In particular, we focus on two important managerial selfperceptions: state sense of power and managerial self-efficacy. Sense of power refers to individuals’ perceptions of their ability to influence others and is determined by the extent to which they perceive others as dependent on themselves (Anderson, John, & Keltner, 2012). Sense of power comes not only from formal hierarchical positions but also from behaviors that demonstrate individuals’ volition to control and influence others (Anderson et al., 2012; Schaerer, Tost, Huang, Gino, & Larrick, 2018). State sense of power describes individuals’ perceptions of power at a particular moment. Integrating self-perception theory and power-dependence theory, we propose that abusive supervisory behavior is positively related to supervisors’ perceptions that they control subordinate-valued resources and goals and that subordinates are dependent on them, which in turn is positively related to their state sense of power. Moreover, self-perception theory suggests that individuals’ behaviors that are salient to themselves are more likely to influence their self-perceptions (Bem, 1982; Casciaro et al., 2014). Thus, we argue that the positive relationship between abusive supervisory behavior and supervisors’ state sense of power is moderated by supervisors’ chronic sense of power (i.e., supervisors’ chronic perceptions of their ability to influence subordinates in their work teams), as it determines the extent to which supervisors view abusive supervisory behavior as salient actions demonstrating power (Anderson et al., 2012; Williams, Gruenfeld, & Guillory, 2017). Furthermore, self-perception theory also suggests that by observing their own behaviors (e.g., abusive supervision), managers may also form an assessment of how effective and competent they are as managers (Bem, 1982; Casciaro et al., 2014). The main construct that captures the extent to which managers perceive that they are effective and competent in fulfilling their managerial roles is managerial selfefficacy (Fast, Burris, & Bartel, 2014; Lu, Siu, & Cooper, 2005; Wood & Bandura, 1989). Thus, we further contend that abusive supervisory behavior has a positive indirect relationship with supervisors’ managerial self-efficacy via state sense of power. Managerial self-efficacy is an important aspect of supervisors’ self-perceptions that affects how managers fulfill their managerial responsibilities and display leadership behaviors (for a review, see Hannah, Avolio, Luthans, & Harms, 2008), such as task-oriented leadership behaviors (Fleishman, 1973; Judge, Piccolo, & Ilies, 2004). We further propose that managerial self-efficacy is positively related to subsequent task-oriented leadership behavior, which refers to leader behavior that clarifies follower expectations, roles, and responsibilities; is oriented toward getting work done; and establishes well-defined patterns and channels of communication (DeRue, Nahrgang, Wellman, & Humphrey, 2011; Judge et al., 2004; Lambert, Tepper, Carr, Holt, & Barelka, 2012). Task-oriented leadership behavior represents the fundamental day-to-day behavior that is important across all types of leaders (Fleishman, 1953; Yukl, 1971). To

2. Theoretical grounding and hypothesis development 2.1. Self-perception theory and power-dependence theory We draw upon self-perception theory (Bem, 1982; Casciaro et al.,

Fig. 1. Theoretical model. 81

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

environments or attitudes and influences others to do things that meet his or her own interests (e.g., Magee, 2009; Schaerer et al., 2018; van Kleef, Homan, Finkenauer, Gündemir, & Stamkou, 2011). As such, we hypothesize the following:

2014) and power-dependence theory (Emerson, 1962; Molm, 1991; Wee et al., 2017) to examine the potential effects of abusive supervision on supervisors in terms of how they perceive themselves as managers. Self-perception theory suggests that people make inferences about themselves “from observations of their own overt behavior and/or the circumstances in which this behavior occurs” (Bem, 1982, p. 5). For instance, engaging in instrumental networking behaviors in the workplace has been found to be positively related to feeling dirty (Casciaro et al., 2014). Accordingly, in light of self-perception theory, abusive supervisory behavior may shape how supervisors perceive themselves. However, the theory does not provide specific clues and rationale as to what specific self-perceptions may emerge from abusive supervisory behavior. Therefore, we further build on power-dependence theory (Emerson, 1962; Molm, 1991; Wee et al., 2017) to understand the relationship between abusive supervisory behavior and supervisors’ self-perceptions of power. Power-dependence theory posits that one’s power “resides implicitly in the other’s dependency” (Emerson, 1962, p. 32). Furthermore, actor A’s dependence on actor B is “(1) directly proportional to A’s motivational investment in goals mediated by B, and (2) inversely proportional to the availability of those goals to A outside of the A-B relation” (Emerson, 1962, p. 32). Accordingly, in a supervisor-subordinate relationship, the supervisor perceives power over the subordinate when the subordinate depends on the supervisor to obtain subordinate-valued resources and goals (e.g., favorable performance evaluations, desirable work conditions, salary increase, career-development opportunities; Wee et al., 2017). In sum, integrating these two theories leads to an overarching theoretical framework for understanding the effects of supervisors’ abusive supervisory behavior on their perceptions of power and subsequent behaviors.

Hypothesis 1. A supervisor’s engagement in abusive supervisory behavior is positively related to his or her state sense of power. 2.3. Abusive supervisory behavior, state sense of power, and managerial self-efficacy Self-efficacy refers to “beliefs in one’s capabilities to organize and execute the courses of action required to manage prospective situations” (Bandura, 1995, p.2). Self-efficacy is particularly important for individuals as it is positively related to perseverance “in the face of difficulties, the quality of their well-being and their vulnerability to stress and depression, and the choices they make at important decision points” (Bandura & Locke, 2003, p. 87). As a specific form of efficacy, managerial self-efficacy refers to the extent to which managers perceive that they are effective and competent in fulfilling their managerial roles (Fast et al., 2014; Lu et al., 2005; Wood & Bandura, 1989). In other words, managerial self-efficacy represents a manager’s self-perception regarding his or her managerial role (Fast et al., 2014; Lu et al., 2005). Previous research has shown that managerial self-efficacy is particularly crucial for managers to fulfill their managerial responsibilities (for a review, see Hannah et al., 2008). Self-perception theory suggests that by observing their own behaviors, managers may form an assessment of how effective and competent they are as managers (Bem, 1982; Casciaro et al., 2014). Accordingly, supervisors’ managerial self-efficacy may also be affected by their own behavior such as abusive supervisory behavior. In this research, integrating self-perception theory with power-dependence theory, we contend that abusive supervisory behavior has a positive indirect relationship with supervisors’ managerial self-efficacy via state sense of power. Specifically, supervisors’ self-efficacy increases when they perceive themselves as highly capable, significant, successful, and worthy (Judge & Bono, 2001). Research has shown that supervisors with elevated state sense of power tend to believe that their capabilities are superior to others and hold confidence in their ability to fulfill their managerial roles (Kipnis, 1976; Tost, Gino, & Larrick, 2013). In particular, research has shown that feeling powerful is related to having more confidence in one’s own knowledge (Fast, Sivanathan, Mayer, & Galinsky, 2012; Tost, Gino, & Larrick, 2012) and decisions (Anderson & Galinsky, 2006; Briñol, Petty, Valle, Rucker, & Becerra, 2007), higher perceptions of personal control (Fast, Gruenfeld, Sivanathan, & Galinsky, 2009), and lower perceptions of environmental constraints (Mourali & Yang, 2013). Such positive self-evaluations and self-perceptions enhance supervisors’ confidence in making managerial decisions, leading subordinates, and coping with difficulties in their managerial work (Judge & Bono, 2001). Accordingly, elevated state sense of power is likely to be positively related to supervisors’ perceptions that they are competent in their managerial roles. The above arguments suggest that state sense of power is positively related to managerial self-efficacy. In addition, because we expect abusive supervisory behavior to be positively related to state sense of power, abusive supervisory behavior may have a positive indirect relationship with managerial self-efficacy via state sense of power. Thus, we propose the following:

2.2. Abusive supervisory behavior and state sense of power As stated above, based on self-perception theory and power-dependence theory, supervisors’ perception of power over subordinates can be boosted by their own abusive supervisory behaviors. When exercising abusive supervision, supervisors are more likely to perceive their control over subordinate-valued resources and goals, thereby perceiving more subordinate dependence on them (Emerson, 1962; Wee et al., 2017). Specifically, abusive supervision involves destructive behaviors against the desire of the subordinates, such as ridiculing subordinates, putting subordinates down in front of others, not giving credits for subordinates’ efforts, and inappropriately blaming subordinates (Tepper, 2000, 2007; Tepper, Duffy, & Shaw, 2001). Abusive supervision also involves manipulative behaviors and falls within the domain of willful behaviors, demonstrating that abusive supervisors can manipulate subordinate-valued resources and goals in ways against subordinates’ desires. For example, subordinates value support, recognition, and constructive feedback provided by their supervisors (Eisenberger, Stinglhamber, Vandenberghe, Sucharski, & Rhoades, 2002). However, abusive supervision shows that supervisors can arbitrarily deprive subordinates of their valued treatments in the workplace. Consequently, subordinates cannot achieve the workplace goals they desire. Abusive supervision reminds supervisors that they can withhold subordinates’ valued resources (e.g., not providing helpful advice to employees) and hinder subordinates’ goal attainment (e.g., giving negative performance evaluations that may decrease employees’ likelihood of salary increase and promotions). Therefore, the above theorizing suggests that performing abusive supervision may lead supervisors to perceive that they have control over subordinates’ valued resources and goals and subordinates are dependent on them. Consequently, a stronger state sense of supervisor power may emerge following abusive supervision. Our arguments have received some empirical support from previous research. For example, researchers have demonstrated that an individual gain a stronger sense of power when he or she engages in behavior that regulates others’

Hypothesis 2. A supervisor’s abusive supervisory behavior has a positive indirect relationship with his or her managerial self-efficacy via state sense of power. 2.4. The moderating role of chronic sense of power Although abusive supervisory behavior may be positively related to 82

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

(i.e., mediating effect of state sense of power) and when (i.e., moderating effect of chronic sense of power) abusive supervision predict the actors’ self-perception as a competent manager. Thus, we propose the following:

state sense of power, this positive relationship may vary for different supervisors. Specifically, self-perception theory suggests that individuals form perceptions about themselves by observing their own overt behaviors (Bem, 1982; Casciaro et al., 2014). Accordingly, individuals’ behaviors that are salient to themselves are more likely to influence their self-perceptions. Thus, drawing upon self-perception theory and power-dependence theory, we argue that the effect of abusive supervision on supervisors’ state sense of power is moderated by supervisors’ chronic sense of power, as it determines the extent to which supervisors view abusive supervisory behavior as salient actions demonstrating their power (Anderson et al., 2012; Williams et al., 2017). Specifically, Anderson et al. (2012) suggested that one’s state sense of power may stem from both momentary social interactions (e.g., abusive supervision) and long-term dyadic relationships (e.g., supervisor-subordinate dyads). In other words, a supervisor’s moments of abusive supervision and his or her chronic sense of power from longterm interactions with subordinates may jointly shape the supervisor’s state sense of power. Supervisors’ chronic sense of power refers to supervisors’ general perception of their ability to influence their subordinates, which is formed gradually based on their numerous relational experiences with their subordinates (Anderson & Galinsky, 2006; Anderson et al., 2012). Compared to those with high chronic sense of power, supervisors with low chronic sense of power generally consider themselves as having limited ability and few opportunities to influence their subordinates and control subordinate-valued goals. For such supervisors, actions that demonstrate their power or control over subordinate-valued goals (e.g., abusive supervisory behavior) are likely to be salient to them and to be perceived as signals of power. In contrast, supervisors with high chronic sense of power frequently demonstrate their ability to influence subordinates and control subordinate-valued goals. Consequently, momentary actions that demonstrate their power will not be as salient to them (Williams et al., 2017). In other words, supervisors with high (versus low) chronic sense of power are less likely to view their abusive supervisory behavior as salient actions that demonstrate power and are thus less likely to feel more powerful after abusing subordinates. The above theorizing has received some support from previous research, which has suggested that being in a leadership position while also having low chronic sense of power sensitizes supervisors to any action that may increase their power (Fiske & Dépret, 1996; Goodstadt & Hjelle, 1973; Hays & Goldstein, 2015). Taking the above arguments and evidence together, we suggest that the positive relationship between abusive supervisory behavior and state sense of power is stronger for supervisors with low chronic sense of power.

Hypothesis 4. A supervisor’s chronic sense of power moderates the indirect effect of the supervisor’s abusive supervisory behavior on his or her managerial self-efficacy via state sense of power such that the positive indirect effect is stronger when the supervisor has a low, rather than a high, level of chronic sense of power. 2.5. Managerial self-efficacy and task-oriented leadership behavior Task-oriented leadership refers to the degree to which leaders define and organize followers’ and their own roles and are oriented toward defining performance, goal, and role expectations and constraints (Fleishman, 1973). Task-oriented leadership is a fundamental day-today behavior performed by all types of leaders (Fleishman, 1953; Yukl, 1971). The meta-analytical study of Judge et al. (2004) revealed that task-oriented leadership has significant impacts on employee, group, and organizational performance (Judge et al., 2004). They also called for a renewal of research interest in this important but forgotten leadership style. In this research, we argue that supervisors’ managerial self-efficacy is positively related to their task-oriented leadership behavior. Specifically, individuals with high (versus low) self-efficacy tend to perceive more capability to mobilize resources and take action to regulate their own behavior and control others’ behavior to accomplish goals (Bandura, 1977; Wood & Bandura, 1989). In particular, supervisors with high managerial self-efficacy believe that they are competent in mobilizing resources and regulating their own and their followers’ behavior to accomplish managerial tasks and fulfill managerial role expectations (Fast et al., 2014). Hence, they are more effective in making work plans, setting directions and rules, assigning tasks, and leading their subordinates by example (Ng, Ang, & Chan, 2008). These are all key components of task-oriented leadership. Accordingly, we argue that supervisors with higher managerial self-efficacy engage in more task-oriented leadership behavior. This claim is also consistent with previous findings that managerial self-efficacy is critical for supervisors to influence their subordinates (Kane, Zaccaro, Tremble, & Masuda, 2002; Ng et al., 2008). In sum, we propose the following: Hypothesis 5. A supervisor’s managerial self-efficacy is positively related to his or her task-oriented leadership behavior.

Hypothesis 3. A supervisor’s chronic sense of power moderates the positive relationship between his or her abusive supervisory behavior and state sense of power such that the positive relationship is stronger when the supervisor has a low, rather than a high, level of chronic sense of power.

3. Research overview To test our model, we conducted a field diary study (Study 1) and three experiments (Studies 2a, 2b and 3). In Study 1, we adopted a multi-level diary design with data collected at multiple time points to test the overall moderated mediation model in a field setting. In Studies 2a and 2b, we conducted two experiments with participants recruited from mainland China and the United States to constructively replicate our test of Hypothesis 1 and to establish the causal relationship between abusive supervisory behavior and state sense of power. In Study 3, using an experimental design, we further replicated the findings from Studies 1, 2a, and 2b, and tested the full model. Based on this multimethod design, we were able to enhance both the internal and external validity of our findings.

Thus far, based on self-perception theory and power-dependence theory, we have proposed that abusive supervisory behavior positively relates to supervisors’ managerial self-efficacy via state sense of power (Hypotheses 1 and 2) and that the relationship between abusive supervisory behavior and state sense of power is bounded by supervisors’ chronic sense of power (Hypothesis 3). Our final hypothesis integrates these ideas into a single moderated mediation model, in which chronic sense of power also moderates the indirect relationship between abusive supervisory behavior and managerial self-efficacy via state sense of power. Specifically, congruent with the theoretical expectation that chronic sense of power weakens the positive relationship between abusive supervision and state sense of power, the positive indirect relationship between abusive supervision and managerial self-efficacy via state sense of power is expected to be stronger for supervisors with low rather than high levels of chronic sense of power. By proposing and testing this moderated mediation model, we are able to explore why

4. Study 1 method 4.1. Participants The participants for Study 1 were recruited from the authors’ alumni networks. We used this approach because it enabled us to recruit 83

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

4.3. Within-individual measures

participants from various industries and organizations and thus enhances the generalizability of the findings. Indeed, such an approach has been commonly adopted in management research (e.g., Bernerth, Taylor, Walker, & Whitman, 2012; Qin, Huang, Hu, Schminke, & Ju, 2018; Sikora, Ferris, & Van Iddekinge, 2015). Initially, 89 supervisors agreed to participate in our study. Each participant was compensated with three managerial books worth USD 11 in total. In addition, to further facilitate participation, we offered to provide each participant with a report on their personality and leadership style after the study. Ultimately, 72 supervisors provided matched data. Of the final sample of participants, 54.2% were male, and their average age was 29.6 years. On average, they had received 17.2 years of education. They worked in different industries, with 23.6% working in manufacturing, 20.8% working in service industries, and 55.6% working in other industries. Professionally, 16.7% of the participants worked in technology-related departments, 12.5% worked in administration-related departments, 45.8% worked in marketing-related departments, and 25% worked in other departments. Their average managerial tenure was 3.1 years. On average, they supervised seven subordinates and worked for 8.9 h per weekday. There were no significant differences regarding the demographic characteristics between our final sample and those excluded.

4.3.1. Abusive supervisory behavior (T1) Abusive supervisory behavior was assessed using Tepper (2000) five-item abusive supervision scale. We invited supervisors to report their own abusive supervisory behavior toward subordinates for the following reasons. First, supervisors are in the best position to know their own overall behavior toward their subordinates as a group, especially within a day (Courtright, Gardner, Smith, Mccormick, & Colbert, 2016; Johnson, Venus, Lanaj, Mao, & Chang, 2012). A particular subordinate may not be able to detect his or her supervisor’s behavior toward other subordinates. As a result, abusive supervisory behavior toward a whole group would likely be missed or underestimated. Second, a recent meta-analysis revealed that there was a moderate to high correlation between self- and other-reports of counterproductive work behavior as well as similar patterns and magnitudes of relationships with a set of common correlates (Berry, Carpenter, & Barratt, 2012). This indirect evidence indicates that self- and otherreports of abusive supervision may also have a moderate to high correlation. Third, even though self-report measures may be biased due to supervisors’ social desirability or stable individual traits, this should not affect the results because we followed the recommendations of Hofmann, Griffin, and Gavin (2000) and previous diary research (e.g., Barnes et al., 2015; Koopman et al., 2016) and mean-centered all daily variables at the person level, which effectively removes between-person confounds, such as response tendencies and stable traits. Accordingly, this approach has been widely applied to measure abusive supervisory behavior in diary research (e.g., Foulk et al., 2018; Liao, Yam, Johnson, Liu, & Song, 2018). Sample items included “This morning, I ridiculed my subordinates” and “This morning, I told my subordinates their thoughts or feelings were stupid” (1 = “Never,” 5 = “Always”). The average Cronbach’s α across days was 0.93.

4.2. Procedures The data collection in this study consisted of two phases across three weeks. In each phase, we sent participants links to online surveys via WeChat, a widely used online communication platform in China. All surveys were conducted in Chinese, and we followed a translation and back-translation procedure (Brislin, 1980) to translate scales that were not originally developed in Chinese. In the first phase, we sent an initial survey to participants in which they reported their demographic information and chronic sense of power. One week later, in the second phase, we sent two daily surveys to participants (one in the morning and the other in the afternoon) for 10 consecutive workdays (spanning two weeks) following prior daily studies (e.g., Barnes, Lucianetti, Bhave, & Christian, 2015; Johnson, Lanaj, & Barnes, 2014; Scott & Barnes, 2011). To provide a full day-level observation, participants were required to complete both daily surveys. The morning survey (Time 1, T1) was sent at 11:00 a.m. At this point in time, participants would have had enough opportunities to display leadership behavior during the first hours of their workday. In the morning survey, participants were asked to report their abusive supervisory behavior in the morning and state sense of power (T1). In the afternoon survey (Time 2, T2), sent at 4:30p.m., participants reported their state sense of power (T2) and managerial self-efficacy. The average time lapse between responses to the morning and afternoon surveys was 6.1 h. It is worth noting that in line with prior related studies (e.g., Koopman, Lanaj, & Scott, 2016), we used the independent variable at T1 to predict the mediator and the dependent variable at T2 while controlling for the mediator at T1. This design with a time lag between the independent variable and both the mediator and the dependent variable not only alleviates concerns of common method variance but also helps infer causality and rule out alternative explanations (Brewer, 2000; Johnson, Rosen, & Djurdjevic, 2011). Without controlling for the mediator at T1, we would not have been able to rule out the impact of the mediator at T1 on the mediator at T2. Thus, we would not have been able to fully establish the causal relationship between the independent variable at T1 and the mediator at T2. Overall, we obtained 560 daily-level data points out of 720 potential data points, representing a response rate of 77.8%. There were no significant differences in individual characteristics (e.g., gender, age, education, and tenure) between participants who completed the survey and those who dropped out.

4.3.2. State sense of power (T2) State sense of power was assessed using four items adapted from the eight-item power scale developed by Anderson et al. (2012). We chose to include only the four items that fit the research context and had high factor loadings to reduce the burden on participants. This practice is common in daily research designs (Fisher & To, 2012; Johnson et al., 2014). Participants rated the extent to which each item described their feelings at that moment on a five-point Likert scale (1 = “Strongly disagree”, 5 = “Strongly agree”). This approach is the conventional way of measuring participants’ momentary feelings in diary research (e.g., Koopman et al., 2016). The four items were “In my interactions with my subordinates, I can get them to listen to what I say”; “In my interactions with my subordinates, I can get them to do what I want”; “In my interactions with my subordinates, I think I have a great deal of power”; and “In my interactions with my subordinates, if I want to, I get to make the decisions.” The average Cronbach’s α across days was 0.95. To address concerns about the validity of the shorter scale, we conducted a pilot study in which the full eight-item sense of power scale was administered to 81 full-time supervisors. Participants in the sample had an average age of 32.7 years, an average organizational tenure of 7.1 years, and an average managerial tenure of 4.4 years. They were from different industries: 2.5% from manufacturing, 30.1% from service industries, and 67.4% from other industries. They work in different departments, including technology (30.0%), administration (37.0%), finance (8.6%), marketing (8.6%), and other departments (15.8%). Participants responded to the items using the same five-point scale as in the primary study. The results indicated that scores on the short fouritem scale (α = 0.88) were highly correlated (r = 0.84, p < .001) with scores on the full eight-item scale (α = 0.89). Thus, the short scale was a suitable substitute. 4.3.3. Managerial self-efficacy (T2) Consistent with the approach adopted by Fast et al. (2014), we 84

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

Table 1 Percentage of within-individual variance among the daily variables (Study 1). Variables

Intercept (b00)

Within-Individual Variance (e2)

Between-Individual Variance (r2)

Percentage of Within-Individual Variance (%)

State sense of power (T1) Abusive supervisory behavior (T1) State sense of power (T2) Managerial self-efficacy (T2)

3.78 1.16 3.73 3.74

0.13 0.06 0.11 0.12

0.18 0.15 0.23 0.21

42% 29% 32% 36%

Note. b00 = average level of the variable across individuals, e2 = within-individual variance, r2 = between-individual variance. Percentage of within-individual variance was calculated as e2/(e2 + r2).

4.6. Analytic strategy

assessed managerial self-efficacy with four items adapted from Chen, Gully, and Eden (2001) eight-item perceived managerial self-efficacy scale. Sample items were “In managerial work, compared to others, I am confident that I can perform effectively on many different tasks” and “In managerial work, I am certain that I will be able to successfully overcome many challenges.” The average Cronbach’s α across days was 0.95. Again, we used the shorter scale to reduce the burden on participants, as is common in studies using a diary/experience-sampling design (Fisher & To, 2012; Rothbard & Wilk, 2001). To establish the validity of the shorter scale, we conducted a short pilot study in which the full eight-item managerial self-efficacy scale was administered to 81 full-time supervisor participants. Participants responded to the items using the same five-point scale as in the primary study. The results indicated that scores on the short four-item scale (α = 0.81) were highly correlated (r = 0.96, p < .001) with scores on the full eightitem scale (α = 0.90). Thus, the short scale was a suitable substitute.

Considering the nested structure of our data (i.e., daily data nested within each person), we used hierarchical linear modeling (HLM; Bryk, Raudenbush, & Congdon, 1996) to test the hypotheses. Following the recommendation by Hofmann et al. (2000), we group-mean centered all the Level 1 variables, including state sense of power (T1), abusive supervisory behavior (T1), state sense of power (T2), and managerial selfefficacy (T2), to examine daily within-individual fluctuations while controlling for between-individual confounds. The Level 2 variable (i.e., chronic sense of power) was grand-mean centered. To test the within-individual hypotheses (Hypotheses 1 and 2), we constructed fixed-effect models in which we controlled for betweengroup variance but did not use any specific group-level variable as a predictor. We used RMediation (Tofighi & Mackinnon, 2011), a commonly used method (e.g., Qin, Ren, Zhang, & Johnson, 2018) that estimates Type I error rates more accurately than traditional mediation tests (e.g., the Sobel test; Mackinnon, Fritz, Williams, & Lockwood, 2007), to test the indirect effect (Hypothesis 2). To test the betweenindividual hypotheses (Hypotheses 3 and 4), we constructed mixedeffect models in which we added a group-level variable and its interaction with an individual-level variable as predictors. We used Edwards and Lambert (2007) moderated path analysis approach to test the moderated mediation hypothesis (Hypothesis 4).

4.4. Between-individual measure 4.4.1. Chronic sense of power Chronic sense of power was assessed in the first phase using the eight-item sense of power scale developed by Anderson et al. (2012). In line with previous research (e.g., Morrison, See, & Pan, 2015), participants were instructed to think about their typical interactions with the subordinates on their teams and then indicate their agreement with the statements (1 = “Strongly disagree”, 5 = “Strongly agree”). Sample items included “In my team, I can get my subordinates to do what I want” and “In my team, my wishes carry much weight” (α = 0.90).

5. Study 1 results and discussion Table 1 reports the ratios of within-individual variance. The results revealed that 29% of the variance in abusive supervisory behavior resided at the within-person level. For state sense of power at T1 and T2 and managerial self-efficacy at T2, the percentages of within-person variation were 42%, 32%, and 36%, respectively. These results suggested that there was sufficient within-individual variance in our data. Table 2 displays the descriptive statistics and the within- and between-individual correlations among the variables. We conducted a multilevel confirmatory factor analysis (CFA) before testing the hypotheses. Specifically, we included three within-individual variables (i.e., abusive supervisor behavior [T1], state sense of power [T2], and managerial self-efficacy [T2]) and one between-individual variable (i.e., chronic sense of power). The results indicated that the four-factor

4.5. Control variables 4.5.1. State sense of power (T1) We controlled for state sense of power (T1) to reveal the increased effect stemming from abusive supervisory behavior, which helped establish the causal relationship between abusive supervisory behavior and state sense of power. State sense of power (T1) was assessed using the same measure as state sense of power (T2). The average Cronbach’s α across days was 0.94.

Table 2 Descriptive statistics and correlations (Study 1). Variables Level 1 variables 1. State sense of power (T1) 2. Abusive supervisory behavior (T1) 3. State sense of power (T2) 4. Managerial self-efficacy (T2) Level 2 variable 5. Chronic sense of power

Mean

SD

1

2

3

3.80 1.10 3.76 3.76

0.55 0.37 0.57 0.56

−0.06 0.14*** 0.03

0.21*** 0.12**

0.36***

3.90

0.47

0.32***

−0.24***

0.30***

+

4

5

0.19***

0.25***

Note. ***p < .001, **p < .01, *p < .05, p < .10. N = 560 at the individual-day level, N = 72 at the individual level. Correlations for the Level 1 variables represent group-mean centered relationships among the daily variables at the within-individual level of analysis. The Level 1 variables were neither centered nor aggregated to provide estimates of the between-individual relationships with the Level 2 variable. 85

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

Table 3 HLM results for predictors of state sense of power (Study 1). Variables

State Sense of Power (T2) Main Effect Model

State sense of power (T1) Abusive supervisory behavior (T1) Chronic sense of power Abusive supervisory behavior (T1) × Chronic sense of power Constant Between-individual variance Within-individual variance Deviance

Moderating Effect Model

b

SE

t

b

SE

t

0.12 0.32

0.04 0.12

2.98** 2.62**

3.73

0.06

62.98***

0.12 0.16 0.43 −0.66 3.74

0.04 0.08 0.11 0.13 0.05

2.99** 2.04* 3.95*** −5.17*** 69.44***

0.24 0.10

0.19 0.10

507.16

478.54

Note. N = 560 at the individual-day level, N = 72 at the individual level. Model deviance = −2 × log-likelihood of the full maximum-likelihood estimate. *** p < .001. ** p < .01. * p < .05.

model had desirable fit (χ2(183) = 470.64, p < .001; standardized root mean square residual [SRMR] = 0.06, root mean square error of approximation [RMSEA] = 0.05, comparative fit index [CFI] = 0.90; Hu & Bentler, 1999) and fit better than alternative models (these results are available upon request), suggesting that these variables were distinguishable.

Table 4 HLM results for predictors of managerial self-efficacy (Study 1). Variables

Managerial Self-Efficacy (T2)

Abusive supervisory behavior (T1) State sense of power (T2) Constant

5.1. Tests of the hypotheses 5.1.1. Within-individual hypotheses In Hypothesis 1, we propose that engaging in abusive supervisory behavior is positively related to supervisors’ state sense of power. The results (shown in Table 3) revealed that daily abusive supervisory behavior positively predicted daily state sense of power (b = 0.32, p < .01). Thus, Hypothesis 1 was supported. Hypothesis 2 predicts that daily state sense of power mediates the relationship between daily abusive supervisory behavior and daily managerial self-efficacy. As shown in the indirect effects model of Table 4, daily state sense of power was significantly related to daily managerial self-efficacy (b = 0.36, p < .001). RMediation was further used to test the indirect effects, and the results showed that abusive supervisory behavior had significant indirect effects on daily managerial self-efficacy (estimate = 0.12, 95% confidence interval [CI] = 0.03, 0.22) via daily state sense of power.2 Thus, Hypothesis 2 was supported.

Between-individual variance Within-individual variance Deviance

Total Effects Model

Indirect Effects Model

b

SE

t

b

SE

t

0.20

0.08

2.40*

0.08

0.08

0.94

3.74

0.06

65.60***

0.36 3.74

0.05 0.06

8.00*** 65.48***

0.22

0.22

0.12 588.17

0.10 528.10

Note. N = 560 at the individual-day level (Level 1), N = 72 at the individual level (Level 2). Model deviance = -2 × log-likelihood of the full maximumlikelihood estimate. **p < .01. *** p < .001. * p < .05. 4.25

State Sense of Power 3.75

5.1.2. Between-individual hypotheses Hypothesis 3 predicts that chronic sense of power attenuates the relationship between abusive supervisory behavior and state sense of power. The results (shown in Table 3) indicate a significant two-way interaction between abusive supervisory behavior and chronic sense of power in predicting state sense of power (b = −0.66, p < .001). We illustrate this between-individual interaction in Fig. 2. Simple slope

Low chronic sense of power High chronic sense of power 3.25 Low High Abusive Supervisory Behavior

Fig. 2. The moderating role of chronic sense of power on the relationship between abusive supervisory behavior and state sense of power (Study 1).

2

To better leverage our existing data, we also tested whether abusive supervisory behavior predicts next-day state sense of power and managerial selfefficacy. The results revealed that abusive supervisory behavior predicted state sense of power on the same day (as the results of our main text showed) but did not predict next-day state sense of power (b = 0.12, ns) or managerial selfefficacy (b = 0.13, ns), implying that the impact of abusive supervisory behavior on state sense of power and managerial self-efficacy was relatively immediate. Furthermore, we conducted additional analyses to test the potential reciprocal relationships. The results revealed that neither state sense of power (b = −0.02, ns) nor managerial self-efficacy (b = −0.01, ns) predicted nextday abusive supervisory behavior. Detailed results are available from the authors upon request.

tests revealed that the relationship between abusive supervision and state sense of power was significant and positive when chronic sense of power was low (b = 0.47, p < .001) but not when chronic sense of power was high (b = −0.15, ns). Thus, Hypothesis 3 was supported. In Hypothesis 4, we suggest that chronic sense of power moderates the indirect effect of engaging in abusive supervisory behavior on daily managerial self-efficacy via daily state sense of power. We calculated the conditional indirect effects at higher (+1 SD) and lower (−1 SD) 86

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

5) of the second phase of this study and included it as the independent variable to predict supervisors’ state sense of power and managerial self-efficacy on Day 10. We also conducted OLS with the same controls as above to test these proposed effects, and the results showed that the aggregated abusive supervisory behavior from Day 1 to Day 5 was negatively related to supervisors’ state sense of power on Day 10 (b = −0.68, p < .001) and also had a significant negative indirect relationship with supervisors’ managerial self-efficacy on Day 10 through their state sense of power on Day 10 (estimate = −0.39, 95% CI = −0.66, −0.18). Taken together, the results of these supplementary analyses suggested that in the short term (i.e., within one day), abusive supervisory behavior benefitted supervisors’ state sense of power and managerial self-efficacy, whereas in the long term (e.g., in a week), such behavior weakened supervisors’ state sense of power and managerial self-efficacy. To sum up, with a multi-wave and multi-level diary design, Study 1 supported our Hypotheses 1–4 in a field setting, but it may not generate strong causal inferences. Thus to address this limitation, we conducted two experiments (Studies 2a and 2b) to establish the causality between abusive supervisory behavior and state sense of power.

levels of chronic sense of power (Edwards & Lambert, 2007). The results indicated that the indirect effect of abusive supervisory behavior on daily managerial self-efficacy via daily state sense of power was significant when chronic sense of power was low (b = 0.17, p < .01) but not when chronic sense of power was high (b = −0.06, ns). The difference between these indirect effects was significant (Δb = −0.22, p < .05). Also, the results revealed that chronic sense of power did not significantly moderate the relationship between state sense of power and managerial self-efficacy (b = 0.04, ns). Thus, Hypothesis 4 was supported. 5.2. Supplementary analyses: long-term effects of abusive supervisory behavior As stated above, we demonstrated that abusive supervisory behavior was positively related to supervisors’ state sense of power in the short term (i.e., within one day). However, according to prior research, it is possible that from a long-term perspective, abusive supervisory behavior is negatively related to supervisors’ sense of power. Specifically, in the long term, abusive supervisory behavior is likely to trigger subordinates’ deviant behavior and supervisor-directed aggression (Ferris, Yan, Lim, Chen, & Fatimah, 2016; Hershcovis & Barling, 2010; Robinson & Bennett, 1995). Such anti-abuse behaviors may make supervisors feel that their subordinates are not dependent on them for resource acquisition and goal achievement, thereby weakening supervisors’ sense of power. To investigate this possibility, we conducted several supplementary analyses. First, we tested the effect of average abusive supervisory behavior over time on supervisors’ state sense of power. Specifically, in the initial survey, we measured supervisors’ average abusive supervisory behavior over time, or how frequently they engaged in abusive supervisory behavior during the previous three months, with five items developed by Mitchell and Ambrose (2007) (α = 0.86). We then conducted HLM to test the proposed long-term effects of average abusive supervisory behavior on supervisors’ subsequent state sense of power measured in the afternoon of the second phase of this study. The results showed that average abusive supervisory behavior was negatively associated with supervisors’ subsequent state sense of power (b = −0.30, p < .01). Next, we used RMediation to test the mediating role of state sense of power. The results showed that supervisors’ state sense of power measured in the afternoon significantly mediated the effect of average abusive supervisory behavior measured in the initial survey on their managerial self-efficacy measured in the afternoon (estimate = −0.11, 95% CI = −0.21, −0.02). These findings suggested that over a longer time (the time lags between the independent variable and dependent variables ranged from seven to 19 days), abusive supervisory behavior was negatively related to supervisors’ sense of power and managerial self-efficacy. Second, we included average abusive supervisory behavior in the initial survey as the independent variable to predict the outcomes measured on the last day of the second-phase survey—that is, supervisors’ state sense of power and managerial self-efficacy measured on Day 10 in the second phase of this study. We tested this effect using ordinary least squares (OLS) regression. We included supervisors’ gender, age, education, managerial tenure, and chronic sense of power as control variables. The results indicated that average abusive supervisory behavior was negatively associated with supervisors’ state sense of power on Day 10 (b = −0.42, p < .05). Furthermore, the mediation analysis showed that average abusive supervisory behavior had a significant indirect relationship with supervisors’ managerial self-efficacy on Day 10 via state sense of power on Day 10 (estimate = −0.22, 95% CI = −0.48, −0.01). Third, to verify the long-term effects of average abusive supervisory behavior, we also used another way to measure average abusive supervisory behavior over time. We averaged the data on abusive supervisory behavior collected from the first five days (i.e., Day 1 to Day

6. Study 2a method 6.1. Participants For Study 2a, we recruited supervisors from Sojump.com, a professional online survey platform that is widely used in management research (for previous uses of this survey panel in management research, see Jin, Ford, & Chen, 2013; Liu, Keller, & Hong, 2015). Each respondent was rewarded with RMB 11 (approximately USD 2) for participation. The sample consisted of 102 full-time supervisors (58.0% male) with an average age of 35.2 years, an average organizational tenure of 7.3 years, and an average managerial tenure of 4.6 years. The supervisors worked in a variety of industries, including manufacturing (11.8%), service industries (37.3%), and other industries (50.9%), and departments, including technology (28.4%), administration (24.5%), finance (5.9%), marketing (40.2%), and other departments (1.0%). 6.2. Procedures and experimental design Participants were instructed to complete a task that involved recalling a recent interaction with their subordinates and describing the interaction. Supervisors were randomly assigned to one of two conditions: the abusive supervisory behavior condition or the control condition. 6.2.1. Abusive supervisory behavior manipulation To manipulate abusive supervisory behavior (it is not ethical or appropriate to have subjects exercise abusive supervision toward real subordinates), we used the critical incident technique in line with related research (e.g., Aquino, Tripp, & Bies, 2001; Bobocel, 2013). Specifically, participants were first asked to recall a time when one of their subordinates exhibited poor performance at work and they treated the subordinate in either an abusive way (experimental condition, n = 52) or a non-abusive way (control condition, n = 50). In the instructions, we gave participants some examples of abusive or nonabusive supervisory behavior. The examples for abusive supervisory behavior were adopted from the abusive supervision scale (Tepper, 2000). Specifically, in the experimental condition, participants read the following instruction: Please recall a particular incident in which one of your subordinates exhibited poor performance at work and you treated him or her in one or several of the following ways: ridiculed him or her, told him or her that his or her thoughts or feelings were stupid, and/or made negative 87

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

comments about him or her to others.

Study 2a established that engaging in abusive supervisory behavior led to a higher level of state sense of power. In this study, we manipulated abusive supervision by asking participants to recall past interactions with subordinates. Using this critical incident technique is consistent with related research (e.g., Aquino et al., 2001; Bobocel, 2013). However, previous studies have suggested using different manipulations across experiments to verify the robustness of findings (e.g., Dubois, Rucker, & Galinsky, 2015; Rothman, 2011; Waytz, Chou, Magee, & Galinsky, 2015). Therefore, in Study 2b, we manipulated abusive supervision in a different way to further establish the causal relationship between abusive supervision and state sense of power. In addition, the participants for Study 2a were from China, a country with high power distance (Hofstede, 2001) where people believe that organizational, institutional, and societal power should be distributed unequally (Hofstede, 2001). Therefore, in Study 2b, we recruited participants from Western cultures to increase generalizability.

In the control condition, participants read the following instruction: Please recall a particular incident in which one of your subordinates exhibited poor performance at work and you treated him or her in one or several of the following ways: pointed out his or her low performance at work, discussed the reasons behind his or her low performance, and/or expressed your expectation for him or her to improve. Immediately after the recall task, participants were invited to write a paragraph to describe what happened during the incident in concrete detail, re-living the experience as much as possible as they described the event. After completing the task, participants were invited to fill out a survey in which they rated their state sense of power and reported demographics. 6.3. Measures

8. Study 2b method

6.3.1. State sense of power We assessed state sense of power with the same scale used in Study 1 (α = 0.74). Specifically, participants rated how they evaluated their relationship with their subordinates at that moment. The items were rated using a five-point Likert-type scale (1 = “Strongly disagree”, 5 = “Strongly agree”).

8.1. Participants For Study 2b, we recruited supervisors from the United States and Canada through Mechanical Turk (MTurk; Buhrmester, Kwang, & Gosling, 2011). Research has shown that data obtained via MTurk has psychometric properties similar to data obtained using other convenient sampling methods (Buhrmester et al., 2011; Lin & Johnson, 2015; Paolacci, Chandler, & Ipeirotis, 2010). Researchers’ investigations into MTurk as a data-collection source for psychological and social science studies have shown that “the data obtained are at least as reliable as those obtained via traditional methods” and that “MTurk can be used to obtain high-quality data inexpensively and rapidly” (Buhrmester et al., 2011, p. 3). Participants were paid USD 1 for participation. Moreover, following the recommendations of Meade and Craig (2012), we added an attention-check item (i.e., “Please respond with ‘strongly agree’”) in the survey to exclude participants who responded carelessly. It is reasonable to infer that those who failed to choose “strongly agree” in this question might not have read the other questions carefully. To guarantee the quality of the data, we excluded 12 respondents who failed the attention check. The final sample consisted of 194 participants who came from a wide range of professions and were all supervisors in their firms. Among the participants, the average age was 37 years, 56.7% were male, 77.8% were Caucasian, and the average managerial tenure was 6.4 years. The participants were from different industries, including 5.7% from manufacturing, 27.8% from service industries, and 66.5% from other industries. They worked in different departments, including technology (30.4%), administration (28.4%), finance (10.8%), marketing (10.8%), and other departments (19.6%).

7. Study 2a results and discussion 7.1. Manipulation check Two independent research assistants who were blind to this study’s hypotheses and conditions coded participants’ supervisory behavior from the written paragraphs using Tepper (2000) five-item abusive supervision scale, which captures verbal aggression exhibited by supervisors (Mitchell & Ambrose, 2007). The two coders’ scores achieved high inter-rater reliability (r = 0.82, p < .001); we thus averaged them to form an overall assessment of participants’ abusive supervisory behavior. The t-test results revealed that abusive supervisory behavior was rated significantly higher in the abusive condition (M = 3.90, SD = 0.63) than in the control condition (M = 1.95, SD = 0.39); t (100) = 18.99, p < .001, indicating that our manipulation was successful in the expected direction. 7.2. Tests of the hypotheses To test Hypothesis 1, which predicts that engaging in abusive supervisory behavior is positively related to supervisors’ state sense of power, we conducted an analysis of variance (ANOVA). In support of Hypothesis 1, the results revealed that supervisors assigned to the experimental condition reported significantly higher levels of state sense of power (M = 4.13, SD = 0.41) compared with supervisors in the control condition (M = 3.88, SD = 0.67; F(1, 100) = 5.13, p < .05).3

8.2. Experimental procedures In this study, participants read and engaged in a scenario instructing them to adopt the role of a team leader who has five subordinates. Participants were also instructed to finish a writing task. Specifically, participants first read a paragraph describing their team and the leader role they played on the team. Participants were then required to draft

3

As stated previously, we propose that abusive supervisory behavior enhances supervisors’ state sense of power by enhancing their perceptions of subordinate dependence. Therefore, we conducted a supplementary analysis to examine the mediating role of supervisors’ perceptions of subordinate dependence. Participants were asked to report their perceptions of subordinate dependence before they rated their state sense of power. Based on the definition of dependence (Emerson, 1962) and in line with Wee et al. (2017), we asked all participants to rate the extent to which they perceived that their subordinates depended on them for valued resources and goal attainment using four items (α= 0.73): “In my team, I determine the rewards and punishment given to the members”; “In my team, the members’ personal development at work depends on me”; “In my team, the members cannot finish their work without my direction”; and “In my team, the members depend on me to achieve their work goals.” The regression results showed that abusive supervisory behavior was positively related to supervisor perceived subordinate dependence (b = 0.20,

(footnote continued) p < .001), which was positively related to supervisor state sense of power (b = 0.55, p < .001). We used RMediation (Tofighi & Mackinnon, 2011) to further test the mediating role of supervisors’ perceptions of subordinate dependence in the relationship between abusive supervisory behavior and state sense of power, and the results revealed a significant indirect effect (estimate = 0.11, 95% CI = 0.06, 0.18). These findings provided evidence for our argument that abusive supervisory behavior enhances supervisors’ state sense of power by increasing perceptions of subordinate dependence. 88

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

10. Study 3 method

an email according to provided instructions (as the manipulation of abusive supervision) to one of the team members—Jess—who was making little progress in his or her work. Following the writing task, participants completed a questionnaire assessing their state sense of power and reported their demographics.

10.1. Participants In Study 3, we recruited 282 supervisors from the United States through MTurk (Buhrmester et al., 2011). Participants were paid USD 1.50 for participation. Among participants, the average age was 35.3 years, 46.8% were male, 72.7% were Caucasian, and the average managerial tenure was 4.6 years. The participants were from different industries, including manufacturing (9.2%), service industries (12.3%), and other industries (78.5%). They worked in different departments, including technology-related (27.0%), administration-related (33.0%), finance-related (7.8%), marketing-related (5.3%), and other departments (26.9%).

8.2.1. Abusive supervision manipulation We manipulated abusive supervision by instructing participants to draft an email to Jess with abusive or non-abusive content. Specifically, they were first instructed to carefully read and consider four key sentences (and try to act in line with them) and then draft an email using at least three of the key sentences. The key sentences for the abusive supervision condition were created based on the definition of abusive supervision (Tepper, 2000): “All of your ideas are stupid”; “I doubt your ability and value to the team”; “If you don’t catch up and work harder, you will get punished”; and “Don’t bring my team down.” The key sentences for the control (nonabusive) condition were “I would like to discuss the ideas you proposed,” “Let’s find a time to talk about your progress,” “You could’ve worked harder,” and “Let’s finish this project on time.” After completing the email-drafting task, participants were asked to fill out a survey in which they rated their state sense of power and reported their demographics.

10.2. Procedures and experimental design Data were collected in two phases separated by a week. In the first phase, participants were required to finish a survey in which they reported their chronic sense of power and demographics. In the second phase, participants were randomly assigned to one of two conditions: the abusive supervisory behavior condition (n = 138) or the control condition (n = 144). The experiment involved the same task as in Study 2a with one main difference. Specifically, the abusive supervision manipulation in Study 2a included poor follower performance as the trigger of abusive supervision in the recall exercise. This trigger was selected because poor performance is one of the most common reasons leaders abuse their employees (Liang et al., 2016). In addition, specifying the trigger helped ensure that recalled incidents were comparable between the two conditions. However, including the trigger of abusive supervision in the manipulation may also have created a potential confound. Therefore, as suggested by an anonymous reviewer, in this experiment, we did not specify the trigger of abusive supervision in the manipulation. After the experimental manipulation, participants were asked to fill out a survey in which they rated their state sense of power and managerial self-efficacy. Then, we measured participants’ subsequent task-oriented leadership behavior using a scenario task.

8.3. Measures 8.3.1. State sense of power For our measure of state sense of power, participants rated how they evaluated their relationship with their subordinates at that moment with the same scale used in Study 1 (α = 0.82). 9. Study 2b results and discussion 9.1. Manipulation check We invited two independent research assistants who were blind to this study’s hypotheses and conditions to code participants’ abusive supervisory behavior based on the emails that participants sent to Jess using Mitchell and Ambrose (2007) five-item abusive supervision scale (α = 0.96, 0.97, respectively). The two coders’ scores achieved high inter-rater reliability (r = 0.96, p < .001), so we averaged them to form an overall score. The results from a t-test revealed that abusive supervisory behavior was rated higher in the abusive supervision condition (M = 4.50, SD = 0.29) than in the control condition (M = 1.50, SD = 0.36), t(192) = 64.43, p < .001, Cohen’s d = 9.25. These results indicated that our manipulation was successful.

10.3. Measures 10.3.1. Chronic sense of power Chronic sense of power was measured using Anderson et al. (2012) scale as in Study 1 (α = 0.80). 10.3.2. State sense of power We assessed state sense of power with Anderson et al. (2012) scale as in Study 1 (α = 0.78).

9.2. Tests of the hypotheses 10.3.3. Managerial self-efficacy We assessed managerial self-efficacy using Chen et al. (2001) scale as in Study 1 (α = 0.85).

The ANOVA results revealed that supervisors who drafted abusive emails (M = 4.23, SD = 0.59) experienced significantly higher levels of state sense of power compared with those who drafted non-abusive emails (M = 3.94, SD = 0.67; F (1, 202) = 13.72, p < .001), thus supporting Hypothesis 1.4

10.3.4. Task-oriented leadership behavior To measure participants’ task-oriented leadership behavior, we did not use a direct measure as doing so may have led to concerns about common method variance. Therefore, in line with previous research (e.g., Antes & Mumford, 2012; Parker, Williams, & Turner, 2006; Zaccaro et al., 2015), we assessed participants’ task-oriented leadership behavior using problem-solving scenarios (Zaccaro, Mumford, Connelly, Marks, & Gilbert, 2000). In this approach, participants were

4 As in Study 2a, we conducted an additional analysis to examine whether the relationship between abusive supervisory behavior and state sense of power was mediated by supervisors’ perceptions of subordinate dependence. Participants rated their perceptions of subordinate dependence using the same scale as in Study 2a (α = 0.73). The regression results showed that abusive supervisory behavior was positively related to supervisors’ perceptions of subordinate dependence (b = 0.16, p < .001), which was positively related to supervisors’ state sense of power (b = 0.49, p < .001). RMediation (Tofighi & Mackinnon, 2011) was used to further test the mediating role of supervisors’ perceptions of subordinate dependence. The results revealed a significant indirect effect (estimate = 0.08, 95% CI = 0.04, 0.12). Therefore, Study 2b

(footnote continued) provided additional support for our argument that abusive supervisory behavior enhances supervisors’ state sense of power by boosting their perceptions of subordinate dependence. 89

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

Hypothesis 3 predicts that the positive relationship between abusive supervisory behavior and state sense of power is stronger when chronic sense of power is low (versus high). As shown in Table 5, the interaction term between abusive supervisory behavior and chronic sense of power in predicting state sense of power was significant (b = −0.22, p < .05). Simple slope tests indicated that the relationship between abusive supervision and state sense of power was significantly positive when chronic sense of power was low (b = 0.21, p < .01) but not when chronic sense of power was high (b = 0.01, ns). These results are depicted in Fig. 3. Thus, Hypothesis 3 was supported. Hypothesis 4 proposes that chronic sense of power moderates the indirect effect of abusive supervisory behavior on managerial self-efficacy via state sense of power. We adopted Hayes (2013) procedure to calculate the conditional indirect effects at higher (+1 SD) and lower (−1 SD) levels of chronic sense of power. The results indicated that the indirect effect was significant when chronic sense of power was low (b = 0.05, p < .05) but not when chronic sense of power was high (b = 0.001, ns). The difference between these indirect effects was significant (Δb = 0.05, p < .05), therefore supporting Hypothesis 4. Furthermore, supporting Hypothesis 5, the results revealed that supervisors’ managerial self-efficacy was positively related to their taskoriented leadership behavior (b = 0.33, p < .001).

instructed to read and engage in a scenario in which they were required to adopt the role of a team leader. Specifically, participants first read a paragraph describing a team leader who was confronted with challenging work issues (see Appendix A for the detailed description of the scenario). Then, they were asked to assume that they were the leader and describe how they would deal with the situation. We invited two independent research assistants who were blind to this study’s hypotheses to rate participants on the degree to which they exhibited task-oriented leadership behavior when solving the problems in their written paragraphs using the 10-item Leadership Behavior Description Questionnaire (LBDQ XII) from Stogdill (1963) (α = 0.85, 0.84, respectively). This scale has frequently been used in prior studies that have supported its reliability and validity (e.g., Judge et al., 2004; Neubert, Kacmar, Carlson, Chonko, & Roberts, 2008). Sample items included “This supervisor decided what shall be done and how it shall be done” and “This supervisor let group members know what is expected of them.” The two coders’ scores achieved high inter-rater reliability (r = 0.88, p < .001); we thus averaged them to form an overall assessment of participants’ task-oriented leadership behavior. 11. Study 3 results and discussion 11.1. Manipulation check

12. General discussion As in Study 2a, we invited another two independent research assistants who were blind to this study’s hypotheses and conditions to code participants’ abusive supervisory behavior from the written paragraphs using Mitchell and Ambrose (2007) five-item scale (α = 0.92, 0.95, respectively). We aggregated the two research assistants’ responses to form an overall assessment of participants’ abusive supervisory behavior since the two raters achieved high inter-rater reliability (r = 0.93, p < .001). The t-test results revealed that abusive supervisory behavior was rated significantly higher in the abusive condition (M = 3.25, SD = 0.06) than in the non-abusive condition (M = 1.52, SD = 0.04); t (280) = 23.7364, p < .001, Cohen’s d = 2.84, suggesting that our manipulation was successful.

The existing literature has strongly established the detrimental effects of abusive supervision on subordinates (Martinko et al., 2013; Tepper, 2007). However, we have limited knowledge about the effects of abusive supervision on actors (i.e., supervisors). Recognizing that understanding the impacts of abusive supervisory behavior on actors themselves is crucial for developing a more comprehensive understanding of abusive supervision and for finding potential ways to reduce abusive supervision, we investigate the impact of abusive supervisory behavior on supervisors themselves. Drawing upon self-perception theory and power-dependence theory, we propose that abusive supervisory behavior is positively related to supervisors’ state sense of power which in turn, is positively related to managerial self-efficacy. Furthermore, we argue that the positive indirect relationship between abusive supervisory behavior and managerial self-efficacy via state sense of power is moderated by supervisors’ chronic sense of power such that the positive indirect relationship is stronger for supervisors with low (rather than high) levels of chronic sense of power. Finally, supervisors’ managerial self-efficacy is positively related to subsequent task-oriented leadership behavior. To test our predictions, we conducted a field diary study and three experiments. The results of these studies provided substantive support for our hypotheses.

11.2. Tests of the hypotheses We conducted an ANOVA to test Hypothesis 1. The results revealed that supervisors assigned to the abusive condition reported significantly higher levels of state sense of power (M = 4.06, SD = 0.46) compared with those in the control condition (M = 3.94, SD = 0.49; F(1, 280) = 4.32, p < .05). Thus, Hypothesis 1 was supported in Study 3.5 Hypothesis 2 proposes an indirect effect of abusive supervisory behavior on managerial self-efficacy via state sense of power. As shown in Table 5, the results revealed that abusive supervisory behavior was significantly related to state sense of power (b = 0.11, p < .05) and that state sense of power was significantly related to managerial selfefficacy (b = 0.42, p < .001). The RMediation results further showed that the indirect effect of abusive supervisory behavior on managerial self-efficacy via state sense of power was significant (estimate = 0.05, 95% CI = 0.001, 0.097). Thus, Hypothesis 2 was supported.

12.1. Implications for theory The present research makes several key theoretical contributions to literatures on abusive supervision and power. First, by linking abusive supervision to actors’ state sense of power, managerial self-efficacy, and task-oriented leadership behavior, our findings enrich the understanding of the consequences of abusive supervision on actors (i.e., supervisors) rather than recipients (i.e., subordinates). Specifically, previous studies have largely taken an interpersonal approach to examine how abusive supervision affects subordinates (for reviews, see Martinko et al., 2013; Tepper, 2007), but few studies have explored how it affects actors. Our research, which uses an intrapersonal approach, represents a novel attempt to shift the predominant focus on recipients to actors by proposing and showing how abusive supervisory behavior affects actors. Interestingly, although previous research has provided substantial evidence for the detrimental effects of abusive supervision on subordinates, our research extends the literature by showing that from an actor-centric perspective, abusive supervision can be beneficial in the

5 As in Studies 2a and 2b, we conducted additional analyses to examine whether supervisors’ perceptions of subordinate dependence mediated the effect of abusive supervisory behavior on state sense of power. Supervisors’ perceptions of subordinate dependence were measured using the same scale as in Study 2a (α = 0.70). The regression results showed that abusive supervisory behavior was positively related to supervisors’ perceptions of subordinate dependence (b = 0.21, p < .01), which was positively related to supervisors’ state sense of power (b = 0.28, p < .01). The RMediation (Tofighi & Mackinnon, 2011) results further revealed a significant indirect effect (estimate = 0.06, 95% CI = 0.02, 0.11). These findings provided further evidence for our argument that abusive supervisory behavior enhances supervisor state sense of power by increasing perceived subordinate dependence.

90

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

Table 5 Regression results for the prediction of state sense of power and managerial self-efficacy (Study 3). Variables

State Sense of Power

Managerial Self-Efficacy

Main Effect Model

Age Gender Education Job tenure Abusive supervision Chronic sense of power Abusive supervision × Chronic sense of power State sense of power Constant R2 △R2

Moderating Effect Model

Main Effect Model

b

SE

t

b

SE

t

b

SE

t

−0.01 −0.00 −0.01 0.02 0.11

0.01 0.06 0.01 0.01 0.06

−2.11* −0.04 −1.34 2.69 2.02*

−0.01 0.01 −0.01 0.01 0.11 0.61 −0.22*

0.00 0.05 0.01 0.01 0.05 0.05 0.10

−2.18* 0.21 −1.07 1.33 2.36* 12.79*** −2.28*

0.00 −0.01 −0.01 −0.00

0.00 0.04 0.01 0.01

0.95 −0.12 −1.69 −0.05

3.94 .

0.04 0.05

98.70***

3.94

0.03 0.41 0.36***

124.81***

0.42 2.3

0.05 0.26

9.02*** 177.54*** 0.25

Note. N = 282. For abusive supervision, 1 = the abusive condition and 0 = the non-abusive condition. **p < .01. *** p < .001. * p < .05. 4.00

work focused on the impact of abusive supervision on supervisors’ power perceptions and found that within a day, abusive supervisory behavior is positively related to positive psychological outcomes related to power (i.e., state sense of power and managerial self-efficacy). These two pieces of work show that abusive supervisory behavior is a doubleedged sword for supervisors, producing both positive and negative psychological consequences. Second, our findings highlight the importance of considering the window of time when investigating the effects of abusive supervisory behavior on supervisors. Specifically, while engaging in abusive supervision is positively associated with supervisors’ state sense of power and managerial self-efficacy in the short term, the results from our supplementary analyses show that in the long term (i.e., a week and beyond), abusive supervisory behavior is negatively associated with sense of power and managerial self-efficacy. Actually, simultaneously capturing different temporal effects of abusive supervision is a key strength of our research. Investigating abusive supervision on a daily basis, we focus on leader behavior dynamics: that is, the ebb and flow of leader behavior (McClean, Barnes, Courtright, & Johnson, 2019). In the meantime, to capture the stable facet of abusive supervisory behavior, we also investigate abusive supervision from a long-term perspective. Therefore, we capture different facets of leadership by using different time windows and thus uncover both short-term and long-term effects of abusive supervision on supervisors. Similarly, Qin et al. (2018) showed that while abusive supervision is positively related to supervisors’ recovery level and work engagement in the short term, these relationships turn negative in the long term. Both our and Qin et al. (2018) research investigate the consequences of abusive supervision for supervisors themselves. However, we drew upon different theoretical lenses to develop different theoretical models. Taken together, these two pieces of work highlight the importance of considering the window of time for understanding the consequences of abusive supervision for actors. Third, by revealing the moderating role of chronic sense of power in the relationship between abusive supervisory behavior and state sense of power, our research contributes to the literature on the supervisory outcomes of abusive supervision by identifying an important boundary condition. Specifically, our findings show that only supervisors with low chronic sense of power can benefit psychologically (i.e., higher state sense of power and managerial self-efficacy) from their abusive supervisory behavior. Research on the strategic use of abusive supervision (e.g., Khan et al., 2018; Tepper et al., 2012) has implied that supervisors may abuse their subordinates to establish power within their teams. Our findings extend these studies by demonstrating that

3.50 State Sense of Power

3.00 Low chronic sense of power High chronic sense of power 2.50 Low High Abusive Supervisory Behavior

Fig. 3. The moderating role of chronic sense of power on the relationship between abusive supervisory behavior and state sense of power (Study 3).

short term (e.g., enhancing daily state sense of power and managerial self-efficacy). The findings also shed light on why supervisors may engage in abusive supervisory behavior by revealing the possibility that they do so because they can gain some immediate psychological benefits afterward. Related research has shown that supervisors may abuse their subordinates as a result of negative experiences, such as facing exceedingly difficult goals (Mawritz, Folger, & Latham, 2014), organizational injustice (Aryee, Chen, Sun, & Debrah, 2007; Tepper, Duffy, Henle, & Lambert, 2006), coworker conflicts (Harris, Harvey, & Kacmar, 2011), and family-to-work conflicts (Courtright et al., 2016). Contrary to these studies, our findings reveal the possibility that such behavior may also be motivated by positive outcomes. This notion is in line with the findings of recent studies showing that abusive supervision can be strategic such that supervisors may abuse their subordinates with calculated premeditation (Ferris, Zinko, Brouer, Buckley, & Harvey, 2007; Khan et al., 2018; Krasikova, Green, & Lebreton, 2013; Walter, Lam, van der Vegt, Huang, & Miao, 2015). Our findings extend these studies by showing that abusive supervision does have a strategic benefit for supervisors because they may benefit psychologically (i.e., feeling more powerful and having more managerial self-efficacy) from such behavior in the short term. In addition, together with the recent work of Liao et al. (2018), our research reveals that abusive supervision exerts various effects on actors. While Liao et al. (2018) work focused on the moral consequences of abusive supervision and found that within a day, abusive supervisory behavior results in negative psychological outcomes related to ethics (i.e., guilt and loss of moral credits) for supervisors themselves, our 91

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

respond to employees’ opinions negatively, such as by being unwilling to solicit and implement voice and even by denigrating employees who speak up. From the power perspective, an effective way to build supervisors’ managerial efficacy is to enhance their sense of power. In this vein, organizations should consider designing intervention mechanisms, such as involving supervisors in firm decision making, to achieve this purpose.

abusive supervision only helps supervisors with low chronic sense of power feel more powerful. However, we do not suggest that supervisors with low power should engage in abusive supervisory behavior. Rather, they should gain power in more effective ways, such as by establishing their status as experts in their area of work and by providing individualized support for their subordinates. Fourth, this research also “gives back” (Whetten, Felin, & King, 2009) to power-dependence theory, upon which it draws. Specifically, power-dependence theory posits that individuals’ power originates from other individuals’ dependence on them for valued goals and resources (Emerson, 1962). Studies drawing upon this theory have generally focused on objective control over goals and resources that others value as the source of dependence and power (e.g., Tepper et al., 2009). However, with the guidance of self-perception theory, our research shows that with the same level of objective control over resources, individuals’ state sense of power is further dependent on their own behavior. By examining the relationship between abusive supervision and state sense of power, we also answered the calls for more research on the antecedents of sense of power, especially from the behavioral perspective (for reviews, see Anderson & Brion, 2014; Magee & Galinsky, 2008; Schaerer et al., 2018; Tost, 2015). Finally, our finding of the moderating role of chronic sense of power also extends previous research (Williams et al., 2017), which provided evidence that aggressive behavior (e.g., sexual harassment) may enable acute powerholders with low chronic power to experience a feeling of power. In addition, the results of our supplementary analyses indicate that although abusive supervisory behavior is positively related to supervisors’ sense of power on the same day, the relationship turns negative in the long term. Thus, we further extend the power literature by suggesting that whether a particular behavior is power inducing or power inhibiting may actually depend on the window of time considered. Future power research should continue to consider the roles of temporal components in theorizing and empirical tests.

12.3. Strengths, limitations, and future directions This research has several key strengths, such as including different designs (diary field study versus experiment) and contexts (Western versus Chinese supervisors) to establish causality and achieve a high level of generalizability. However, our studies also have some limitations and future directions that should be discussed. First, considering that in the daily context, supervisors themselves are typically best able to report on their abusive supervisory behavior (e.g., subordinate ratings are unlikely to capture all instances of leader behavior) and other variables related to daily perceptions, the data we collected in Study 1 were self-reported. However, self-reporting can lead to concerns of common method variance (Podsakoff, Mackenzie, Podsakoff, & Lee, 2003) and the potential for reverse causality. Thus, we took several measures to mitigate these concerns in the design and analysis for Study 1. Specifically, in line with prior related studies (e.g., Koopman et al., 2016), we collected the predictors in the T1 (i.e., morning) survey and the mediator and the dependent variable in the T2 (i.e., afternoon) survey because the time lag between the independent variable and both the mediator and the dependent variable effectively alleviates concerns of common method variance and alternative explanations. We also measured the mediator at T1 and controlled for it when testing the impact of the predictors on the T2 mediator. Similarly, Koopman et al. (2016) noted that in this way, we “were able to assess change in this variable [i.e., the mediator] over the course of the day…. Examining change over the day in this manner alleviates (although not entirely) concerns about the direction of causality for [the] hypotheses” (p. 421). In addition, following other daily research (e.g., Butts, Becker, & Boswell, 2016; Koopman et al., 2016), we group-mean centered all the Level 1 variables. This approach enabled us to examine daily within-individual fluctuation while controlling for between-individual confounds. In our case, the observed moderating effects were probably not produced by common method variance (Siemsen, Roth, & Oliveira, 2010). More importantly, our experimental studies (i.e., Studies 2a, 2b, and 3) helped reduce the concern of reverse causality. Nevertheless, our design could be expanded in future research by measuring daily abusive supervisory behavior through observers’ (e.g., subordinates’) reports. Relatedly, the time lag between T1 and T2 in the diary study was nearly six hours. We think this time lag is desirable to alleviate the concern of common method variance and establish causality (Brewer, 2000), and it is in line with prior related diary studies (e.g., Lanaj, Johnson, & Lee, 2016). However, measuring state sense of power in closer proximity to abusive supervision also has some advantages as doing so may help capture the more proximal or immediate consequences of abusive supervision. Therefore, taking the length of the time lag into consideration may help minimize the possibility that our findings are contingent on the time lag. We did so as part of our supplementary analyses, in which we controlled for the time lag. The results were comparable in terms of regression coefficients and significance levels to those reported in the main text (detailed results are available from the authors upon request). Although the time-lag difference within one day may not be an issue in our research, we welcome future research to consider the length of the time lag and replicate our findings using closer proximity measures. Second, we found that abusive supervisory behavior is a predictor of supervisors’ state sense of power, but Foulk et al. (2018) found the reverse relationship. However, we do not suggest a positive loop between sense of power and abusive supervisory behavior for two reasons.

12.2. Implications for practice Our research also provides several important managerial implications. First, we find that engaging in abusive supervision is positively related to the state power perceptions of supervisors with low chronic sense of power. However, we do not recommend that supervisors abuse employees given the negative impact of abusive supervision on subordinates (Martinko et al., 2013). Rather, based on the research findings of this paper, we offer two suggestions. On the one hand, we recommend that higher-level managers reduce supervisors’ attempts to abuse subordinates to feel more powerful by enhancing supervisors’ chronic sense of power. For example, supervisors could be given reasonably higher discretion in evaluating their subordinates’ performance or allocating task assignments. Under such circumstances, they are less likely to use abusive supervisory behavior as a means to enhance their personal sense of power. On the other hand, abusive supervisors should be informed that abusive supervisory behaviors will only make them feel powerful in the short time, as suggested by our supplementary analyses. Supervisors need to be told that abusing subordinates will reduce their sense of power in the long term, which will in turn hinder their leadership effectiveness. Second, our findings show the importance of power perceptions for supervisors. The results from our daily surveys indicate that supervisors’ daily state sense of power is positively associated with their managerial self-efficacy. Supervisors’ belief that they can perform well in their managerial position matters because it results in many positive outcomes (Bandura, 1977). For example, they can engage in more taskoriented leadership behavior, which is essential to establish common goals with subordinates, clarify ambiguity, and coordinate communication (Fleishman, 1973; Judge et al., 2004). Lack of self-efficacy, however, leads to detrimental consequences. For instance, Fast et al. (2014) found that supervisors with low managerial self-efficacy tend to 92

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

supervisory outcomes may vary when subordinates react in different ways.

On the one hand, the relationship between power and abuse is contingent upon leaders’ traits. Specifically, Foulk et al. (2018) found that psychological power increases abusive supervisory behavior only when supervisors are low in agreeableness. Our studies show that abusive supervisory behavior is positively related to psychological power only for supervisors with low chronic sense of power. Therefore, the positive loop is unlikely for supervisors with high agreeableness or chronic sense of power. On the other hand, although we found that abusive supervisory behavior is positively related to state sense of power within a day, this relationship may change over a longer window of time. The results of our supplementary analyses show that abusive supervisory behavior does not predict next-day state sense of power (b = 0.12, ns) and that in the long term (e.g., a week), abusive supervision is negatively related to sense of power. Third, abusive supervision involves several important aspects including expressing anger or aggression, humiliating subordinates, treating subordinates unfairly, and so on. It is possible that the relationship between abusive supervision and state sense of power is due to the effects of one of these aspects such as anger or aggression expression, which has been shown to be associated with certainty, optimism, and control (Lerner & Keltner, 2001; Lerner, Gonzalez, Small, & Fischhoff, 2003). Thus, we encourage future research to test how different aspects of abusive supervision affect supervisors’ sense of power. Fourth, although we specified managerial self-efficacy as an outcome of sense of power, it seems plausible that high managerial selfefficacy could contribute to sense of power. As we measured managerial self-efficacy and sense of power at the same time point in both Studies 1 and 3, we are unable to rule out the possibility of reverse causality using the current data. We thus welcome future research to rule out this possibility and replicate our findings. Fifth, while we found chronic sense of power to be an important boundary condition of the relationship between abusive supervisory behavior and state sense of power, others may wonder why a leader with low chronic sense of power would engage in abusive supervisory behavior as a prior study showed that psychological power is a precondition of abusive supervisory behavior (Foulk et al., 2018). We argue that leaders with low chronic sense of power may still engage in abusive supervisory behavior for three reasons. First, all leaders have a certain degree of latitude to engage in abusive supervision as they usually possess structural (or positional) power over subordinates (Tepper, 2000). Second, leaders with low chronic sense of power may be even more motivated to engage in abusive supervision in order to restore or experience a sense of power. This notion is in line with previous research suggesting that supervisors with low (versus high) chronic sense are more sensitive to actions that may increase their power (Fiske & Dépret, 1996; Hays & Goldstein, 2015). Third, besides psychological power, there are other factors, such as lack of self-regulatory resources (Courtright et al., 2016; Yam, Fehr, Keng-Highberger, Klotz, & Reynolds, 2016), that trigger leaders’ abusive supervisory behavior. Thus, a supervisor with low chronic sense of power can still engage in abusive supervision. Finally, future studies may explore more actor-centric effects of abusive supervision and more boundary conditions for such effects. For example, abusive supervision may be related to other positive supervisor psychological states, such as self-esteem, perceived status, satisfaction of competence needs, and work passion (Chen, Liu, & He, 2015). More interestingly, abusive supervision may also be related to negative supervisor psychological states, such as feeling dirty. In addition, while we explored task-oriented leadership in this research, it would also be interesting to explore how abusive supervision relates to other types of leadership behavior, such as person-oriented leadership. Moreover, subordinates may react to supervisor abuse in different ways, such as silence and obedience (Nandkeolyar, Shaffer, Li, Ekkirala, & Bagger, 2014; Schat, Frone, & Kelloway, 2006) or direct confrontation. Therefore, the ways subordinates react to abusive supervision may be potential moderators such that the effects of abusive supervision on

12.4. Conclusion This research investigates the impacts of abusive supervisory behavior on actors’ (i.e., supervisors’) self-perceptions as managers by integrating self-perception theory and power-dependence theory. Using a daily field study and three experiments, this research highlights the short-term (versus long-term) benefits of abusive supervisory behavior for actors and the boundary conditions. We hope this work sparks scholars’ interests in examining the impacts of abusive supervisory behavior on actors and provides more refined suggestions for reducing such behavior in organizational settings. Acknowledgements We thank Associate Editor Linn Van Dyne and three anonymous reviewers for their tremendously insightful and constructive comments. This research was supported by grants funded by National Natural Science Foundation of China (Grant No. 71702175, 71602032, 71872190, 71502179, and 71902172), and Guangdong Province Higher Vocational Colleges & Schools Pearl River Scholar Funded Scheme (2018). Appendix A. Problem-solving scenario for measuring taskoriented leadership behavior NTSoft is a corporation specializing primarily in developing and marketing enterprise software products. Currently, the company is developing a management information system (MIS) software for a customer company. To complete this project, there are several sub-teams, including a developer team, a test engineer team, an interface designer team, etc. Alex is the leader of the developer sub-team, which has seven team members. The whole project is supposed to be completed in one year. However, the progress is way behind schedule now, and the team may not be able to finish on time. The slow progress is caused by many factors. In particular, the developer sub-team led by Alex has not been performing very well. The progress of this sub-team is also behind schedule. Three of the seven members on Alex’s team have been performing below expectations. There are also some collaboration problems within this sub-team and coordination problems with other subteams in this project. Appendix B. Supplementary material Supplementary data to this article can be found online at https:// doi.org/10.1016/j.obhdp.2019.09.003. References Anderson, C., & Brion, S. (2014). Perspectives on power in organizations. Annual Review of Organizational Psychology and Organizational Behavior, 1, 67–97. Anderson, C., & Galinsky, A. D. (2006). Power, optimism, and risk-taking. European Journal of Social Psychology, 36, 511–536. Anderson, C., John, O. P., & Keltner, D. (2012). The personal sense of power. Journal of Personality, 80, 313–344. Antes, A. L., & Mumford, M. D. (2012). Strategies for leader cognition: Viewing the glass “half full” and “half empty”. Leadership Quarterly, 23, 425–442. Aquino, K., Tripp, T. M., & Bies, R. J. (2001). How employees respond to personal offense: The effects of blame attribution, victim status, and offender status on revenge and reconciliation in the workplace. Journal of Applied Psychology, 86, 52–59. Aryee, S., Chen, Z. X., Sun, L. Y., & Debrah, Y. A. (2007). Antecedents and outcomes of abusive supervision: Test of a trickle-down model. Journal of Applied Psychology, 92, 191–201. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191–215.

93

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

& J. G. Hunt (Eds.). Current developments in the study of leadership (pp. 1–40). Carbondale, IL: Southern Illinois University Press. Foulk, T., Lanaj, K., Tu, M. H., Erez, A., & Archambeau, L. (2018). Heavy is the head that wears the crown: An actor-centric approach to daily psychological power, abusive leader behavior, and perceived incivility. Academy of Management Journal, 61, 661–684. Goodstadt, B. E., & Hjelle, L. A. (1973). Power to the powerless: Locus of control and the use of power. Journal of Personality and Social Psychology, 27, 190–196. Harris, K. J., Harvey, P., & Kacmar, K. M. (2011). Abusive supervisory reactions to coworker relationship conflict. Leadership Quarterly, 22, 1010–1023. Hannah, S. T., Avolio, B. J., Luthans, F., & Harms, P. D. (2008). Leadership efficacy: Review and future directions. Leadership Quarterly, 19, 669–692. Hays, N. A., & Goldstein, N. J. (2015). Power and legitimacy influence conformity. Journal of Experimental Social Psychology, 60, 17–26. Hayes, A. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York: Guilford. Hershcovis, M. S., & Barling, J. (2010). Towards a multi-foci approach to workplace aggression: A meta-analytic review of outcomes from different perpetrators. Journal of Organizational Behavior, 31, 24–44. Hofmann, D. A., Griffin, M. A., & Gavin, M. B. (2000). The application of hierarchical linear modeling to organizational research. In K. J. Klein, & S. W. J. Kozlowski (Eds.). Multilevel theory, research, and methods in organizations (pp. 467–511). San Francisco: Jossey-Bass. Hofstede, G. H. (2001). Culture’s consequences: Comparing values, behaviors, institutions and organizations across nations (2nd ed.). Thousand Oaks, CA: Sage. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1–55. Jin, J. F., Ford, M. T., & Chen, C. C. (2013). Asymmetric differences in work-family spillover in North America and China: Results from two heterogeneous samples. Journal of Business Ethics, 113, 1–14. Johnson, R. E., Lanaj, K., & Barnes, C. M. (2014). The good and bad of being fair: Effects of procedural and interpersonal justice behaviors on regulatory resources. Journal of Applied Psychology, 99, 635–650. Johnson, R. E., Rosen, C. C., & Djurdjevic, E. (2011). Assessing the impact of common method variance on higher-order multidimensional constructs. Journal of Applied Psychology, 96, 744–761. Johnson, R. E., Venus, M., Lanaj, K., Mao, C. G., & Chang, C. H. (2012). Leader identity as an antecedent of the frequency and consistency of transformational, consideration, and abusive leadership behaviors. Journal of Applied Psychology, 97, 1262–1272. Judge, T. A., Piccolo, R. F., & Ilies, R. (2004). The forgotten ones? The validity of consideration and initiating structure in leadership research. Journal of Applied Psychology, 89, 36–51. Judge, T. A., & Bono, J. E. (2001). Relationship of core self-evaluations traits–self-esteem, generalized self-efficacy, locus of control, and emotional stability–with job satisfaction and job performance: A meta-analysis. Journal of Applied Psychology, 86, 80–92. Kane, T. D., Zaccaro, S. J., Tremble, T. R., & Masuda, A. D. (2002). An examination of the leader's regulation of groups. Small Group Research, 33, 65–120. Khan, A. K., Moss, S., Quratulain, S., & Hameed, I. (2018). When and how subordinate performance leads to abusive supervision: A social dominance perspective. Journal of Management, 44, 2801–2826. Kipnis, D. (1976). The powerholders. Chicago: University of Chicago Press. Koopman, J., Lanaj, K., & Scott, B. A. (2016). Integrating the bright and dark sides of OCB. Academy of Management Journal, 59, 414–435. Krasikova, D. V., Green, S. G., & Lebreton, J. M. (2013). Destructive leadership: A theoretical review, integration, and future research agenda. Journal of Management, 39, 1308–1338. Lambert, L. S., Tepper, B. J., Carr, J. C., Holt, D. T., & Barelka, A. J. (2012). Forgotten but not gone: An examination of fit between leader consideration and initiating structure needed and received. Journal of Applied Psychology, 97, 913–930. Lanaj, K., Johnson, R. E., & Lee, S. M. (2016). Benefits of transformational behaviors for leaders: A daily investigation of leader behaviors and need fulfillment. Journal of Applied Psychology, 101, 237–251. Lerner, J. S., Gonzalez, R. M., Small, D. A., & Fischhoff, B. (2003). Effects of fear and anger on perceived risks of terrorism: A national field experiment. Psychological Science, 14, 144–150. Lerner, J. S., & Keltner, D. (2001). Fear, anger, and risk. Journal of Personality and Social Psychology, 81, 146–159. Liang, L. H., Lian, H., Brown, D. J., Ferris, D. L., Hanig, S., & Keeping, L. M. (2016). Why are abusive supervisors abusive? A dual-system self-control model. Academy of Management Journal, 59, 1385–1406. Liao, Z., Yam, K. C., Johnson, R. E., Liu, W., & Song, Z. (2018). Cleansing my abuse: A reparative response model of perpetrating abusive supervisor behavior. Journal of Applied Psychology, 103, 1039–1056. Lin, S. H., & Johnson, R. E. (2015). A suggestion to improve a day keeps your depletion away: Examining promotive and prohibitive voice behaviors within a regulatory focus and ego depletion framework. Journal of Applied Psychology, 100, 1381–1397. Liu, X. X., Keller, J., & Hong, Y. Y. (2015). Hiring of personal ties: A cultural consensus analysis of China and the United States. Management and Organization Review, 11, 145–169. Lord, R. G., & Hall, R. J. (2005). Identity, deep structure and the development of leadership skill. Leadership Quarterly, 16, 591–615. Lu, C. Q., Siu, O. L., & Cooper, C. L. (2005). Managers’ occupational stress in China: The role of self-efficacy. Personality & Individual Differences, 38, 569–578. Mackey, J. D., Frieder, R. E., Brees, J. R., & Martinko, M. J. (2017). Abusive supervision: A meta-analysis and empirical review. Journal of Management, 43, 1940–1965.

Bandura, A. (1995). Self-efficacy in changing societies. New York: Cambridge University Press. Bandura, A., & Locke, E. A. (2003). Negative self-efficacy and goal effects revisited. Journal of Applied Psychology, 88, 87–99. Barnes, C. M., Lucianetti, L., Bhave, D. P., & Christian, M. S. (2015). You wouldn't like me when I'm sleepy: Leaders’ sleep, daily abusive supervision, and work unit engagement. Academy of Management Journal, 58, 1419–1437. Bem, D. J. (1982). Self-perception theory. In L. Berkowitz (Ed.). Advances in experimental social psychology (pp. 1–62). New York: Academic Press. Bernerth, J. B., Taylor, S. G., Walker, H. J., & Whitman, D. S. (2012). An empirical investigation of dispositional antecedents and performance-related outcomes of credit scores. Journal of Applied Psychology, 97, 469–478. Berry, C. M., Carpenter, N. C., & Barratt, C. L. (2012). Do other-reports of counterproductive work behavior provide an incremental contribution over self-reports? A meta-analytic comparison. Journal of Applied Psychology, 97, 613–636. Bobocel, D. R. (2013). Coping with unfair events constructively or destructively: The effects of overall justice and self-other orientation. Journal of Applied Psychology, 98, 720–731. Brewer, M. B. (2000). Research design and issues of validity. In H. T. Reis, & C. M. Judd (Eds.). Handbook of research methods in social and personality psychology (pp. 3–16). New York: Cambridge University Press. Briñol, P., Petty, R. E., Valle, C., Rucker, D. D., & Becerra, A. (2007). The effects of message recipients' power before and after persuasion: A self-validation analysis. Journal of Personality and Social Psychology, 93, 1040–1053. Brislin, R. W. (1980). Translation and content analysis of oral and written material. In H. C. Triandis, & J. W. Berry (Eds.). Handbook of cross-cultural psychology (pp. 389–444). Boston: Allyn & Bacon. Bryk, A. S., Raudenbush, S. W., & Congdon, R. T. (1996). HLM: Hierarchical linear and nonlinear modeling with the HLM/2L and HLM/3L programs. Chicago: Scientific Software International. Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechanical Turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6, 3–5. Butts, M. M., Becker, W. J., & Boswell, W. R. (2016). Hot buttons and time sinks: The effects of electronic communication during nonwork time on emotions and worknonwork conflict. Academy of Management Journal, 58, 763–788. Casciaro, T., Gino, F., & Kouchaki, M. (2014). The contaminating effects of building instrumental ties: How networking can make us feel dirty. Administrative Science Quarterly, 59, 705–735. Chen, G., Gully, S. M., & Eden, D. (2001). Validation of a new general self-efficacy scale. Organizational Research Methods, 4, 62–83. Chen, X. P., Liu, D., & He, W. (2015). Does passion fuel entrepreneurship and job creativity? A review and preview of passion research. The Oxford Handbook of creativity, innovation and entrepreneurship (pp. 159–175). Courtright, S. H., Gardner, R. G., Smith, T. A., Mccormick, B. W., & Colbert, A. E. (2016). My family made me do it: A cross-domain, self-regulatory perspective on antecedents to abusive supervision. Academy of Management Journal, 59, 1630–1652. Day, D. V., Harrison, M. M., & Halpin, S. M. (2009). An integrative approach to leader development: Connecting adult development, identity, and expertise. New York: Routledge. DeRue, D. S., Nahrgang, J. D., Wellman, N., & Humphrey, S. E. (2011). Trait and behavioral theories of leadership: An integration and meta-analytic test of their relative validity. Personnel Psychology, 64, 7–52. Dubois, D., Rucker, D. D., & Galinsky, A. D. (2015). Social class, power, and selfishness: When and why upper and lower class individuals behave unethically. Journal of Personality and Social Psychology, 108, 436–449. Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation and mediation: A general analytical framework using moderated path analysis. Psychological Methods, 12, 1–22. Eisenberger, R., Stinglhamber, F., Vandenberghe, C., Sucharski, I. L., & Rhoades, L. (2002). Perceived supervisor support: Contributions to perceived organizational support and employee retention. Journal of Applied Psychology, 87, 565–573. Emerson, R. M. (1962). Power-dependence relations. American Sociological Review, 27, 31–41. Fast, N. J., Burris, E. R., & Bartel, C. A. (2014). Managing to stay in the dark: Managerial self-efficacy, ego defensiveness, and the aversion to employee voice. Academy of Management Journal, 57, 1013–1034. Fast, N. J., Gruenfeld, D. H., Sivanathan, N., & Galinsky, A. D. (2009). Illusory control: A generative force behind power’s far-reaching effects. Psychological Science, 20, 502–508. Fast, N. J., Sivanathan, N., Mayer, N. D., & Galinsky, A. D. (2012). Power and overconfident decision-making. Organizational Behavior and Human Decision Processes, 117, 249–260. Ferris, D. L., Yan, M., Lim, V. K. G., Chen, Y., & Fatimah, S. (2016). An approachavoidance framework of workplace aggression. Academy of Management Journal, 59, 1777–1800. Ferris, G. R., Zinko, R., Brouer, R. L., Buckley, M. R., & Harvey, M. G. (2007). Strategic bullying as a supplementary, balanced perspective on destructive leadership. Leadership Quarterly, 18, 195–206. Fisher, C. D., & To, M. L. (2012). Using experience sampling methodology in organizational behavior. Journal of Organizational Behavior, 33, 865–877. Fiske, S. T., & Dépret, E. (1996). Control, interdependence and power: understanding social cognition in its social context. European Review of Social Psychology, 7, 31–61. Fleishman, E. A. (1953). The description of supervisory behavior. Journal of Applied Psychology, 37, 1–6. Fleishman, E. A. (1973). Twenty years of consideration and structure. In E. A. Fleishman,

94

Organizational Behavior and Human Decision Processes 154 (2019) 80–95

D. Ju, et al.

Siemsen, E., Roth, A., & Oliveira, P. (2010). Common method bias in regression models with linear, quadratic, and interaction effects. Organizational Research Methods, 13, 456–476. Sikora, D. M., Ferris, G. R., & Van Iddekinge, C. H. (2015). Line manager implementation perceptions as a mediator of relations between high-performance work practices and employee outcomes. Journal of Applied Psychology, 100, 1908–1918. Stogdill, R. M. (1963). Manual for the leader behavior description questionnaire-form XII: An experimental revision. Columbus, Ohio: Bureau of Business Research, Ohio State University. Tepper, B. J. (2000). Consequences of abusive supervision. Academy of Management Journal, 43, 178–190. Tepper, B. J. (2007). Abusive supervision in work organizations: Review, synthesis, and research agenda. Journal of Management, 33, 261–289. Tepper, B. J., Carr, J. C., Breaux, D. M., Geider, S., Hu, C. Y., & Hua, W. (2009). Abusive supervision, intentions to quit, and employees’ workplace deviance: A power/dependence analysis. Organizational Behavior and Human Decision Processes, 109, 156–167. Tepper, B. J., Duffy, M. K., & Breaux-Soignet, D. M. (2012). Abusive supervision as political activity: Distinguishing impulsive and strategic expressions of downward hostility. In G. Ferris, & D. Treadway (Eds.). Politics in organizations: Theory and research considerations (pp. 191–212). New York: Routledge. Tepper, B. J., Duffy, M. K., Henle, C. A., & Lambert, L. S. (2006). Procedural injustice, victim precipitation, and abusive supervision. Personnel Psychology, 59, 101–123. Tepper, B. J., Duffy, M. K., & Shaw, J. D. (2001). Personality moderators of the relationship between abusive supervision and subordinates’ resistance. Journal of Applied Psychology, 86, 974–983. Tofighi, D., & Mackinnon, D. P. (2011). RMediation: An R package for mediation analysis confidence intervals. Behavior Research Methods, 43, 692–700. Tost, L. P. (2015). When, why, and how do powerholders “feel the power”? Examining the links between structural and psychological power and reviving the connection between power and responsibility. Research in Organizational Behavior, 35, 29–56. Tost, L. P., Gino, F., & Larrick, R. P. (2012). Power, competitiveness, and advice taking: Why the powerful don’t listen. Organizational Behavior and Human Decision Processes, 117, 53–65. Tost, L. P., Gino, F., & Larrick, R. P. (2013). When power makes others speechless: The negative impact of leader power on team performance. Academy of Management Journal, 56, 1465–1486. van Kleef, G. A., Homan, A. C., Finkenauer, C., Gündemir, S., & Stamkou, E. (2011). Breaking the rules to rise to power: How norm violators gain power in the eyes of others. Social Psychological and Personality Science, 2, 500–507. Walter, F., Lam, C. K., van der Vegt, G. S., Huang, X., & Miao, Q. (2015). Abusive supervision and subordinate performance: Instrumentality considerations in the emergence and consequences of abusive supervision. Journal of Applied Psychology, 100, 1056–1072. Waytz, A., Chou, E. Y., Magee, J. C., & Galinsky, A. D. (2015). Not so lonely at the top: The relationship between power and loneliness. Organizational Behavior and Human Decision Processes, 130, 69–78. Wee, E. X. M., Liao, H., Liu, D., & Liu, J. (2017). Moving from abuse to reconciliation: A power-dependence perspective on when and how a follower can break the spiral of abuse. Academy of Management Journal, 60, 2352–2380. Whetten, D. A., Felin, T., & King, B. G. (2009). The practice of theory borrowing in organizational studies: Current issues and future directions. Journal of Management, 35, 537–563. Williams, M. J., Gruenfeld, D. H., & Guillory, L. E. (2017). Sexual aggression when power is new: Effects of acute high power on chronically low-power individuals. Journal of Personality and Social Psychology, 112, 201–223. Wood, R., & Bandura, A. (1989). Impact of conceptions of ability on self-regulatory mechanisms and complex decision making. Journal of Personality and Social Psychology, 56, 407–415. Yam, K. C., Fehr, R., Keng-Highberger, F. T., Klotz, A. C., & Reynolds, S. J. (2016). Out of control: A self-control perspective on the link between surface acting and abusive supervision. Journal of Applied Psychology, 101, 292–301. Yukl, G. (1971). Toward a behavioral theory of leadership. Organizational Behavior and Human Performance, 6, 414–440. Zaccaro, S. J., Connelly, S., Repchick, K. M., Daza, A. I., Young, M. C., Kilcullen, R. N., et al. (2015). The influence of higher order cognitive capacities on leader organizational continuance and retention: The mediating role of developmental experiences. Leadership Quarterly, 26, 342–358. Zaccaro, S. J., Mumford, M. D., Connelly, M. S., Marks, M. A., & Gilbert, J. A. (2000). Assessment of leader problem-solving capabilities. Leadership Quarterly, 11, 37–64.

Mackinnon, D. P., Fritz, M. S., Williams, J., & Lockwood, C. M. (2007). Distribution of the product confidence limits for the indirect effect: Program PRODCLIN. Behavior Research Methods, 39, 384–389. Magee, J. C. (2009). Seeing power in action: The roles of deliberation, implementation, and action in inferences of power. Journal of Experimental Social Psychology, 45, 1–14. Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self-reinforcing nature of power and status. Academy of Management Annals, 2, 351–398. Martinko, M. J., Harvey, P., Brees, J. R., & Mackey, J. (2013). A review of abusive supervision research. Journal of Organizational Behavior, 34, S120–S137. Mawritz, M. B., Folger, R., & Latham, G. P. (2014). Supervisors’ exceedingly difficult goals and abusive supervision: The mediating effects of hindrance stress, anger, and anxiety. Journal of Organizational Behavior, 35, 358–372. McClean, S., Barnes, C. M., Courtright, S. H., & Johnson, R. E. (2019). Resetting the clock on dynamic leader behaviors: a conceptual integration and agenda for future research. Academy of Management Annals, 13, 479–508. Meade, A. W., & Craig, S. B. (2012). Identifying careless responses in survey data. Psychological Methods, 17, 437–455. Mintzberg, H. (1983). Power in and around organizations. Englewood Cliffs, N.J.: Prentice Hall. Mitchell, M. S., & Ambrose, M. L. (2007). Abusive supervision and workplace deviance and the moderating effects of negative reciprocity beliefs. Journal of Applied Psychology, 92, 1159–1168. Molm, L. D. (1991). Affect and social exchange: Satisfaction in power-dependence relations. American Sociological Review, 56, 475–493. Morrison, E. W., See, K. E., & Pan, C. (2015). An approach-inhibition model of employee silence: The joint effects of personal sense of power and target openness. Personnel Psychology, 68, 547–580. Mourali, M., & Yang, Z. (2013). The dual role of power in resisting social influence. Journal of Consumer Research, 40, 539–554. Nandkeolyar, A. K., Shaffer, J. A., Li, A., Ekkirala, S., & Bagger, J. (2014). Surviving an abusive supervisor: The joint roles of conscientiousness and coping strategies. Journal of Applied Psychology, 99, 138–150. Neubert, M. J., Kacmar, K. M., Carlson, D. S., Chonko, L. B., & Roberts, J. A. (2008). Regulatory focus as a mediator of the influence of initiating structure and servant leadership on employee behavior. Journal of Applied Psychology, 93, 1220–1233. Ng, K. Y., Ang, S., & Chan, K. Y. (2008). Personality and leader effectiveness: A moderated mediation model of leadership self-efficacy, job demands, and job autonomy. Journal of Applied Psychology, 93, 733–743. Paolacci, G., Chandler, J., & Ipeirotis, P. G. (2010). Running experiments on Amazon Mechanical Turk. Judgment and Decision Making, 5, 411–419. Parker, S. K., Williams, H. M., & Turner, N. (2006). Modeling the antecedents of proactive behavior at work. Journal of Applied Psychology, 91, 636–652. Podsakoff, P. M., Mackenzie, S. B., Podsakoff, N. P., & Lee, J. Y. (2003). The mismeasure of management and its implications for leadership research. Leadership Quarterly, 14, 615–656. Qin, X., Huang, M., Hu, Q., Schminke, M., & Ju, D. (2018). Ethical leadership, but toward whom? How moral identity congruence shapes the ethical treatment of employees. Human Relations, 71(8), 1120–1149. Qin, X., Huang, M., Johnson, R., Hu, Q., & Ju, D. (2018). The short-lived benefits of abusive supervisory behavior for actors: An investigation of recovery and work engagement. Academy of Management Journal, 61, 1951–1975. Qin, X., Ren, R., Zhang, Z., & Johnson, R. E. (2018). Considering self-interests and symbolism together: How instrumental and value-expressive motives interact to influence supervisors’ justice behavior. Personnel Psychology, 71(2), 225–253. Robinson, S. L., & Bennett, R. J. (1995). A typology of deviant workplace behaviors: A multidimensional scaling study. Academy of Management Journal, 38, 555–572. Rothbard, N. P., & Wilk, S. L. (2011). Waking up on the right or wrong side of the bed: Start-of-workday mood, work events, employee affect, and performance. Academy of Management Journal, 54, 959–980. Rothman, N. B. (2011). Steering sheep: How expressed emotional ambivalence elicits dominance in interdependent decision making contexts. Organizational Behavior and Human Decision Processes, 116, 66–82. Schat, A., Frone, M. R., & Kelloway, E. K. (2006). Prevalence of workplace aggression in the U.S. workforce: Findings from a national study. In E. K. Kelloway, J. Barling, & J. J. J. Hurrell (Eds.). Handbook of workplace violence (pp. 47–89). Thousand Oaks, CA: Sage. Schaerer, M., Tost, L. P., Huang, L., Gino, F., & Larrick, R. (2018). Advice giving: A subtle pathway to power. Personality and Social Psychology Bulletin, 44, 746–761. Scott, B. A., & Barnes, C. M. (2011). A multilevel field investigation of emotional labor, affect, work withdrawal, and gender. Academy of Management Journal, 54, 116–136.

95