A Resource Perspective on Abusive Supervision and Extra-Role Behaviors: The Role of Subordinates’ Psychological Capital

Abstract

Abusive supervision (perceived enduring hostile verbal and nonverbal behavior) results in a host of detrimental consequences for the individual subordinate and for the organization. In the current research, we tested whether abusive supervision relates negatively to beneficial extra-role behaviors of subordinates (individual-directed and organization-directed citizenship behaviors; OCBI and OCBO) and positively to deviant extra-role behaviors of subordinates (individual-directed and organization-directed counterproductive work behavior; CWBI and CWBO). Moreover, reasoning from a resource perspective, we examined whether subordinates’ psychological capital (PsyCap: hope, resilience, self-efficacy, and optimism) mediates these relations. PsyCap is a resource variable that is amenable to situational influences such as leadership. This makes PsyCap align with a theoretically viable, but previously not explicitly tested, mechanism underlying the effects of abusive supervision. We conducted a time-lagged, multisource study among 408 university faculty members. Abusive supervision and PsyCap were measured at Time 1 from focal participants. At Time 2, data for OCBs were collected from their supervisors and data for CWBs were collected from their peers. Results indicate that PsyCap mediated the relations between abusive supervision and OCBI, OCBO, CWBI, and CWBO. Shedding light on this process helps researchers and practitioners develop ways in which to mitigate the consequences of abusive supervision, for example, by seeking to develop PsyCap using different resources.

Keywords

positive organizational behavior employee behaviors human resources leadership behavior

Abusive supervision (“ . . . subordinates’ perceptions of the extent to which their supervisors engage in the sustained display of hostile verbal and nonverbal behaviors, excluding physical contact”; Tepper, 2000, p. 178) is a type of workplace mistreatment that has gained momentum in research and practice. Indeed, “the boss” may be one of the major factors that can lead employees to experience stress in their jobs (Michie, 2002) and literature shows far-reaching consequences of abusive supervision for employee attitudes and behavior (Martinko, Harvey, Brees, & Mackey, 2013). In addition to affecting employee well-being, abusive supervision affects employees’ discretionary behavior. Abusive supervision predicts reductions in positive discretionary behaviors (e.g., Xu, Huang, Lam, & Miao, 2012) such as organizational citizenship behavior (OCB). These desirable OCBs can entail behaviors such as helping colleagues (individual-directed OCB; OCBI) or attending nonmandatory organizational meetings (organization-directed OCB; OCBO). Abusive supervision also relates positively to negative discretionary behaviors (Mitchell & Ambrose, 2007), such as counterproductive work behaviors (CWB). CWBs can entail behaviors such as ridiculing or embarrassing coworkers (individual-directed CWB) or leaving work early and taking longer breaks (organization-directed CWB).

Recent research suggests that, when employees have better resources to cope with abuse, they may show less of the detrimental behavioral effects (e.g., Frieder, Hochwarter, DeOrtentiis, 2015; Nandkeolyar, Shaffer, Li, Ekkirala, & Bagger, 2014). Working from such findings, we sought to examine a potential mechanism linking abusive supervision to its outcomes. From such findings, it follows that abusive supervision constitutes a resource-consuming factor (Harris, Kacmar, & Zivnuska, 2007; Kacmar, Whitman, & Harris, 2013). Abuse, in and of itself, is a stressful event that individuals need to deal with. Moreover, given the definition of abusive supervision as enduring, coping requires resources expenditure. In turn, a lack of resources implies reduced ability to engage in constructive behaviors, such as OCBs, and a lack of resources increases negative behaviors, such as CWBs (Fox, Spector, & Miles, 2001).

Research has examined related variables such as strain and burnout that arise from abusive supervision (e.g., Carlson, Ferguson, Hunter, & Whitten, 2012). Working from a resource perspective (Hobfoll, 1989), we examine whether psychological capital (PsyCap; Luthans, Avolio, Avey, & Norman, 2007) mediates the relations between abusive supervision and extra-role behaviors. PsyCap is a resource-variable of resilience, optimism, hope, and self-efficacy and affects behavioral and performance outcomes (Avey, Reichard, Luthans, & Mhatre, 2011). Individuals with high PsyCap have a strong belief in their ability to regulate goal pursuit, they form positive anticipations in dealing with challenging situations and they positively influence and contribute in harmonizing adverse situations (Luthans et al., 2007). Moreover, PsyCap is a resource (Luthans & Youssef, 2004), that, rather than being a stable trait, it is susceptible to situational influences. This implies that situational, resource-exhausting factors may be associated with decreases in PsyCap as well. Specifically, we propose that abusive supervision is negatively related to PsyCap, which mediates the relation between abusive supervision and OCBs and CWBs (see Figure 1).

Figure 1.

Conceptual framework of psychological capital as a mediator of the links between abusive supervision and OCBs and CWBs.

This research provides several contributions. First, we examine PsyCap as one resource that may contribute to the outcomes associated with abusive supervision. Doing so broadens accounts for the outcomes of abusive supervision, which aids understanding of this process. Second, this study contributes to the PsyCap literature. A core premise of the concept of PsyCap is that PsyCap is amenable to outside influences. Support for our hypotheses would yield indirect support for this basic notion as it would tentatively suggest that PsyCap is associated negatively with resource-expending situational variables such as abusive supervision. Hence, this research addresses a call for better understanding the antecedents of PsyCap (e.g., Avey, 2014). Finally, this research has practical implications. Abusive supervision is thought to constitute a serious threat and to have spillover effects to organizational functioning as well as to the well-being of individuals. If the current proposition is supported, it indicates specific practical advice for dealing with, or mitigating, the consequences of abusive supervision. That is, understanding the mechanism underlying the effects of abusive supervision will allow intervention into a specific process as well. Such knowledge may enable developing ways in which individuals can restore or strengthen their PsyCap through other sources, such as training and development (e.g., Luthans, Avey, & Patera, 2008), which might be one way of mitigating consequences.

Abusive Supervision

The abusive supervision concept was introduced by Tepper (2000), stating that abusive supervision involves hostile verbal and nonverbal behaviors, but not physical contact (which is part of violence). According to this conceptualization, abusive supervision is a sustained pattern, and abusive supervision centers on the perception of the subordinate concerning such things as public mockery, invasion of privacy, wrongful blame, rudeness, breach of promises, selfishness, and discrimination in information sharing procedures. This implies that a behavior may be abusive in one context, but not in another, and it entails that abusive supervision is a subordinate-level construct. Abusive supervision is particularly insidious because targets of abuse may often be unable to change their position, implying they often (need to) stay in these situations without the prospect of improvement.

Abusive supervision decreases satisfaction and commitment, and increases turnover intentions, stress, burnout, physical symptoms, and feelings of frustration and helplessness (Tepper, Henle, Lambert, Giacalone, & Duffy, 2008). Task performance also suffers (Jian, Kwan, Qiu, Liu, & Yim, 2012) as do other variables such as behaviors individuals engage in that are not part of formal roles such as citizenship behaviors that either help colleagues (OCBI) or help the organization (OCBO; see X. Y. Liu, & Wang, 2013). Moreover, abusive supervision may predict increases in CWB, toward colleagues (CWBI; e.g., ridiculing and embarrassing others) and toward the organization (CWBO; e.g., leaving work early and taking longer breaks; see Mitchell & Ambrose, 2007). Research has provided several explanations for these findings and in the current research we rely on the notion that abusive supervision forms a stress factor in employees’ environment, which may affect subordinates psychological resources (Hobfoll, 1989).

A Resource Perspective

Conservation of resources theory states that people who are confronted with stress will seek to minimize the net loss of resources. Resources are valued for two reasons. First, resources have instrumental value: they help people attain valued outcomes. Second, resources have symbolic value: they help people to define who they are (Hobfoll, 1989). This model, accordingly, posits the existence of a quantity of resources that help people attain their objectives. This quantity can be depleted, replenished, or compensated. Resources used for pursuing one goal may not be available to pursue other goals. Furthermore, people use resources relatively strategically. Hence, they might seek to replenish resources by trying to tap into other sources or saving resources by withholding effort.

Abusive supervision affects subordinates’ resources in terms of (a) the time and effort that goes into managing their work environment and supervisor, (b) subordinates’ sense of (loss of) control over their environment, and (c) the necessity of dealing with the emotional impact of abuse. First, in the presence of an abusive supervision, time and effort may be required simply to prevent as much as possible the abusive behavior from occurring, leaving employees too preoccupied to engage in extra-role behavior (Harris et al., 2007; Kacmar et al., 2013). Second, losing sense of control is a direct consequence of abusive behavior such as being denied a say in decisions affecting one’s own situation (Tepper, 2000) and such as the threat to withhold resources necessary for job performance (Harris et al., 2007). Third, the outcomes of stress, anxiety, burnout, and so forth (Carlson et al., 2012; Harvey, Stoner, Hochwarter, & Kacmar, 2007), indicate that abusive supervision in itself is something that requires coping. Indeed, research indicates that abusive supervision prompts subordinates’ engagement in emotional labor (i.e., “acting”; Carlson et al., 2012).

Recent research also shows several other indications that abusive supervision is resource demanding. First, abusive supervision predicts subordinates’ tendency to engage in certain coping behaviors (Yagil, Ben-Zur, & Tamir, 2011). Second, employees’ tendency to use certain coping strategies plays a moderating role in the abusive supervision process (Nandkeolyar, Shaffer, Li, Ekkirala, & Bagger, 2014). Third, subordinates’ ability to manage resources moderates the link between abusive supervision and detrimental outcomes (Frieder et al., 2015). Abusive supervision, thus, directly implies reduced control over subordinates’ own situation; they are less able to effect changes and actions, implying that their resources to do their jobs are reduced. It also implies that targets of abuse need to cope with this behavior, leaving reduced resources for constructive behaviors. Thus, abusive supervision may predict reduced OCBs.

There is also reason to hypothesize involvement of resources in the relations between abusive supervision and CWBs. First, considering the negative behaviors toward colleagues (CWBI), Duffy, Ganster, Shaw, Johnson, and Pagon (2006) argued that the insidious nature of undermining behaviors that the abusive supervisor shows, will consequently hinder employees’ ability to maintain a good impression in front of their coworkers. Likewise, Schat, Frone, and Kelloway (2006) argued that the irritation, anger, frustration, and helplessness that subordinates of an abusive supervision feel, relate to aggression toward coworkers (Spector, 1998). Second, considering behavior toward the organization (CWBO), variables such as somatic complaints, resulting from an inability to cope and generally representing an exhaustion of mental and physical resources, might explain why individuals call in sick when they are not really sick, take extra breaks, and leave work early. In addition, telling other people outside the organization that it is a lousy place to work and complaining about insignificant things likely also results from experiencing the work place as exhausting.

While our contention is not that the justice explanation is incorrect, the argumentation leads us to conclude that there may also be a resource mechanism in the link between abusive supervision and certain extra-role behaviors. Other negative behaviors, perhaps even more insidious, such as theft and fraud, may nevertheless be better explained by the motivation to retaliate, invoking a justice or exchange mechanism. We sought to contribute to the literature on abusive supervision by testing a resource perspective. Specifically, we suggest that a logical candidate for this resource is PsyCap (Luthans et al., 2007; Youssef & Luthans, 2007), which has been proposed as a key resource variable in the human capital of organizations due to its influence on individual performance (Avey et al., 2011).

Psychological Capital as a Mediator

PsyCap is a composite characteristic consisting of resilience, optimism, hope, and self-efficacy. High levels of PsyCap means

. . . having confidence (self-efficacy) to take on and put in the necessary effort to succeed at challenging tasks; (2) making a positive attribution (optimism) about succeeding now and in the future; (3) persevering toward goals and, when necessary, redirecting paths to goals (hope) in order to succeed; and (4) when beset by problems and adversity, sustaining and bouncing back and even beyond (resilience) to attain success. (Luthans et al., 2007, p. 3)

PsyCap implies the resource to think of many ways to reach goals, to recover from troubles and setbacks, to take stressful events in stride.

Meta-analyses by Avey et al. (2011) showed positive relations between PsyCap and desirable employee attitudes (job satisfaction, organizational commitment, psychological well-being), and between PsyCap and desirable employee behaviors and performance. Negative relations were found between PsyCap and undesirable attitudes such as cynicism, turnover intentions, stress, and anxiety. PsyCap also holds specific relevance for the prediction of extra role, discretionary behaviors, both positive and negative. First, according to broaden and build theory (Fredrickson, 2001), positivity, such as the hope and optimism of PsyCap, has a broadening effect on people’s behavioral repertoires and increases the potential for proactive, extra-role behaviors such as spontaneously helping colleagues and voluntarily attending organizational events. This indicates that lower levels of PsyCap (associated with abusive supervision) should be, in turn, also associated with lower levels of OCBI and OCBO. Second, having a high level of PsyCap is associated with a reduction in CWBs because such optimistic and positive states lead people to behave in a friendly, rather than nasty, way, toward other colleagues. Resources also partly serve to define how the person defines and feels about him or herself, which, in the case of PsyCap, means defining the self as a hopeful and optimistic person. Hence, PsyCap is unlikely to result in behaviors that hurt the organization or that hurt the coworkers. Indeed, these relationships have been supported in prior research in meta-analysis (see Avey et al., 2011).

Importantly, however, a major proposed practical benefit of PsyCap (see Luthans & Youssef, 2004) is that it has developmental potential, meaning people (or supportive environments) can increase this resource. For example, PsyCap has been shown to change over time (Peterson, Luthans, Avolio, Walumbwa, & Zhang, 2011) and variables such as supportive climate are positively related to PsyCap (Luthans, Norman, Avolio, & Avey, 2008). Conversely, this also implies that stressors may affect PsyCap in detrimental ways. Indeed, recent research shows that stressful work environments (L. Liu, Chang, Fu, Wang, & Wang, 2012) and employment uncertainty (Epitropaki, 2013) are associated with reductions in PsyCap. In contrast, PsyCap has been found to mediate relations between positive leadership behaviors and beneficial outcomes (e.g., Gooty, Gavin, Johnson, Frazier, & Snow, 2009). Generally, then, we propose a negative relationship between abusive supervision which is a persistent, sustained stressor and PsyCap a resource with developmental potential. Therefore, we tested the following five hypotheses:

Hypothesis 1: Abusive supervision is negatively related to PsyCap.

Hypothesis 2: PsyCap mediates the negative relation of abusive supervision and OCB1.

Hypothesis 3: PsyCap mediates the negative relation of abusive supervision and OCBO.

Hypothesis 4: PsyCap mediates the positive relation of abusive supervision and CWBI.

Hypothesis 5: PsyCap mediates the positive relation of abusive supervision and CWBO.

It may be important to note that, with regard to our model, there may be similarities between arguing for moderation versus for mediation of PsyCap (see Li, Wang, Yang, & Liu, 2016). PsyCap being a mediator between abusive supervision and its outcomes in no way implies that abusive supervision is the only determinant of PsyCap. Indeed, individuals may draw resources from a variety of sources and it is likely that parts of PsyCap that are not associated with abusive supervision would allow individuals to cope better with abusive behavior (which would imply a moderator; see Cunningham, DiRenzo, & Mawritz, 2013). We consider this possibility exploratively and discuss the issue of how to approach this in more detail below. Hence, we formulated an additional research question:

Research Question 1: Does the element of PsyCap that is not associated with abusive supervision per se moderate the link between abusive supervision and OCBs and CWBs?

Method

Design

Social desirability and common source biases might arise from measurement of independent and dependent variables from the same source, both of which can lead to invalid or inflated results. We sought to remedy such concerns by using a multisource design. Furthermore, it may be argued that CWB represents a type of behavior employees hide from their supervisor, implying colleagues may be a more reasonable source for CWB-measurements. Focal individuals, faculty members, completed questionnaires of abusive supervision (about their supervisor) and their own PsyCap at Time 1. Then, at Time 2, data for focal individuals’ CWBs were collected by asking a colleague of these individuals to rate the relevant items, while data for OCBs were collected from focal individuals’ supervisors. Focal participants, at Time 1, also received a cover letter before agreeing to participate in the study, which explained the procedure for the whole study, including that they would later be rated by their immediate supervisor and by a peer.

There are 183 universities operating in Pakistan and 37,428 academic staff members are working in these universities. Out of 183 universities, 26 institutions were randomly selected. All selected universities have websites on which information about all academic staff is available, which we utilized to prepare department-wise lists of academic staff. Only departments with more than 15 academic staff members were shortlisted for data collection. The permanent faculty members who had at least 2 years of experience in the department were shortlisted for data collection. Using the above criteria, 600 academic staff members were shortlisted. Since the respondents were academic staff, they all had adequate English language skills; therefore, the questionnaires were administered in English. A cover letter was attached with each questionnaire stating the objective of the research and ensuring the confidentiality of the responses. Each questionnaire was allocated a unique code for identification of the focal faculty member. Data were collected through personal visits of the researcher to these academic institutions. A time period of 1 week was given to the faculty members for completion of the questionnaire. After 1 week, the questionnaires were collected.

Out of 600 distributed questionnaires distributed, 523 were returned (87.17%). Out of 523 questionnaires, 38 incomplete questionnaires were excluded. In the next step, questionnaires for OCBO and OCBI of the 485 faculty members were distributed to their supervisors who were the Dean, Chairman, and Heads of Academic Departments and were personally requested by the researcher to participate in data collection. Indeed, 20 supervisors completed a maximum of 15 questionnaires and 6 supervisors completed a maximum of 20 questionnaires to measure OCBO and OCBI of faculty members. For CWBO and CWBI, questionnaires were distributed to their peers. These questionnaires were coded with the same code that had been allocated to respondents at Time 1 of the data collection process. As peer-raters, only those employees who had at least 5 years of experience working in the department were selected to complete the questionnaire. The peers were randomly selected from a list of all coworkers serving in same department. The maximum number of faculty members in a department was 30. Out of 485 sets of questionnaires that had been distributed, 427 (sets or parts of sets of) questionnaires were returned to the researcher. Out of those, 19 were excluded due to missing one part of the complete set (supervisor rating or peer rating). Thus, the overall response rate for complete sets of focal individuals, supervisors, and peers was 408 out of 600, 68%.

Participants

Participants were 408 faculty members from universities in Pakistan (35.0% female; 1.2% did not indicate gender). Participants’ age was measured with five categories: 20 to 25 years (11.0%), 26 to 35 years (63.5%), 36 to 45 years (19.1%), 46 to 55 years (3.9%), and 56 years and older (0.5%), and there were eight nonrespondents (2.0%). We also recorded participants’ work experience with four categories: 1 to 3 years (21.6%), 4 to 6 years (28.7%), 7 to 9 years (17.4%), or 10 years and older (9.1%), but many participants did not complete this question (28.2%).

Measures

Abusive Supervision

We used Tepper’s (2000) instrument consisting of 15 items such as “My supervisor expresses anger at me when he/she is mad for another reason.” Participants responded to these items (M = 1.69, SD = 0.58; α = .90) on the same scale as used in the original questionnaire, where 1 meant “I cannot remember him/her ever using this behavior with me,” 2 meant “He/she very seldom uses this behavior with me,” 3 meant “He/she occasionally uses this behavior with me,” 4 meant “He/she uses this behavior with me moderately often,” and 5 meant “He/she uses this behavior very often with me.”

Psychological Capital

We used Luthans et al.’s (2007) instrument consisting of 24 items (M = 2.26, SD = 0.42; α = .91). Example items are as follows: “If I should find myself in a jam at work, I could think of many ways to get out of it,” “When I have a setback at work, I have trouble recovering from it” (reversed item), “I always look on the bright side of things regarding my job,” and “I feel confident presenting information to a group of colleagues.” We used the same response scale as used in the original version of the questionnaire, from 1 to 6 (1 = strongly disagree, 2 = disagree, 3 = somewhat disagree, 4 = somewhat agree, 5 = agree, 6 = strongly agree). The items were averaged to form the composite scale.

Organizational Citizenship Behaviors

We used Lee and Allen’s (2002) 16-item OCB questionnaire. Participants’ supervisors completed the items on a scale from 1 (never) to 5 (always). Eight items such as “Helps others who have been absent” measured OCBI (M = 3.44, SD = 0.65; α = .82) and eight items such as “Attends functions that are not required but that help the organizational image” measured OCBO (M = 3.44, SD = 0.62; α = .79).

Counterproductive Work Behaviors

We used the 10-item scale by Spector, Bauer, and Fox (2010) to assess the CWBs. For each participant, a colleague responded to these items, on a scale ranging from 1 (never) to 5 (every day). Five items such as “Insulted or made fun of someone at work” were used to measure CWBI (M = 2.58, SD = 0.58; α = .53) and five items such as “Stayed home from work and said they were sick when they weren’t,” were used to measure CWBO (M = 2.59, SD = 0.60; α = .57). The Cronbach’s alphas are relatively low, but item-rest correlations suggested that there were no particularly poor-performing items.

Results

Measurement Model Analysis and Common Method Variance

We examined our six-factor (abusive supervision, PsyCap, OCBI, OCBO, CWBI, and CWBO) measurement model in a confirmatory factor analysis (CFA) and constructed several comparison models. Scale items were used as indicators of the latent variables and we allowed the latent variables to correlate. The hypothesized six-factor model fit the data well, χ² = 2115.66, df = 2,000, p = .036, comparative fit index (CFI) = .984, incremental fit index (IFI) = .984, Tucker–Lewis index (TLI) = .983, root mean score error of approximation (RMSEA) = .012. The hypothesized model also compared favorably to a three-factor model in which the variables were combined according to the source of the data (one factor for abusive supervision and PsyCap, a second factor for OCBO and OCBI, and a third factor for CWBI and CWBO), χ² = 2948.93, df = 2,013, p < .001, CFI = .872, IFI = .873, TLI = .867, RMSEA = .034, and, Δχ² = 833.27, Δdf = 13, p < .001.

We tested three other comparison models. In the first, four factors were created whereby the first factor was abusive supervision, the second was PsyCap, the third factor combined OCBI and OCBO, and the fourth combined CWBI and CWBO. The fit of this model, χ² = 2126.32, df = 2,009, p = .034, CFI = .984, IFI = .984, TLI = .983, RMSEA = .012, was not significantly worse than that of the hypothesized model, Δχ² = 10.66, Δdf = 9, p = .300, which suggests that the OCBs and the CWBs might also be considered to constitute two factors, rather than four. To examine whether this was more pronounced for the OCBs than for the CWBs (or vice versa), we tested two five-factor models wherein either the OCBs or the CWBs represented one factor, while the other represented two factors. A model distinguishing the two forms of OCB, but not distinguishing the two forms of CWB, did not provide a better or worse fit than the hypothesized six-factor model, χ² = 2116.82, df = 2,005, p = .041, CFI = .985, IFI = .985, TLI = .984, RMSEA = .012, and Δχ² = 1.16, Δdf = 5, p = .950. A model distinguishing the two forms of CWB, but not distinguishing the two forms of OCB showed only slightly, but not significantly, worse fit than the hypothesized model, χ² = 2125.18, df = 2,005, p = .031, CFI = .984, IFI = .984, TLI = .983, RMSEA = .012, and Δχ² = 9.52, Δdf = 5, p = .090. All in all, the models that distinguish between the two forms of OCB and the two forms of CWB exhibited no significantly worse (but also no better) fit than the models not making such distinctions. Given the theoretical distinction between the variables, we continued our analyses using the four separate variables.

We examined whether common method variance (CMV) might be an issue in our data (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). First, the results of a principal components analysis that was set to extract one factor indicated that one factor explained 22.94% of the variance. Second, the hypothesized six-factor model compared favorably with a model in which all items loaded onto one factor, χ² = 3970.35, df = 2,015, p < .001, CFI = .732, IFI = .734, TLI = .723, RMSEA = .049, and Δχ² = 1854.69, Δdf = 15, p < .001.

Third, we conducted another CFA with a latent CMV factor added to our hypothesized six-factor model, χ² = 2002.85, df = 1,936, p = .142, CFI = .991, IFI = .991, TLI = .990, RMSEA = .009, which fit the data better than the hypothesized model, Δχ² = 112.81, Δdf = 64, p < .001. However, a comparison of the standardized regression coefficients for the hypothesized model versus those for the latent CMV factor model indicated that the difference between the coefficients in only one item was larger than 0.200. These analyses suggest that CMV was not a major concern.

Analytic Strategy and Hypothesis Tests

The 408 academics were supervised by 26 supervisors. While this suggests, it may be important to examine, whether multilevel analysis is justified (e.g., separating leader-level effects and individual-level effects), abusive supervision is considered an individual-level construct (Tepper, 2000). An analysis of the variance in abusive supervision that could be attributed to the leader level showed the intraclass correlation to be 0.006%, and, F(25, 382) = 1.10, p = .342, indicating abusive supervision should not be aggregated.

Individual cases are nested, nevertheless. We control for this with dummy variables as doing so enables dealing with a potential limitation of the study design, which is that we were unable to collect demographic data from supervisors. An analysis that includes dummy variables for the supervisors provides a conservative test in terms of controlling for supervisor characteristics, as any variance associated with constructs that vary at the supervisor level is systematically partialed out. We did this instead of a random intercept because we had no other way to control for supervisor characteristics, yet this also captures all variance associated with the supervisor level and a random intercept would thus not add anything. Given our sample size, the reduction in statistical power due to these dummy variables was assumed not to be dramatic. We also tested these same models using a random intercept and the results were the same up to three decimals of the coefficients.

Table 1 presents descriptive statistics of the theoretical variables and the zero-order correlations and Table 2 provides an overview of different sets of hypotheses tests. We conducted analyses without control variables, with gender and age of the employees, and, finally, with gender and age as control variables and while including the dummy variables for the supervisors (as age was assessed with five categories, four dummy variables were modelled). We examined the different relations involved in the hypothesized mediation and we statistically tested for mediation by employing the Monte Carlo (bootstrapping) technique implemented by Preacher and Hayes (2004), using 5,000 resamples and 95% confidence intervals, and examining whether confidence intervals included zero. Regardless of the type of control variables included, the results are highly consistent (see Table 2). To illustrate, we explain the results of the analyses that were conducted without control variables.

Table 1.

Descriptive Statistics and Zero-Order Correlations Between the Theoretical Study Variables.

	M	SD	1	2	3	4	5	6
1. Abusive supervision	1.69	0.58	−.90
2. Psychological capital	2.26	0.42	−.62	−.91
3. OCBI	3.44	0.65	−.39	−.56	−.82
4. OCBO	3.44	0.62	−.37	−.54	−.82	−.79
5. CWBI	2.58	0.58	−.21	−.43	−.12	−.10	−.53
6. CWBO	2.59	0.60	−.26	−.50	−.21	−.20	−.47	−.57

Note. OCBI = individual-directed organizational citizenship behavior; OCBO = organization-directed organizational citizenship behavior; CWBI = individual-directed counterproductive work behavior; CWBO = organization-directed counterproductive work behavior.Cronbach’s alphas presented in bold.

rs ≥ .20, ps < .001; rs ≥ .10, ps < .05.

Table 2.

Total Effects, Direct Effects, Indirect Effects, and Tests of Mediation Using Bootstrapped Monte Carlo CIs on All Dependent Variables and With Different Sets of Control Variables.

	No control variables				Age and gender				Age, gender, leader-level dummy variables
	OCBI	OCBO	CWBI	CWBO	OCBI	OCBO	CWBI	CWBO	OCBI	OCBO	CWBI	CWBO
X → Y	−.43 (.05)	−.40 (.05)	−.20 (.05)	−.27 (.05)	−.45 (.05)	−.40 (.05)	−.20 (.05)	−.26 (.05)	−.45 (.05)	−.40 (.05)	−.18 (.05)	−.26 (.05)
X → M	−.44 (.03)	−.44 (.03)	−.44 (.03)	−.44 (.03)	−.44 (.03)	−.44 (.03)	−.44 (.03)	−.44 (.03)	−.43 (.03)	−.43 (.03)	−.43 (.03)	−.43 (.03)
M(X) → Y	−.80 (.08)	.74 (.08)	−.68 (.08)	−.78 (.08)	−.79 (.08)	.73 (.08)	−.69 (.08)	−.79 (.08)	−.82 (.09)	−.74 (.08)	−.70 (.08)	−.78 (.09)
X(M) → Y	−.08 (.06)	−.07 (.06)	−.10 (.06)	−.08 (.06)	−.10 (.06)	−.08 (.06)	−.11 (.06)	−.09 (.06)	−.10 (.06)	−.09 (.06)	−.12 (.06)	−.08 (.06)
Indirect	−.35 (.04)	−.33 (.04)	−.30 (.04)	−.34 (.04)	−.35 (.04)	−.32 (.04)	−.31 (.04)	−.35 (.04)	−.35 (.04)	−.32 (.04)	.30 (.04)	.33 (.04)
95% CI	[–.44, –.27]	[–.41, –.25]	[.23, .39]	[.27, .43]	[–.44, –.27]	[–.41, –.24]	[.23, .39]	[.27, .44]	[–.44, –.27]	[–.40, –.24]	[.22, .39]	[.25, .42]

Note. OCBI = individual-directed organizational citizenship behavior; OCBO = organization-directed organizational citizenship behavior; CWBI = individual-directed counterproductive work behavior; CWBO = organization-directed counterproductive work behavior; CI = confidence interval. Standard errors of the estimated coefficients are included in brackets. Coefficients in bold are significant at p < .001 at the least. None of the other coefficients were significant (ps < .05). M(X) → indicates the effect of PsyCap on the dependent variable while abusive supervision is included in the analysis; X(M) →Y indicates the effect of abusive supervision on the dependent variable while PsyCap is included in the analysis.

PsyCap

Supporting Hypothesis 1, abusive supervision was significantly negatively related to PsyCap. Β = −0.44, SE_Β = 0.03, β = −.62, t(406) = −15.87, p < .001.

OCBI

Abusive supervision negatively related to OCBI, Β = −0.43, SE_Β = 0.05, β = −.39, t(406) = −8.49, p < .001. When including PsyCap as an additional predictor, PsyCap positively related to OCBI, Β = 0.80, SE_Β = 0.08, β = .51, t(405) = 9.76, p < .001, and the link between abusive supervision and OCBI was no longer significant, Β = −0.08, SE_Β = 0.06, β = −.07, t(405) = −1.37, p = .173. Bootstrapping analysis indicated support for Hypothesis 2, that PsyCap mediates the relation between abusive supervision and OCBI: the indirect effect was Β = −0.35, SE_Β = 0.04; the confidence interval did not include zero [−.44, −.27].

OCBO

Abusive supervision negatively related to OCBO, Β = −0.40, SE_Β = 0.05, β = −.37, t(406) = −8.06, p < .001. When including PsyCap as an additional predictor, PsyCap positively related to OCBO, Β = 0.75, SE_Β = 0.08, β = .50, t(405) = 9.38, p < .001, and the link between abusive supervision and OCBO was no longer significant, Β = −0.07, SE_Β = 0.06, β = −.08, t(405) = −1.18, p = .241. Bootstrapping indicated support for Hypothesis 3, that PsyCap mediates the relation between abusive supervision and OCBO: the indirect effect was Β = −0.33, SE_Β = 0.04; the confidence interval did not include zero [−.41, −.25].

CWBI

Abusive supervision positively related to CWBI, Β = 0.21, SE_Β = 0.05, β = .21, t(406) = 4.26, p < .001. When including PsyCap as an additional predictor, PsyCap negatively related to CWBI, Β = −0.68, SE_Β = 0.08, β = −.49, t(405) = −8.64, p < .001, and the link between abusive supervision and CWBI was no longer significant, Β = −0.10, SE_Β = 0.06, β = −.10, t(405) = −1.71, p = .089. Bootstrapping analysis indicated support for Hypothesis 4, that PsyCap mediates the relation between abusive supervision and CWBI: the indirect effect was, Β = 0.30, SE_Β = 0.04; the confidence interval did not include zero [.23, .39].

CWBO

Abusive supervision positively related to CWBO, Β = 0.27, SE_Β = 0.05, β = .26, t(406) = 5.43, p < .001. When including PsyCap as an additional predictor, PsyCap negatively related to CWBO, Β = −0.78, SE_Β = 0.08, β = −.54, t(405) = −9.93, p < .001, and the link between abusive supervision and CWBO was no longer significant, Β = −0.08, SE_Β = 0.06, β = −.08, t(405) = −1.39, p = .165. Bootstrapping indicated support for Hypothesis 5, that PsyCap mediates the relation between abusive supervision and CWBO: the indirect effect was Β = 0.34, SE_Β = 0.04; the confidence interval did not include zero [.27, .43].

Overall, abusive supervision showed significant direct relationships with all four outcomes variables (see Table 2). It may be further noted that the links to OCBs were consistently stronger than those for CWBs, although all were significant. Nevertheless, all four relationships became nonsignificant when including the mediator in the model, suggesting full mediation for all outcomes.

Auxiliary: Explorative Results

Although we explicitly hypothesized PsyCap as being a mediator, our argumentation based on the resource perspective also could be interpreted as implying that PsyCap might be a moderator (see Cunningham et al., 2013). First, technically that is, mediation implies also that the absence of the mediator would prevent the effect of the independent variable on the outcome variable from occurring. Second, individuals’ PsyCap is influenced by many sources of which abusive supervision is just one. On a theoretical level, it is possible that individuals might be able to deal better with abusive supervision if they accrue resources from other sources. The same variable being a mediator and a moderator is problematic; however, since in the theoretical case of full mediation, any effect of the independent variable on the dependent variable is explained by the mediator—including any moderated effect. This is only in the theoretical case, however.

Although one would ideally have an additional variable to model mediator and moderator separately, there might be a statistical way to get an initial indication of whether moderation is also present, without jeopardizing any conclusions about mediation. That is, both mediation and moderation may be true at the same time but apply to different elements of PsyCap. What would be required to explore this possibility would be a version of PsyCap that is cleaned of the variance explained by abusive supervision. To that end, we conducted a regression analysis of PsyCap on abusive supervision and retained the residual variance in PsyCap as a new variable. Hence, this new variable is not associated at all with abusive supervision. The standardized version of this variable, the standardized version of abusive supervision, and the interaction between the two, were then entered in a moderated multiple regression analysis as predictors of the four dependent variables.

The interactions were significant for OCBI, Β = 0.13, SE = 0.03, β = .19, t(404) = 4.34, p < .001, and for OCBO, Β = 0.13, SE = 0.03, β = .20, t(404) = 4.47, p < .001, but not for CWBO, Β = 0.01, SE = 0.03, β = .01, t(404) = 0.30, p = .766 and not for CWBI, Β = −0.02, SE = 0.03, β = −.03, t(404) = −0.66, p = .511. Figure 2 provides the graphical interaction for OCBI, and it may be noted that almost the identical pattern holds for OCBO.

Figure 2.

The element in PsyCap that is not affected by abusive supervision as a moderator of the link between abusive supervision and OCBI.

Discussion

The purpose of this research was to examine a resource perspective on the extra-role behaviors associated with abusive supervision. To that end, we tested whether the relations of abusive supervision, on the one hand, with OCBI, OCBO, CWBI, and CWBO, on the other hand, were mediated by subordinates’ PsyCap—a composite individual resource characteristic consisting of the combination of hope, optimism, resilience, and self-efficacy. The findings from our study, comprising more than 400 academic staff members, ratings from their supervisors (on OCBI and OCBO) and ratings from their peers (on CWBI and CWBO), supported these hypotheses.

From a resource perspective (Hobfoll, 1989), we argued that the lack of control and the need to emotionally cope with abusive supervisory behavior would result in reduced OCBI and OCBO and would result in increased CWBI and CWBO. The necessity to cope with stressors in the environment, in this case abusive supervision, prompts individuals to minimize the loss of resources. This can be done by engaging less in those behaviors that require extra effort and that also cost resources, such as OCBI and OCBO. Similarly, venting in a negative way about the organization or actually trying to cope with resource depletion by taking longer breaks and so forth (CWBO) may also be a result. Furthermore, acting out—not necessarily in retaliation—but out of frustration and anger, toward one’s colleagues (CWBI) is another likely consequence of stress and resource loss. The finding that PsyCap mediates the links between abusive supervision and these beneficial and detrimental extra-role behaviors provides support for this reasoning and, more generally, supports a resource perspective on abusive supervision.

Theoretical Implications

One contribution of this research is the testing of an additional process underlying the outcomes of abusive supervision. Although it should be acknowledged that we are not the first to bring up this resource explanation (Harris et al., 2007; Kacmar et al., 2013), and that research has modeled other related variables such as strain and burnout (e.g., Carlson et al., 2012), we believe specifically examining the state-like resource variable of PsyCap adds to this literature because it suggests something may be done about this detrimental process by intervening. While we explicitly sought to test that PsyCap mediates the outcomes of abusive supervision, our logic also allows for a moderation hypothesis (see Cunningham et al., 2013; Li et al., 2016). At first glance, one may be skeptical about the same variable being treated as a mediator and as a moderator, but a broader resource perspective would indicate that both are reasonable as not all resources are associated with supervision. Individuals accrue resources from other sources as well, and those individuals who do so might be better able to cope with abusive supervision. That is, mediation implies that the absence of the process variable prevents the outcome from occurring. This also can be tested by removing the process and observing that the outcome does not occur even in the presence of the independent variable. Indeed, Cunningham et al. (2013) explicitly noted that PsyCap was a moderator of the outcomes of abusive supervision and our explorative analyses are in most cases aligned with their proposal.

It should nevertheless be acknowledged that there are potential alternative explanations. For instance, people may experience resource loss as an injustice (cf., Tepper, 2000), which then further brings about the consequences of abusive supervision. Moreover, a reciprocity explanation seems to us better suited as an underlying mechanism for certain, even more aggressive, behaviors that we did not examine. For instance, such behaviors as theft, which harm the bottom line of an organization, may not so straightforwardly be linked to a resource perspective. These behaviors seem much more retaliatory than they are compensatory. It is interesting to note, in this regard, that PsyCap, in our explorative analysis, did not moderate the links between abusive supervision and the CWBs. This could be taken to imply tentative indication that other processes than resources are more important in these more retaliatory behaviors. Nevertheless, the current research suggests that a broader model of the abusive supervisory process may be fruitful and should, perhaps, include different pathways in the occurrence of different types of outcomes. As such, this research contributes to a deeper understanding of the mechanisms associated with abusive supervision.

This study found that abusive supervision is positively related to CWBI and CWBO, and that it is also negatively related to OCBI and OCBO. This wholly aligns with the literature and corroborates previous findings using a strong methodology in which the dependent variables were gathered in a time-lagged design from nonself-report sources. The current study adds to the literature on the association between abusive supervision and its outcomes. Also, the current findings add to the literature on abusive supervision by examining this phenomenon in a large and non-Western sample; doing so adds to the generalizability of the theoretical model and responds to a call for studies on abusive supervision from different cultures (see Kernan, Watson, Chen, & Kim, 2011; Mackey, Frieder, Brees, & Martinko, 2017). This study also adds to knowledge about the association between abusive supervision and CWBI and CWBO. Quite a few studies support a link between abusive supervision and different forms of CWB (e.g., Mitchell & Ambrose, 2007) but most of these studies have utilized self-reported CWB, deviance, or resistance. Only one or two exceptions in which supervisor ratings of deviance at the group level (Mawritz, Mayer, Hoobler, Wayne, & Marinova, 2012) or peer ratings (Harvey, Harris, Gillis, & Martinko, 2014) have been used for CWB. The use of peer ratings for CWBI and CWBO as in the current research may thus provide strong support for the association between abusive supervision and CWBI and CWBO.

Finally, this research contributes to the literature on PsyCap. The construct of PsyCap implies that individuals can develop this resource. Indeed, research has tested interventions and human resource development practices that effectively increase individuals’ PsyCap (Luthans, Avey, Avolio, Norman, & Combs, 2006; Luthans, Avey, Avolio, & Peterson, 2010; Luthans, Avey, et al., 2008; Luthans, Vogelgesang, & Lester, 2006) and has found that positive leadership behaviors and positive supportive organizational climates may increase it (Gooty et al., 2009; Luthans, Norman, et al., 2008). The other implication is that contextual influences may decrease PsyCap, but research on this particular implication is limited (see L. Liu et al., 2012). The current research contributes, accordingly, to the validity of the notion that PsyCap is a resource that can be changed by outside influences. More specifically, however, keeping in mind that our research design does not allow to make claims about causality, it also shows that individuals’ PsyCap may be susceptible to negative influences such as abusive supervision.

Practical Implications

OCBI and OCBO comprise some of the extra effort that individuals might exert at work, significantly benefitting overall organizational functioning. In contrast, CWBI and CWBO can severely harm the organization’s culture and its bottom line and can affect others in the organization in negative ways (Dunlop & Lee, 2004). The current research underscores the severity of the problems associated with abusive supervision in relation to these extra-role behaviors. The first and foremost implication is, thus, that organizations need to detect and minimize the occurrence of abusive supervisory behavior.

Providing insight into the process of resources that may partly underlie the outcomes of abusive supervision may also suggest practical implications. That is, if a resource perspective on this matter is considered, then it also follows that individuals might be able to compensate for these in different, more constructive ways. While the most important factor is perhaps to attempt to prevent abusive behavior in the first place, this might not always, and for everyone, be preventable. Individuals and organization might be advised to attempt to develop, and support development of, resources such as PsyCap through other means, which may allow employees to deal better with an abusive supervisor. That is, PsyCap training may be able to enhance individuals’ PsyCap (e.g., Luthans, Avey, et al., 2008), which could, in turn, buffer the effect of variables such as abusive supervision. Since abusive supervision is an individual-level perception, it may be difficult to detect objectively whether someone is experiencing abusive supervision. Training employees to identify issues themselves, and offering ways of increasing employee PsyCap from other sources such as training might allow individuals who perceive their supervisor as abusive, to deal to some extent with the consequences. Our explorative moderation analyses provide some support for the notion that variance in PsyCap not associated with abusive supervision could mitigate its outcomes. Hence, individuals might actively seek for other sources of PsyCap, or training and workshops might be used to help individuals in developing PsyCap resources (see Luthans, Avey, et al., 2008), which could, in turn, help individuals to buffer against the detrimental impact of abusive supervision.

PsyCap is a predictor of a range of beneficial affective and performance outcomes (Avey et al., 2011). The current research focused on a set of behaviors that have far-reaching consequences for organizational functioning. Nevertheless, the finding that the relation between abusive supervision and these outcomes may be partly driven by PsyCap suggests that the effects of abusive supervision may likewise be extremely far reaching. The results therefore underscore the importance of combating this negative abusive supervision in the work place.

Strengths and Limitations

While this study has several strengths, there are certainly also several limitations. One limitation may be that abusive supervision and PsyCap were rated by the same person. This is a limitation that may be impossible to overcome. Abusive supervision is considered an individual perception and should, accordingly, be rated by the subordinate. The same holds true for people’s PsyCap, which refers to their internal sense of their ability to do something successfully at work. Separating the measurements across time would not have dealt with this, as the source of ratings would still be the same. In this sense, it may be important to note that (although a posteriori) the CFAs indicated that CMV was not a major concern. However, this aspect of the design means we are unable to provide causal claims and the conclusion of causal mediation should therefore be made with caution.

This limitation may be partly offset by the fact that other sources of data were used for the dependent variables. We used peer-ratings for CWBI and CWBO, and supervisors for OCBI and OCBO. The usage of such sources can be considered an important strength of this research and yields confidence that the relations between the independent and dependent variables are solid. In any case, they do not seem to be based on common source bias. However, this strength of the current research does not deal with issues of causality, which is still a matter on which we cannot draw conclusions. On the one hand, it may seem that PsyCap is rather an antecedent of performance outcomes than an outcome of performance, which implies that including PsyCap lends some credence to the directionality of our hypotheses. In support of this, Luthans, Avey, et al. (2008) studied the relationship between PsyCap and performance outcomes where the variables were separated in time, which at least lends some support for the temporal order of the occurring variables. On the other hand, it is certainly possible that dysfunctional employee behaviors (or low levels of resources) give rise to abusive supervisory behaviors. However, while this could explain the relationships involving OCBI and OCBO, this is a less plausible alternative explanation for the CWBI and CWBO relations That is, these are particularly behaviors that supervisors might not be able to observe very clearly in that employees will hide these behaviors from supervisors. Moreover, when we consider the chain of variables together, reverse causality becomes less plausible for the mediation chain. The notion that performance of employees affects supervisory behaviors through the perceptions of employees that their own resources are low, is far less logical from a theoretical point of view. While our model logically builds on theory, it is nevertheless not possible to provide causal claims without establishing both temporal and experimental evidence. Finally, while we did not find evidence for common method bias, we note that the correlation between abusive supervision and PsyCap was sizable.

Another limitation of the study is the data collection from academic staff working in universities. The academic organizational structures have unique cultures (with shared governance, faculty who have significant autonomy) as compared with business organizations, which yield findings that could potentially be unique and not broadly generalizable. While this element may limit generalizability, it is also interesting to consider it as a salutary element of this research. That is, even in a context in which there might be a relatively higher degree of autonomy, abusive supervision still has a noticeable, detrimental association with important outcomes.

With regard to the peer-ratings of CWBs, it could be the cases that our method of randomly selecting peer-raters introduced a certain level of measurement inaccuracy in the form of error. That is, while these peers were selected randomly, with the criterion that they should work at the department for more than 5 years, we did not control for how well they knew the focal participant or other indicators of proximity. Another choice could have been to let focal participants themselves choose peers of a certain level of proximity, but allowing them to do so would, in contrast, rather introduce bias. Focal participants are likely to select peers whom they think will provide a positive image of them, yielding positively biased ratings for all participants. Hence, we suggest our choice for selection of the peers by the researchers was the reasonable choice.

Finally, Fox, Spector, Goh, and Bruursema (2007) investigated the convergence between self-ratings and coworker ratings of CWBs. They found little convergence for CWBO, but at the same time they also found that stressor variables predicted both self-ratings of CWBO and peer ratings of CWBO. This implies that both sources could be capturing important elements of this behavior, but that it is unclear which source is addressing which element, or whether one is more accurate than the other. To give an example, only the individual self has completely accurate knowledge about whether they were really sick when they stayed home, so that peers might need to use other sources of information to infer whether individuals were really sick. This implies a limitation of the current research in that the ratings by peers, on CWBO may be inaccurate to some extent, or biased by other variables associated with the peers’ relationship to the participants. Overall, however, the use of different sources, and the examination of several different criterion variables simultaneously, may be seen as salutary characteristics of this study.

Future Research

Taking into accounts the above limitations, the results suggest that PsyCap may play a mediating role between abusive supervision and employee extra-role behaviors. In doing so, this research opens up possibilities for examining a resource perspective on abusive supervision more deeply. The definition of abusive supervision implies that it is not a one-time behavior, but an enduring pattern. An important avenue for future research is, in that sense, to examine the abusive supervision process over time, and in particular from the point at which the relationship starts until the point at which it ends. For instance, when does abuse exactly become enduring and when does it become resource depleting?

Employees may generally cope with abusive supervision by limiting their resource expenditures in other areas, or by acting hostile out of frustration and anger. However, in many occupations, abuse may spillover into different domains as well. In occupations that involve relationships with clients and customers, it could be particularly interesting to examine underlying mechanisms. One might speculate that there are many circumstances in which it does not make sense to show dysfunctional behavior toward clients or customers, as doing so would only harm one’s individual performance. Yet if resources are depleted, customer or client-oriented behaviors that only serve the individual’s own performance may still suffer. More generally, a broader theory of abusive supervision could perhaps include situational moderators for its effect and for the mechanisms that drive such effects.

Given the state-like nature of the concept of PsyCap, it may be reasonable to suspect that people’s PsyCap also varies considerably over time, which is something we did not examine in the current research. Future research is thus needed to explicate the influence of variables such as abusive supervision on PsyCap and other resource variables, and particularly with regard to how incidents of abuse relate to temporary fluctuations in resources. In addition to creating insight into PsyCap, such research may be important to better understand the enduring nature of abusive supervision and the trajectories that might occur in the abusive supervision process over time.

PsyCap has been consistently shown to have a positive relationship with well-being-related variables (e.g., Avey, Luthans, & Jensen, 2009; , Luthans, Smith, & Palmer, 2010). Given that abusive supervision is detrimental to these types of outcomes, an extended model of abusive supervision and employee outcomes could include health and well-being-related variables in addition to the performance variables that we studied. Such a model could provide further support for the importance of developing PsyCap in seeking to mitigate the consequences of abusive supervision and doing so would broaden the theoretical model in important ways.

Conclusion

Our study findings suggest that PsyCap mediates the relationship between abusive supervision and discretionary behaviors such as OCBI, OCBO, and deviant behaviors such as CWBI and CWBO. When employees perceive abusive supervision at work it is associated with decreases in desirable extra-role behaviors (OCBs) and increases in undesirable, deviant work behaviors in the workplace (CWBs). Specifically, in relation to our main aim and research question, our results suggest that decreases in desirable employee behaviors and increases in undesirable employee behaviors can be attributed, at least in part, to the reduced resources (PsyCap) associated with experiencing abusive supervision. Not only does this provide an explanation for the outcomes of abusive supervision and a contribution to the knowledge of antecedents of PsyCap but also tentatively suggest what can be practically done to deal with abusive supervision. Supported by the explorative moderation analyses, the results suggest that buffering against the effects of abusive supervision might be served well by, for example through training, enabling the development of PsyCap.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Melvyn R. W. Hamstra

Author Biographies

Jawwad Ahmad is PhD scholar in University Institute of Management Sciences at Pir Mehr Ali Shah Arid Agriculture University. His research interests are workplace mistreatment, work family conflict, work attitudes, leadership and extra role behaviors.

Muhammad Razzaq Athar is an associate professor in University Institute of Management Sciences at Pir Mehr Ali Shah Arid Agriculture University. His research and teaching interest are employability, performance, training and development.

Rauf I Azam is vice chancellor in University of Education. His research and teaching interest are operational research, consumer behavior, organizational behavior and leadership.

Melvyn R. W. Hamstra is an assistant professor in Organizational Behavior at Maastricht University School of Business and Economics. His research and teaching interests are organizational learning, motivation and performance, and leadership.

Muhammad Hanif is an associate professor in Department of Mathematics & Statistics at Pir Mehr Ali Shah Arid Agriculture University. His area of research interests are statistical inference for stochastic processes, sampling techniques, data analysis and mathematical statistics.

References

Avey

J. B.

(2014). The left side of psychological capital: New evidence on the antecedents of PsyCap. Journal of Leadership & Organizational Studies, 21, 141-149.

Avey

J. B.

Luthans

Jensen

S. M.

(2009). Psychological capital: A positive resource for combating employee stress and turnover. Human Resource Management, 48, 677-693.

Avey

J. B.

Luthans

Smith

R. M.

Palmer

N. F.

(2010). Impact of positive psychological capital on employee well-being over time. Journal of Occupational Health Psychology, 15, 17-28.

Avey

J. B.

Reichard

R. J.

Luthans

Mhatre

K. H.

(2011). Meta-analysis of the impact of positive psychological capital on employee attitudes, behaviors, and performance. Human Resource Development Quarterly, 22, 127-152.

Carlson

Ferguson

Hunter

Whitten

(2012). Abusive supervision and work–family conflict: The path through emotional labor and burnout. Leadership Quarterly, 23, 849-859.

Cunningham

Q. W.

DiRenzo

M. S.

Mawritz

M. B.

(2013, August). Abusive supervision and employee outcomes: The influence of psychological contract violation and psychological capital. Paper presented at the Academy of Management Conference 2013, Orlando, FL.

Duffy

M. K.

Ganster

D. C.

Shaw

J. D.

Johnson

J. L.

Pagon

(2006). The social context of undermining behavior at work. Organizational Behavior and Human Decision Processes, 101, 105-126.

Dunlop

P. D.

Lee

(2004). Workplace deviance, organizational citizenship behavior, and business unit performance: The bad apples do spoil the whole barrel. Journal of Organizational Behavior, 25, 67-80.

Epitropaki

(2013). Employment uncertainty and the role of authentic leadership and positive psychological capital. Academy of Management Proceedings, 2013, 10925. doi:10.5465/AMBPP.2013.10925abstract

10.

Fox

Spector

P. E.

Goh

Bruursema

(2007). Does your coworker know what you’re doing? Convergence of self-and peer-reports of counterproductive work behavior. International Journal of Stress Management, 14, 41-60.

11.

Fox

Spector

P. E.

Miles

(2001). Counterproductive work behavior (CWB) in response to job stressors and organizational justice: Some mediator and moderator tests for autonomy and emotions. Journal of Vocational Behavior, 59, 291-309.

12.

Fredrickson

B. L.

(2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist, 56, 218-226.

13.

Frieder

R. E.

Hochwarter

W. A.

DeOrtentiis

P. S.

(2015). Attenuating the negative effects of abusive supervision: The role of proactive voice behavior and resource management ability. Leadership Quarterly, 26, 821-837.

14.

Gooty

Gavin

Johnson

P. D.

Frazier

M. L.

Snow

D. B.

(2009). In the eyes of the beholder transformational leadership, positive psychological capital, and performance. Journal of Leadership & Organizational Studies, 15, 353-367.

15.

Harris

K. J.

Kacmar

K. M.

Zivnuska

(2007). An investigation of abusive supervision as a predictor of performance and the meaning of work as a moderator of the relationship. Leadership Quarterly, 18, 252-263.

16.

Harvey

Harris

K. J.

Gillis

W. E.

Martinko

M. J.

(2014). Abusive supervision and the entitled employee. Leadership Quarterly, 25, 204-217.

17.

Harvey

Stoner

Hochwarter

Kacmar

(2007). Coping with abusive supervision: The neutralizing effects of ingratiation and positive affect on negative employee outcomes. Leadership Quarterly, 18, 264-280.

18.

Hobfoll

S. E.

(1989). Conservation of resources: A new attempt at conceptualizing stress. American Psychologist, 44, 513-524.

19.

Jian

Kwan

H. K.

Qiu

Liu

Z. Q.

Yim

F. H. K.

(2012). Abusive supervision and frontline employees’ service performance. Service Industries Journal, 32, 683-698.

20.

Kacmar

K. M.

Whitman

M. V.

Harris

K. J.

(2013). The lingering impact of abusive supervision. Journal of Applied Management and Entrepreneurship, 18, 51-71.

21.

Kernan

M. C.

Watson

Chen

F. F.

Kim

T. G.

(2011). How cultural values affect the impact of abusive supervision on worker attitudes. Cross Cultural Management: An International Journal, 18, 464-484.

22.

Lee

Allen

N. J.

(2002). Organizational citizenship behavior and workplace deviance: The role of affect and cognitions. Journal of Applied Psychology, 87, 131-142.

23.

Wang

Yang

L. Q.

Liu

(2016). The crossover of psychological distress from leaders to subordinates in teams: The role of abusive supervision, psychological capital, and team performance. Journal of Occupational Health Psychology, 21, 142-153.

24.

Liu

Chang

Wang

(2012). The mediating role of psychological capital on the association between occupational stress and depressive symptoms among Chinese physicians: A cross-sectional study. BMC Public Health, 12, 219.

25.

Liu

X. Y.

Wang

(2013). Abusive supervision and organizational citizenship behaviour: Is supervisor–subordinate guanxi a mediator? International Journal of Human Resource Management, 24, 1471-1489.

26.

Luthans

Avey

J. B.

Avolio

B. J.

Norman

S. M.

Combs

G. M.

(2006). Psychological capital development: Toward a micro-intervention. Journal of Organizational Behavior, 27, 387-393.

27.

Luthans

Avey

J. B.

Avolio

B. J.

Peterson

S. J.

(2010). The development and resulting performance impact of positive psychological capital. Human Resource Development Quarterly, 21, 41-67.

28.

Luthans

Avey

J. B.

Patera

J. L.

(2008). Experimental analysis of a web-based training intervention to develop positive psychological capital. Academy of Management Learning & Education, 7, 209-221.

29.

Luthans

Avolio

B. J.

Avey

J. B.

Norman

S. M.

(2007). Positive psychological capital: Measurement and relationship with performance and satisfaction. Personnel Psychology, 60, 541-572.

30.

Luthans

Norman

S. M.

Avolio

B. J.

Avey

J. B.

(2008). The mediating role of psychological capital in the supportive organizational climate—employee performance relationship. Journal of Organizational Behavior, 29, 219-238.

31.

Luthans

Vogelgesang

G. R.

Lester

P. B.

(2006). Developing the psychological capital of resiliency. Human Resource Development Review, 5, 25-44.

32.

Luthans

Youssef

C. M.

(2004). Human, social, and now positive psychological capital management: Investing in people for competitive advantage. Organizational Dynamics, 33, 143-160.

33.

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.

34.

Martinko

M. J.

Harvey

Brees

J. R.

Mackey

(2013). A review of abusive supervision research. Journal of Organizational Behavior, 34(Suppl. 1), S120-S137.

35.

Mawritz

M. B.

Mayer

D. M.

Hoobler

J. M.

Wayne

S. J.

Marinova

S. V.

(2012). A trickle-down model of abusive supervision. Personnel Psychology, 65, 325-357.

36.

Michie

(2002). Causes and management of stress at work. Occupational & Environmental Medicine, 59, 67-72.

37.

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.

38.

Nandkeolyar

A. K.

Shaffer

J. A.

Ekkirala

Bagger

(2014). Surviving an abusive supervisor: The joint roles of conscientiousness and coping strategies. Journal of Applied Psychology, 99, 138-150.

39.

Peterson

S. J.

Luthans

Avolio

B. J.

Walumbwa

F. O.

Zhang

(2011). Psychological capital and employee performance: A latent growth modeling approach. Personnel Psychology, 64(2), 427-450.

40.

Podsakoff

P. M.

MacKenzie

S. B.

Lee

J. Y.

Podsakoff

N. P.

(2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879-903.

41.

Preacher

K. J.

Hayes

A. F.

(2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior research methods, instruments, & computers, 36(4), 717-731.

42.

Schat

A. C.

Frone

M. R.

Kelloway

E. K.

(2006). Prevalence of workplace aggression in the U.S. workforce: Findings from a national study. In Kelloway

E. K.

Barling

Hurrell

J. J.

(Eds.), Handbook of workplace violence (pp. 47-89). Thousand Oaks, CA: Sage.

43.

Spector

P. E.

(1998). A control theory of the job stress process. In Cooper

C. L.

(Ed.), Theories of organizational stress (pp. 153-169). Oxford, England: Oxford University Press.

44.

Spector

P. E.

Bauer

J. A.

Fox

(2010). Measurement artifacts in the assessment of counterproductive work behavior and organizational citizenship behavior: Do we know what we think we know? Journal of Applied Psychology, 95, 781-790.

45.

Tepper

B. J.

(2000). Consequences of abusive supervision. Academy of Management Journal, 43, 178-190.

46.

Tepper

B. J.

Henle

C. A.

Lambert

L. S.

Giacalone

R. A.

Duffy

M. K.

(2008). Abusive supervision and subordinates’ organization deviance. Journal of Applied Psychology, 93, 721-732.

47.

Huang

Lam

C. K.

Miao

(2012). Abusive supervision and work behaviors: The mediating role of LMX. Journal of Organizational Behavior, 33, 531-543.

48.

Yagil

Ben-Zur

Tamir

(2011). Do employees cope effectively with abusive supervision at work? An exploratory study. International Journal of Stress Management, 18, 5-23.

49.

Youssef

C. M.

Luthans

(2007). Positive organizational behavior in the workplace: The impact of hope, optimism, and resilience. Journal of Management, 33, 774-800.