Abusive retaliation of low performance in low-quality LMX relationships

Abstract

In the current study, we investigated the occurrence of abusive retaliation in response to low performance. More specifically, we hypothesize an association between low work performance and abusive supervision, as well as leader-member exchange (LMX) as a moderator of the association. Furthermore, we explored a potential link between abusive supervision and social loafing, and whether work-related negative affect would mediate this link. Data were collected from business organizations in Norway. To alleviate potential common method variance, data were collected separately from leaders and followers. In line with our expectations, we present results indicating that low performing followers are more prone to be abused by their supervisor, particularly if they are in a low quality LMX relationship. Conversely, LMX relationships can have a “buffering” effect on the performance/abusive supervision association. Furthermore, our results suggest abused followers reacted with both work-related negative affect and social loafing. The implications of this latter finding are discussed in light of the ongoing conversation regarding the role of emotions in exchange processes.

Keywords

abusive supervision performance LMX work-related negative affect social loafing counter-productive behavior

Introduction

Research on leadership for decades has explored the question of what makes leaders effective in influencing and mobilizing followers (Bass and Riggio, 2006; Ferris et al., 2007). Somewhat surprisingly, however, despite its long history, research has primarily focused on “productive” leadership, and until the past decade paid little attention to its dark side and its adverse effects on followers in terms of job satisfaction, affective commitment, and psychological well-being (Benson and Hogan, 2008). Research has looked at different kinds of destructive leader behavior that hinder the ability to establish and maintain positive interpersonal relationships (Duffy et al., 2002) or behavior that is intended to physically or psychologically harm workers (Schat et al., 2006). The present study limits its scope to the examination of abusive leadership by testing a multivariate model that expands our knowledge of predictors and outcomes of abusive leadership.

Tepper (2000) defines abusive supervision as “subordinates’ perceptions of the extent to which supervisors engage in the sustained display of hostile verbal and nonverbal behaviors, excluding physical contact” (p. 178). Examples of prolonged hostile treatment of subordinates include explosive outbursts, ignoring subordinates, humiliation in front of others, sabotaging, public ridicule, and coercing others (Duffy et al., 2002; Hoobler and Brass, 2006; Mitchell and Ambrose, 2007). Furthermore, a number of studies report greater destructive impacts of abusive supervision on subordinates’ job and life satisfaction, psychological distress, turnover intent, and reduced citizenship behavior than their non-abused colleagues (Tepper, 2000, 2007; Zellars et al., 2002). Previous studies have also examined antecedents of abusive supervision (Aryee et al., 2007; Gabler et al., 2014; Harris et al., 2011; Hoobler and Brass, 2006; Lian et al., 2014; Tepper et al., 2006). According to Tepper et al. (2011), most studies focus on followers’ perception of mistreatment by their employer, and not on the reasons why leaders abuse specific followers. Tepper et al. (2011), who studied predictors of abusive supervision, identified follower performance as a predictor and demonstrated that followers perceived to be low performers may become targets of abusive supervision. This finding seems to fit Olweus’s (1978) concept of “provocative victims,” describing how pupils in schools who were considered annoying and difficult to work with might become targets of victimization. In similar manner, we believe that low performers at work may be perceived as provocative victims by some of their leaders and thus become victims of abusive supervision. However, several situational factors may moderate the association between follower performance and abusive supervision. In the present study, we apply a relational perspective using social exchange theory (Blau, 1964) and expand on Tepper et al. (2011) by developing a theoretical model suggesting that leader member exchange (LMX) relationships can have a “buffering” effect on the performance/abusive supervision association. Our rationale for applying LMX theory is that it allows us to investigate how different relationships that vary in quality influence the extent to which leaders engage in abusive supervision as a result of low follower performance. Specifically, we expect that low-performing followers to a lesser extent will be abused by their leaders in high-quality LMX relationships. In the present article, we therefore examine to what extent leaders with poor LMX relationships may expose their low-performing followers to abusive supervision.

However, in the present study, we not only investigate antecedents to abusive supervision but aim to contribute to the extant literature by exploring uncharted outcomes of abusive supervision. A number of studies have identified different types of consequences of nonphysical supervisory hostility, such as antisocial and deviant behavior, and psychological well-being and performance consequences (Tepper, 2007). Schyns and Schilling (2012), in their meta-analysis, have documented negative associations between destructive forms of leadership and attitudes toward the leader, subordinate well-being, and individual performance and positive correlations with turnover, resistance toward the leader, and counterproductive work behavior (see also Einarsen et al., 2013). Regarding subordinate resistance, numerous studies have suggested that subordinates frequently directly retaliate against their abusive leaders (Hershcovis & Barling, 2010; Tepper and Almeda, 2012). There are studies including direct retaliation, such as confronting the leader, to more indirect retaliation, such as decreased organizational citizenship behavior and reduced job performance (see also Lian et al. (2015) for an overview). In addition, Tepper et al. (2001) found that abused subordinates used destructive resistance tactics (ignoring supervisor’s requests and questioning the supervisor’s authority) and constructive resistance tactics (e.g. being willing to break the cycle of hostility by responding to abusive supervision in a nondestructive manner) more often than did their non-abused counterparts. In accordance with affective events theory (Weiss and Cropanzano, 1996), we believe that job-related factors at work, such as abusive supervision, can generate negative affect (NA), which in turn influences follower behavior. This notion is in line with the stressor-emotion model, which suggests that negative emotions or affect can trigger dysfunctional behavior (Spector and Fox, 2005). Including work-related NA as a potential mediator between abusive supervision and destructive resistance tactics represents an extension of the Tepper et al. (2001) study. In addition, it represents a potential contribution to social exchange theory. Specifically, social exchange theory has been criticized for ignoring the role of emotions (ref.). By exploring the mediating role of work-related affect, we attempt to contribute to social exchange theory by exploring the extent to which negative reciprocity reflects a rational cognitive versus a more emotional process.

A further extension and contribution to the literature would be to investigate a different resistance strategy by followers experiencing hostile behavior from their leader. Follower resistance may consist of various dysfunctional behaviors, such as limiting personal effort and contribution to the firm, withholding information, giving bad advice, or providing misleading information (Fox et al., 2001). In the present study, we use the concept of follower “social loafing,” that is a behavioral withdrawal response to circumstances perceived by the individual—as measured by their leader. Social loafing is a motivational construct according to Liden et al. (2004) which can be defined as “the reduction in motivation and effort when individuals work collectively compared with when they work individually or coactively” (Karau and Williams, 1993, p. 681). Specifically, we expect followers who experience NA, as a response to abusive supervision, to resort to social loafing, which includes reducing their physical, perceptual, and cognitive effort and allowing others to pick up the slack (Murphy et al., 2003). Based on our literature review and our suggested extensions regarding predictors and affective and behavioral outcomes of abusive supervision, we have developed a new research model depicted in Figure 1. The present investigation tests this model using two independent samples, hereafter referred to as study 1 and study 2.

Figure 1.

Predictors and outcomes of abusive supervision.

Theory and hypotheses

According to Tepper et al. (2011), little is known about the factors that predict the occurrence of abusive supervision. Studies on antecedents of abusive supervision have found that organizational injustice and breach of the psychological contract are associated with subordinate reports of supervisory abuse (Aryee et al., 2007; Hoobler and Brass, 2006; Tepper et al., 2006). However, Tepper et al. (2011) focused on factors that could explain why leaders abuse specific followers and found a negative association between follower performance and abusive supervision, consistent with moral exclusion theory (Opotow, 2001). In addition, the victim precipitation literature predicts that abusive supervision is more likely to occur when subordinates do not meet required performance standards and are perceived as annoying and difficult to work with (Olweus, 1978). In the present investigation, we rely on social exchange theory (Blau, 1964) as an overarching framework when developing our hypotheses. As noted by Dulebohn et al. (2012), although LMX theory was rooted in role theory, it has evolved to rely increasingly on social exchange theory. The basic premise of social exchange theory is that individuals, on the basis of the norm of reciprocity (Gouldner, 1960), feel an obligation to reciprocate when they receive beneficial treatment. As noted by Blau (1964, p. 16) “If we feel grateful and obligated to an associate for favors received, we shall seek to reciprocate his kindness by doing things for him.” Furthermore, according to social exchange theory and psychological contract theory (e.g. Rousseau, 1995), when one party to the exchange fails to fulfil promised obligations, the other party is likely to reciprocate by reducing positive attitudes or behavior or by getting even through negative actions (e.g. De Ruiter et al., 2016; Zhao et al., 2007). Building on this, we argue that when the follower fails to achieve expected performance, the leader is likely to reciprocate by getting even through negative actions. Specifically, LMX theory contends that followers should repay their leader for benefits received in a high-quality LMX relationship by performing on behalf of the leader. When subordinates do not perform, leaders may perceive that followers have not fulfilled the promised obligations inherent in the exchange relationship. Leaders may then get frustrated or angry and feel let down by the follower. In such situations, some kind of abusive behavior might be a likely outcome. Accordingly, we expect that followers with low work performance are more likely to be targets of abusive supervision. Hence, we hypothesize:

Hypothesis 1: There is a negative relationship between follower work performance and abusive leadership.

The moderating role of LMX

Even though followers who fail to meet superiors’ performance expectations may experience abusive behavior, the extent of the abuse may depend on the quality of the social exchange relationship between the leader and the follower. Some followers have a close working relationship with their leader (high-quality LMX relationship) based on mutual trust, respect, and obligation, while others have a poor working relationship solely based on a contractual work/pay basis with the same leader. According to Graen and Uhl-Bien (1995, p. 232), the motivations of a follower in a low-quality LMX relationships “are based upon the satisfaction of his/her own self-interests, without consideration of the good of the group.” These latter followers will probably experience less socio-emotional support from their leader, show less trust in their leader, and respond less positively when they experience setbacks. We expect leaders in high-quality working relationships with low-performing followers to perceive these employees as more deserving of fair treatment, and more likely to be delegated interesting tasks, provided access to important resources, and coached to enhance their performance. Specifically, according to the buffering hypothesis (e.g. Bal et al., 2010), lower performing followers may be less likely to become the target of abusive supervision when in a high-quality LMX relationship with their leader. Furthermore, as noted by Aselage and Eisenberger (2003, p. 496), “the norm of reciprocity holds parties more responsible for favorable or unfavorable treatment when such treatment is voluntary as opposed to being accidental or due to factors beyond an individual’s control.” For instance, leaders having a high-quality LMX relationship with followers may perceive low performance as less intentional and rather more attributable to circumstances than personal disposition. Followers in a low-quality LMX relationship, on the other hand, may more likely become targets of abusive supervision as their leader may interpret their lack of performance as yet another signal that the follower does not value the relationship with the leader, or even the job as a whole. In addition, supervisors may perceive these employees as annoying and difficult to work with and target them for victimization, especially if they are perceived as weak and defenseless (Aquino, 2000). Since abusive supervision reduces follower motivation and negatively affects one’s attitude toward the job (Duffy et al., 2002), it is less likely that leaders would risk these negative consequences with high LMX followers. Rather, leaders would likely be motivated to enhance follower performance. However, in a poor LMX relationship, leaders are probably not equally motivated to behave constructively, and, presumably, they perceive they have less to lose through hostility because these followers are unwilling to put much effort into their work anyway. Hence, we suggest the following hypothesis:

Hypothesis 2: The negative relationship between follower work performance and abusive leadership is moderated by the quality of the LMX relationship—the lower the quality of the LMX, the stronger the negative relationship.

Outcomes of abusive supervision

Below, we argue that when followers experience abusive supervision, they respond by means of social loafing, that is, by reducing their physical, perceptual, and cognitive effort such as letting others to pick up the slack (Murphy et al., 2003). Furthermore, we argue that NA is the mechanism through which abusive supervision results in social loafing.

As a response to abusive supervision, social exchange theory suggests that followers can be expected to retaliate for perceived harm or punishment. Blau (1964) even postulated that “retaliation for the pain inflicted on them tends to become an end-in-itself, for the sake of which they are willing to sacrifice other advantages” (p. 227). On this basis, researchers have suggested that abusive supervision should elicit leader directed aggression (Lian et al., 2014). Similarly, Gouldner (1960) argued that targets of harm will react in accordance with what he labeled “return of injuries” (p. 172): negative norms of reciprocity, that is, in sentiments of retaliation where the emphasis is placed not on the return of benefit but on the return of harm. However, social exchange theory has been criticized for conceptualizing the exchange as a rational calculative process (e.g. Korsgaard et al., 2010) in which employees compare their inputs and outputs in relation to some standard of reward and punishment and adjust their actions accordingly. Predictions resulting from social exchange theory have been largely supported (e.g. Dulebohn et al., 2012; Rhoades and Eisenberger, 2002). However, as pointed out by Zhao et al. (2007), a major limitation of social exchange theory is minimizing the role of emotion and assume that reactions are based on rational cognitive and objective judgments. Building on the affective events theory of Weiss and Cropanzano (1996), Zhao et al. (2007) argue that individuals who experience a negative workplace event will not necessarily engage in a rational appraisal of the situation, but rather react with negative emotional affect, which may progress to overwhelm the individual involved. Accordingly, we expect abusive supervision to result in negative work-related affect, which manifests in employee outcomes such as social loafing. In addition, the importance and relevance of the event, in this case, abusive supervision, we expect will determine the intensity of the negative emotional response (Zhao et al., 2007). Furthermore, Vecchio (1995) suggests that emotional responses can trigger strong outbursts and physical attacks such as sabotage, harassment, or ostracism. Specifically, mistreated low performers feeling nervous, irritable, hostile, and upset may engage in interpersonal counterproductive work behaviors directed at the leader. So, we expect that abused low-performance followers experiencing work-related NA may react to their situation perhaps by limiting their personal effort and contribution to the organization through a process of social loafing (Fox et al., 2001; Vecchio, 2007). In the present study, we have chosen the target variable of counterproductive work behavior to be represented by social loafing, where low-performing followers may sacrifice personal beneficial outcomes to retaliate against the organization and/or the leader. Accordingly, we hypothesize:

Hypothesis 3: Abusive supervision is positively related to subordinate social loafing via work-related NA.

Finally, from a deontic justice perspective (e.g. He et al., 2016), followers are motivated to react not only to how they themselves are treated by their leader but also to how their fellow coworkers are treated by their leader. Followers can be expected to be concerned about how others are treated and react negatively when they perceive violations of their moral and ethical standards, even if they themselves are not the victims, as emphasized by Skarlicki and Kulik (2005) and He et al. (2016). Research reviewed by He et al. (2016) provides strong support for the deontic justice perspective and shows that employees who see coworkers treated unfairly react with negative emotions and attempt to retaliate against the instigator (Barclay et al., 2005; O’Reilly et al., 2016; Spencer and Rupp, 2009). Hence, we hypothesize that abusive supervision is also likely to result in social loafing at the group level via work-related affect. Therefore, we hypothesize:

Hypothesis 4: Abusive supervision is positively related to subordinate social loafing via work-related NA at the group level.

The current investigation of predictors and outcomes of abusive supervision is conducted in two studies. The first study consists of an empirical examination of hypotheses 1 and 2, where the purpose is to identify predictors of abusive supervision. The second study examines hypotheses 3 and 4 focusing on outcomes of abusive supervision. By applying two studies, investigating different aspects of abusive supervision, we provide a more comprehensive understanding of the destructive dynamics of abusive supervision in the context of business organizations.

Study 1

Method

Participants and procedures

Data were collected from medium to large Norwegian business organizations within the telecommunications, oil, shipping, and pharmaceutical industries. Questionnaires were distributed electronically, and each leader and all their direct reports were given a unique access link to the questionnaires. Responses to the items were collected electronically. Participants completed the surveys during work hours. The covering letter informed respondents that data collection was conducted solely for academic research purposes, with the goal of better understanding various aspects of the leader–follower relationship. Respondents were assured of the confidentiality of their responses. Respondents were not compensated for their participation in the study.

Responses were obtained from 59 leaders and 493 followers. The response rate was nearly 77%. The leader sample consisted of 56.7% males. The leaders’ mean age was 41.73 years, reporting an average education of 14.8 years. They averaged 4.57 years in their current job in the company and were employed as shift supervisor, team leader, executive director, production supervisor, technical manager, sales director, service center manager, managing director, factory manager, marketing supervisor, and logistics manager. The follower sample consisted of 57.8% males, and their mean age was 38.84 years, with average education of 13.4 years. Their average tenure in their current position in the company was 7.29 years, and the average years working for their supervisor was 3.49 years. They held various positions including production worker, process operator, plant controller, operating technician, engineer, systems engineer, market developer, customer support, designer, production planner, sales consultant, quality controller, laboratory engineer, and high-level employee reporting directly to middle and upper management supervisors.

Instruments

To avoid same source bias, the leaders were asked to rate their followers’ work performance, which was measured using a five-item performance rating scale developed by Liden and Graen (1980) (item stems: Dependability, Planning, Know-how and Judgment, Overall Present Performance, and Expected Future Performance; anchors: 1 = unsatisfactory, 7 = outstanding). Responses to these five items were then averaged to provide a measure of Performance for each subordinate.

Each follower rated LMX, which was measured with the seven-item version of the LMX (Graen and Uhl-Bien, 1995). Sample item: “My supervisor has enough confidence in me that he/she would defend my decisions if I were not present to do so”; anchors: 1 = strongly disagree, 2 = disagree, 3 = undecided, 4 = agree, 5 = strongly agree. LMX was included in the survey to provide an index of employee assessment of his/her working relationship with the leader.

Abusive supervision was measured with the 15-item version of Tepper’s (2000) instrument. Sample items: “My supervisor tells me my thoughts and feelings are stupid”; “My supervisor doesn’t give me credit for jobs requiring a lot of effort.” Respondents indicated their agreement with each item using a five-point scale; 1 = never, 2 = seldom, 3 = occasionally, 4 = often, 5 = always.

Control variables

We controlled for the age and gender of both the leader and the follower which could potentially provide alternative explanations for the relationships outlined in our hypotheses. In addition, because relational demographics may influence abusive supervision (e.g. Tepper et al., 2011), we controlled for gender (similar genders coded “0”; dissimilar genders coded “1”) and age differences (using an absolute difference score). Furthermore, the length of time the follower had worked in the organization and under the same leader could also affect the hypothesized relationships. Accordingly, we controlled for both organizational tenure and dyad tenure.

Results

Confirmatory factor analysis

We conducted a preliminary analysis to test the measurement models for the included variables to determine whether the items reflected the construct they were intended to measure and performed a confirmatory factor analysis (CFA) on a four-factor model. A model with 32 items loading on four latent variables was specified without cross-loadings. CFA revealed acceptable fit (Hair et al., 2010): χ ²[458] = 616.76, p < 0.01; RMSEA = 0.027; CFI = 0.99; NNFI/TLI = 0.98. According to Chin (1998, p. xiii), “Most of the loadings should be at least 0.60, and ideally at 0.70 or above, indicating that each measure is accounting for 50 percent or more of the variance of the underlying LV.” Convergent validity is supported, as all factor loadings were statistically significant with a mean standardized loading of 0.82.

Descriptive statistics

Table 1 provides descriptive statistics and intercorrelations among the study variables. All the scales demonstrated high internal consistency, with composite reliability estimates ranging from 0.87 to 0.97. The composite reliability is similar to Cronbach’s α (Nunnally, 1978) but additionally accounts for the possibility that the items may have different error variances and loadings (Bagozzi and Yi, 1988). According to Bagozzi and Yi (1988), the composite reliability estimate should be higher than 0.60. The coefficient estimates for the multi-item scales are listed on the primary diagonal of the intercorrelation matrix. As expected, follower work performance was significantly and negatively correlated with abusive supervision (r = −0.20, p < 0.01) and significantly and positively correlated with LMX (r = 0.31, p < 0.01); and LMX was significantly and negatively correlated with abusive supervision (r = −0.42, p < 0.01).

Table 1.

Means, standard deviations, reliabilities, and intercorrelations for study 1.

		Mean	SD	1	2	3	4	5	6	7	8	9	10	11	12
1.	Leader age	41.73	7.46
2.	Follower age	38.84	10.66	0.28**
3.	Age difference	9.40	6.70	0.21**	−0.04
4.	Leader gender	0.43	0.50	0.01	−0.10	−.01
5.	Follower gender	0.42	0.50	0.04	−0.07	0.01	0.21**
6.	Gender difference	0.39	0.49	0.08	−0.06	0.03	0.13*	0.12*
7.	Dyad tenure	65.87	48.01	0.09*	0.28**	−0.15**	−0.15**	−0.14*	−0.06
8.	Organizational tenure	62.91	49.43	0.01	0.44**	0.09	−0.09*	−0.03	0.02	0.50**
9.	Follower work performance	5.52	1.04	0.00	0.00	0.03	0.13**	0.07	−0.04	0.05	−0.04	(0.95)
10.	LMX	3.84	0.64	−0.10*	−0.08	−0.09	0.10	0.02	0.03	0.09	−0.03	0.31**	(0.92)
11.	Willingness	6.56	0.99	0.01	−0.04	−0.08	0.06	−0.01	0.02	−0.08	−0.17**	0.21**	0.30**	(0.87)
12.	Abusive supervision	1.23	0.38	0.19**	0.18**	0.07	−0.13*	−0.09	−0.08	0.17**	0.19**	−0.20**	−0.42**	−0.30**	(0.97)

Note: LMX: leader member exchange; SD: standard deviation. n = 493. Composite reliability estimates on primary diagonal.

*p < 0.05; **p < 0.01.

HLM analyses

Because our data are hierarchical (i.e. individuals nested within leaders), we tested our hypotheses 1 and 2 using hierarchical linear modeling (HLM) via SPSS (SPSS statistics 25) mixed. In these analyses, we estimated a fully unconditional model (null model) for the dependent variable (abusive supervision). The results of this null model indicated significant between-supervisor variability in subordinate ratings of abusive supervision (τ₀₀ = 0.02, p < 0.01). Furthermore, we calculated the intraclass correlation coefficients (ICC) (Hofmann et al., 2000) of abusive supervision (ICC = 0.14), which suggested that a significant proportion of the variability in abusive supervision (14%) was attributable to between-group variability, thus underlining the appropriateness of HLM. Table 2 presents the results of the HLM analyses. In step 1, we entered the control variables. In step 2, we entered the independent variable follower performance. Results showed support of hypothesis 1, as follower performance was negatively related to abusive supervision (standardized coefficient = −0.24, p < 0.001). Next, in step 3, results showed that follower performance (standardized coefficient = −0.15, p < 0.001) and LMX (standardized coefficient = −0.27, p < 0.001) related negatively to abusive supervision concurrently. Finally, in step 4, we entered the interaction term (follower performance × LMX). The introduction of the interaction term (standardized coefficient = 0.14, p < 0.001) resulted in increase in the pseudo R ² (ΔR ² = 0.02) and a significant decrease in model deviance (Δχ ² = 10.65, p < 0.01), thus showing that the relationship between follower performance and abusive supervision is moderated by LMX. To investigate the nature of the statistically significant interaction, we plotted low versus high scores on leader LMX (one standard deviation below and above the means using nonstandardized scores; Cohen et al., 2003). To determine the significance of the simple slopes, we used the HLM two-way interaction tool developed by Preacher et al. (2006). The results displayed in Figure 2 demonstrate a negative relationship between follower work performance and abusive supervision only for individuals who perceive a low-quality LMX relationship. Specifically, for followers with low LMX, the relationship between performance and abusive supervision was negative and significant (−0.09, p < 0.001). In contrast, the relationship was not statistically significant for followers with high LMX (0.01, n.s.), suggesting that LMX represents a boundary condition under which performance relates to abusive supervision. Accordingly, we received support for hypothesis 2. To further examine the validity of our results, we performed a supplemental analysis where we excluded all the control variables that were not significantly related to abusive supervision. We did so because control variables can sometimes lead to spurious effects, in particular with respect to nested data (Becker et al., 2016). The results of this supplemental analysis which included only one control variable (dyad tenure) did not differ substantially from the results that included all the control variables.

Table 2.

Results of hierarchical linear modeling analyses for the two-way interaction in study 1.

Variables	Abusive supervision
Variables	Step 1	Step 2	Step 3	Step 4
Intercept	0.87***	0.84***	0.92***	0.92***
Leader age	0.09	0.11*	0.09	0.08
Follower age	0.06	0.07	0.05	0.05
Age difference^a	0.06	0.05	0.04	0.04
Leader gender^b	−0.04	−0.02	−0.01	−0.02
Follower gender^b	−0.04	−0.02	−0.02	−0.03
Gender difference^c	−0.05	−0.07	−0.07	−0.06
Dyad tenure	0.07	0.08	0.10*	0.10*
Follower organizational tenure	0.07	0.09	0.09	0.08
Follower performance		−0.24***	−0.15**	−0.12*
LMX			−0.27***	−0.27***
Follower Performance × LMX				0.14***
Pseudo R ²	0.25	0.26	0.33	0.35
ΔR ²		0.01	0.07	0.02
Model deviance χ ²	173.54	152.76	120.64	109.99
Decrease in deviance: Δχ ^2d		20.78***	32.12***	10.65**

Note. LMX: leader member exchange; SD: standard deviation. n = 493. Standardized coefficients are displayed.

^a We calculated age difference as the absolute difference in age between the leader and follower.

^b Male = 0; female = 1.

^c We coded similar genders between leaders and followers as “0” and different genders as “1.”

^d The full ML estimator was used to calculate this decrease in deviance (Δχ ²) which can be considered a way of expressing effect size in multilevel modeling.

*p < 0.05; **p < 0.01; ***p < 0.001.

Figure 2.

The moderating effect of LMX on the relationship between follower work performance and abusive supervision for study 1. LMX: leader member exchange.

We also conducted ancillary analyses to examine whether abusive supervision could occur in high-quality LMX relationships. That is, we proposed a more fine-grained interaction model where the moderating role of LMX depended on employees’ willingness. As expected, willingness was negatively associated to abusive supervision (r = −0.30, p < 0.01) and positively related to follower work performance and LMX (rs = 0.21 and 0.30, respectively, all p < 0.01). For the measurement of willingness, we used a five-item scale (Fernandez & Vecchio, 1997; sample item: “Willingness to take responsibility”). The ancillary analysis showed that the relationship between performance and abusive supervision was moderated by LMX and willingness. To determine the form of the significant three-way interaction, we created four combinations of performance and abusive supervision (at one standard deviation above and below the mean) and plotted one performance–abusive supervision slope for each group. To determine the significance of the simple slopes, we used the HLM three-way interaction tool developed by Preacher et al. (2006). As shown in Figure 3, contrary to our expectations, but in line with the initial findings, the relationship between performance and abusive supervision among employees with high LMX was not significant for followers with high (−0.05, n.s.) or low (0.06, p < 0.10) willingness.

Figure 3.

Ancillary analyses for study 1.

Study 2

Method

Participants and procedures

Data were collected from 29 Norwegian business organizations, including forestry and wood construction, agricultural, real estate, automotive, and dairy producing companies. Questionnaires were distributed electronically to each respondent. Participants completed the questionnaire during working hours and responses to the items were collected electronically. Respondents were informed that the data were collected for academic research and were assured of the confidentiality of their responses.

The questionnaire was distributed to 88 leaders and 532 followers, and nearly 86% completed the survey. The leader sample contained 85.2% males, mean age was 42.4 years and they averaged 4.57 years in their current job in the company. The leaders were employed as shift supervisor, operations manager, executive director, production supervisor, managing director, sales director, project director, director of operations, production leader, and customer service manager. The follower sample consisted of 61% males. Their mean age was 44.6 years. The average time working for their organization (organizational tenure) was 7.29 years, and average years working for their supervisor (dyad tenure) was 3.49 years. They were employed in various positions such as process operator, sales, coordinator, process techniques, system engineer, laboratory engineer, laboratory assistant, advisor, and high-level employee reporting directly to middle and upper management supervisors.

Instruments

Like study 1, followers completed Tepper’s (2000) 15-item version instrument to measure abusive supervision, with sample items: “My supervisor tells me my thoughts and feelings are stupid”; “My supervisor doesn’t give me credit for jobs requiring a lot of effort”; anchors: 1 = never, 2 = seldom, 3 = occasionally, 4 = often, 5 = always.

Social loafing scale was measured by each supervisor, using four items adapted from Kidwell and Robie (2003). Sample items: “This group member takes it easy if others are around to do the work”; “This group member gives less than 100 percent effort”; anchors: 1 = very inaccurate, 7 = very accurate. The social loafing scale measures the extent to which an individual tends to do less than his or her share of work when other group members are available.

Work-related NA was assessed with six items from Watson et al.’s (1988) and Clark, Watson’s and Mineka (1994) Positive and Negative Affect Schedule. Participants were instructed to indicate the amount to which any part of their job (e.g. the work, coworkers, supervisor, clients, and pay) had elicited a particular emotion in the past 30 days. Items were rated on a five-point frequency scale ranging from 1 (not at all) to 5 (extremely).This shortened NA scale has been used and supported by previous abusive supervision studies (e.g. Tepper et al., 2006). Specifically, subordinates were asked to indicate the extent to which they had emotionally felt “afraid,” “nervous,” “irritable,” “hostile,” “upset,” and “distressed.”

Control variables

Like Study 1, we controlled for the age and gender of both the leader and the follower. Furthermore, we controlled for gender differences (similar genders coded “0”; dissimilar genders coded “1”) and age differences (using an absolute difference score), because relational demographics may relate to abusive supervision. In addition, we controlled for both organizational tenure and dyad tenure because the length of time the follower had worked in the organization and under the same leader could influence the hypothesized relationships.

Results

Descriptive statistics

Descriptive statistics and intercorrelations among the study variables are outlined in Table 3. The coefficient α estimates for the multi-item scales are listed on the primary diagonal of the intercorrelation matrix. The composite reliability coefficients were in an acceptable range for all the variables of interest, ranging from 0.82 to 0.97. As expected, abusive supervision was significantly and positively correlated with work-related NA and social loafing (rs = 0.18 and 0.18, respectively, all p < 0.01). Furthermore, work-related affect was positively related to social loafing (r = 0.13, p < 0.01).

Table 3.

Means, standard deviations, reliabilities, and intercorrelations for study 2.

		Mean	SD	1	2	3	4	5	6	7	8	9	10	11
1.	Leader age	42.36	7.82
2.	Follower age	44.55	14.03	0.07
3.	Age difference	10.61	11.83	−0.08	0.58**
4.	Leader gender	0.16	0.36	−0.19**	−0.03	0.00
5.	Follower gender	0.30	0.46	−0.16**	0.00	0.01	0.23**
6.	Gender difference	0.28	0.45	−0.08	−0.01	−0.03	0.16**	0.64**
7.	Dyad tenure	77.36	76.50	0.25**	0.11*	−0.03	−0.10*	−0.06	0.01
8.	Organizational tenure	87.42	86.78	0.02	0.32**	0.09	0.05	−0.00	−0.02	0.24**
9.	Abusive supervision	1.23	0.35	0.19**	0.02	0.01	−0.08	−0.09	−0.09	0.14**	0.16**	(0.97)
10.	Work-related affect	2.05	0.52	−0.10*	0.00	0.05	0.06	0.01	−0.09	−0.07	−0.09	0.18**	(0.82)
11.	Social loafing	2.57	1.44	0.00	0.04	0.03	−0.10*	−0.12*	−0.05	−0.01	0.10*	0.18**	0.13**	(0.91)

Note. SD: standard deviation. n = 532. Composite reliability estimates on primary diagonal.

*p < 0.05; **p < 0.01.

Confirmatory factor analysis

To determine the adequacy of our measurement model, a CFA was performed on a model with 25 items loading on three latent variables specified without cross loadings. The results revealed acceptable fit (Hair et al., 2010): χ ²[272] = 682.90, p < 0.01, RMSEA = 0.053, CFI = 0.94, NNFI/TLI = 0.94. Furthermore, convergent validity is supported, as all factor loadings were statistically significant with a mean standardized loading of 0.78 (Chin, 1998).

MSEM analyses

Because of the multilevel nature of our data where the employees are nested within groups, we tested the mediated relationship via a multilevel structural equation model (MSEM) using Mplus (Mplus 8). Although we could have tested our “1–1–1 mediation model” using multilevel modeling (MLM), a drawback of MLM is that the estimated indirect effect is conflated by the within and between components of the mediation model (Preacher et al., 2010, 2011). An advantage of MSEM, however, is that it separates the within and between components and allows for an examination of the hypothesized mediated relationship at each level (Preacher et al., 2011).

Prior to testing the hypotheses, we examine group-level variability by estimating the intraclass correlations of the latent variables using the Excel tool which accompanies Biemann et al. (2012). The ICC₁s of abusive supervision (ICC = 0.22), social loafing (ICC = 0.23), and work-related affect (ICC = 0.35) suggested that a significant proportion of the variability in abusive supervision (22%), social loafing (23%), and work-related affect (35%) was attributable to between-group variability, thus underlining the appropriateness of MSEM. In addition, the group-level reliability as shown by the ICC₂s was acceptable for abusive leadership (ICC₂ = 0.59), work-related affect (ICC₂ = 0.74), and social loafing (ICC₂ = 0.64). That is, “Common practice suggests that values of 0.70 and higher are acceptable, values between 0.50 and 0.70 are marginal, and values lower than 0.50 are poor” (Klein et al., 2000, p. 518). This indicated enough agreement between the employees (nested within groups) to continue with the MSEM approach to test the hypotheses.

The MSEM path model we estimated demonstrated excellent model fit (χ ²[0] = 0.00, n.s.; RMSEA = 0.00; CFI = 1.00; SRMR_within = 0.000; SRMR_between = 0.001). As shown in Figure 4, the results unveiled a positive relationship between abusive supervision and work-related affect at the within level (0.35, p < 0.01), as well as a positive relationship between work-related affect and social loafing at the between level (0.74, p < 0.05). However, no significant relationship was obtained between work-related affect and social loafing at the within level (0.16, n.s.) or between abusive supervision and work-related affect at the between level (−0.11, n.s.). Accordingly, we did not receive support for hypothesis 3, which predicted a positive association between abusive supervision and subordinate social loafing via work-related NA. Similarly, we did not receive support for hypothesis 4, which predicted a positive association between abusive supervision and social loafing via work-related NA at the group level. It should, however, be noted that the results also revealed a direct positive association between abusive supervision and social loafing at the within (0.81, p < 0.05), but not the between (0.12, n.s.) level.

Figure 4.

MSEM path model. Note: n = 532 (employees) and n = 88 (groups). Nonstandardized coefficients are shown as only these are available in conjunction with “type = two-level random” in Mplus. MSEM: multilevel structural equation model.

Discussion

The present study aimed first to examine the validity of Tepper et al.’s (2011) finding that subordinates perceived to be lower performers become targets of abusive supervision. The second aim was to examine whether LMX relationships can have a “buffering” effect on the performance/abusive supervision association, and third, to investigate whether abusive supervision can generate NA, which in turn can trigger dysfunctional behavior such as social loafing. And finally, based on the deontic justice perspective, to investigate whether employees who see their coworkers being treated unfairly would attempt to retaliate against the instigator. We relied on social exchange theory as an overarching framework when developing our hypotheses. In this section, we outline the theoretical and practical implications of the results, present the limitations of our study, and suggest directions for future research.

Targets of abusive supervision

Social exchange theory predicts that when one party fails to fulfil promised obligations, the other party is likely to reciprocate either by reducing positive attitudes or behavior or by getting even through negative actions (De Ruiter et al., 2016). Comparable with the Tepper et al. (2011) study, we found support from the hypothesized negative relationship between follower work performance and abusive leadership. This finding is consistent with social exchange theory. Followers who fail to perform as expected may find that their leader is likely to reciprocate by getting even through negative actions. This finding is also consistent with the victim precipitation literature, which predicts that abusive supervision is more likely to occur if followers do not meet required performance standards and are perceived as annoying and difficult to work with (Olweus, 1978). The Tepper et al. (2011) study was based on the data adopted from health-care organizations in the United States and suggested future research should conduct cross-industry comparisons to determine whether their results generalize to different industries. The present study uses data from business settings, a population not examined in previous research. The data analysis shows that the suggested prediction of abusive supervision is valid in these for-profit settings.

With respect to our examination of whether LMX relationships can have a “buffering” effect on the performance/abusive supervision association, it seems that leaders’ abusive behavior depends on the relationship between leader and their followers. Followers in high-quality working relationships with their leaders will experience more constructive treatment, where leaders avoid abusive behavior and probably focus on more constructive approaches to enhance follower performance. Such treatment will best serve leaders’ interests, as improved follower performance benefits the organization and also helps to achieve individual leaders’ goals. This finding is also consistent with the buffering hypothesis (e.g. Bal et al., 2010), where lower performing followers may be less likely to become targets of abusive behavior by their leader when they are in a high-quality LMX relationship. As such, when the relationship is of high quality, leaders may perceive low performance as less intentional and more attributable to circumstances. In contrast, when the relationship is of poor quality, the leader may attribute a follower’s low performance to personal disposition. As leaders who attribute the cause of poor performance to personal dispositions rather than the work situation (an external cause) are likely to react more harshly (Nelsen and Quick, 1997), attribution theory may provide a potential explanation for the observed interaction effects in the present study. We therefore encourage future research to explore the role of attribution when investigating similar relationships as those investigated in the present study.

Affective and behavioral responses

In our second study, we addressed two aspects of abusive supervision—the role of follower affect engendered by hostile behavior by the leader and followers’ response to rectifying the feeling of being mistreated. MSEM analyses of the data showed a positive association between abusive supervision and social loafing at the within level. This is in line with social exchange theory, suggesting that followers are likely to retaliate for harm inflicted upon them. In accordance with affective events theory (Weiss and Cropanzano, 1996), we argued that individuals who have experienced a negative workplace event (i.e. exposure to abusive supervision) would not immediately engage in a rational appraisal of the situation but react with negative emotional affect which in turn would trigger revenge such as social loafing. Our results showed a positive association between abusive supervision and work-related affect at the within level. However, and to our surprise, no significant relationship was obtained for the suggested association between work-related affect and social loafing at the within level. Our finding seems to unveil two separate consequences of abusive supervision. First, consistent with social exchange theory which conceptualizes the exchanges as a relatively rational calculative process, followers feel compelled to reciprocate by limiting their personal effort and contribution to the firm through the process of social loafing, even though such action may jeopardize their careers. Second, abusive behavior, where leaders ignore followers, humiliate them in front of others, sabotage, publicly ridicule, and coerce them, has affective consequences, where followers feel overwhelmed and upset by emotions like fear, nervousness, irritability, and hostility. According to affective event theory, emotional reactions to work events accumulate over time and influence individuals’ attitudes and work behavior. However, according to our data, negative work-related affect does not seem to mediate the abusive supervision/social loafing relationship. Although Zhao et al. (2007) criticized social exchange theory for ignoring the role of emotions and assuming that all employee reactions result from cognitive and objective judgments, our data suggest that the examined association is a rational process. Still, emotions are important as we did find a link between abusive supervision and work-related affect on the group level.

Finally, we suggested that abusive supervision is positively related to subordinate social loafing via work-related NA at the group level. The hypothesis is based on a deontic justice perspective (e.g. He et al., 2016), which posits that followers react not only to how they are themselves treated by their leader but also to how their coworkers are treated by their lender. The MSEM analysis found a positive relationship between work-related affect and social loafing at the group level, but no significant group-level relationship was obtained between abusive supervision and social loafing via work-related affect. Perhaps this result indicates that the destructive dynamic that occurs from abusive supervision is primarily present at the individual level. Although tentative, many coworkers might avoid taking part in confronting their leader with increasingly intensive resistance, thus putting their own careers at risk. Due to power differences, the likelihood that followers would retaliate directly against their immediate leader is small (Glasø et al., 2006; Zellars et al., 2002).

Practical implications

The present study found support for the suggestion that leaders may react harshly and engage in abusive supervision when followers perform below expectations. Accordingly, managers should be aware of the attribution processes by which they attribute the cause of poor performance to personal dispositions (internal cause) rather than the work situation (an external cause) itself as this may cause them to react more harshly (Nelsen and Quick, 1997). In this respect, leaders could be coached to be more self-aware of their perception and attribution processes and thereby handle the situation more constructively.

Furthermore, followers may experience different reactions from their leaders according to the quality of the LMX relationship. Leaders who have established a higher quality working relationship with their followers may give low performers fair treatment and focus on developing, challenging, and coaching these followers to enhance their performance. Followers experiencing poor LMX with their leader and not meeting expectations are liable to suffer abusive behavior. A potential devastating consequence is that low-performing followers do even more poorly and thus may confront leaders with increasingly intensive resistance. Our findings therefore suggest a “buffering effect” which general managers should make use of by actively seeking to develop high-quality relationships with their followers to prevent acceleration of this destructive dynamic. In this respect, previous research suggests that managers can help facilitate high-quality relationships with their followers by acting as a role model to their followers, provide them with challenges, delegate tasks, and inspire them to take ownership for their job (Gottfredson and Aguinis, 2016). Managers can also seek to understand the strengths and weaknesses of their followers, so that the leader is better able to assign tasks in line with their competence level.

Strengths, limitations, and future research

As with all empirical studies, a number of strengths and limitations can be identified here too. First, the present study has reduced the potential influence from common method variance by applying multisource data (supervisory rating of follower work performance and follower social loafing matched with follower self-rating of LMX, abusive supervision, and work-related NA) (MacKenzie and Podsakoff 2012). Furthermore, significant moderator effects are unlikely to be the result of common method variance (Siemsen et al., 2010). Second, like the Tepper et al. (2011) study, a cross-sectional study research design was applied, where all variables were measured at the same point in time. We, therefore, cannot exclude alternative causal ordering, such that abusive supervision affects follower work, which has also been demonstrated in previous research (Harris et al., 2007; Tepper, 2007). Still, from a theoretical point of view, low performance may influence the extent to which leaders engage in abusive supervision, which in turn leads to lower follower performance, thus becoming as a vicious cycle. However, our research indicates that this process is affected by the quality of the relationship between the two parties. Specifically, we believe that leaders in high-quality LMX relationships choose a more constructive approach to dealing with low follower work performance than leaders in low LMX relationships. Regrettably, due to the cross-sectional nature of our studies, we cannot substantiate the causality of these findings. Hence, future research should apply longitudinal data to identify a reciprocal effect between abusive supervision and subordinate resistance. Third, rigorous experimental studies might also be more appropriate for studying the causal relationships between abusive supervision, work-related NA, and social loafing. Additionally, the validity and the degree of confidence in the findings could be enhanced in future research by applying in-depth interviews or observation to check findings generated using questionnaires. Future research could also integrate both field and experimental design to explore the causality issues with a better research design.

Conclusion

Taken together our findings contribute to the abusive supervision literature by demonstrating the moderating role of LMX on the performance/abusive supervision association. In this way, the study integrates a relational perspective into abusive supervision literature. Future studies on the antecedents of abusive supervision can be extended by considering other plausible contextual variables such as organizational culture, lack of resources, or poor training. Furthermore, no support was found for the critique of social exchange theory, which posits that individuals who experience a negative workplace event will not immediately engage in rational appraisals of the situation, but will react with negative emotional affect, which in turn relates to employee behavioral outcomes such as social loafing. Rather, our findings seem to reveal two separate consequences of abusive supervision. As such, we believe these results may imply that the relationships between the cognitive, emotional, and behavioral aspects are more complex than previously assumed by prior research.

Footnotes

Declaration of conflicting interests

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

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Geir Thompson

References

Aquino

(2000) Structural and individual determinants of workplace victimization: the effects of hierarchical status and conflict management style. Journal of Management 26(2): 171–193.

Aryee

Chen

Sun

, et al. (2007) Antecedents and outcomes of abusive supervision: test of a trickle-down model. Journal of Applied Psychology 92(1): 191.

Aselage

Eisenberger

(2003) Perceived organizational support and psychological contracts: a theoretical integration. Journal of Organizational Behavior 24(5): 491–509.

Bagozzi

(1988) On the evaluation of structural equation models. Journal of Academy of Marketing Science 16(1): 74–94.

Bal

Chiaburu

Jansen

(2010) Psychological contract breach and work performance: Is social exchange a buffer or an intensifier? Journal of Managerial Psychology 25(3): 252–273.

Barclay

Skarlicki

Pugh

(2005) Exploring the role of emotions in injustice perceptions and retaliation. Journal of Applied Psychology 90: 629–643.

Bass

Riggio

(2006) Transformational Leadership, 2nd ed. Mahwah: L. Erlbaum Associates.

Becker

Atinc

Breaugh

, et al. (2016) Statistical control in correlational studies: 10 essential recommendations for organizational researchers. Journal of Organizational Behavior 37(2): 157–167.

Benson

Hogan

(2008) How dark side leadership personality destroys trust and degrades organisational effectiveness. Organisations and People 15(3): 10–18.

10.

Biemann

Cole

Voelpel

(2012) Within-group agreement: on the use (and misuse) of r _WG and r _WG(J) in leadership research and some best practice guidelines. The Leadership Quarterly 23(1): 66–80.

11.

Blau

(1964) Exchange and Power in Social life. Piscataway: Transaction Publishers.

12.

Chin

(1998) Issues and opinion on structural equation modeling. MIS Quarterly 22(1): 7–16.

13.

Clark

Watson

Mineka

(1994). Temperament, personality, and the mood and anxiety disorders. Journal of Abnormal Psychology 103(1): 103.

14.

Cohen

West

, et al. (2003) Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences, 3rd ed. London: Erlbaum.

15.

De Ruiter

Schalk

Schaveling

, et al. (2016) Manager psychological contract breach and employee outcomes: the role of economic and social LMX. Presented at the Annual Meeting of the Academy of Management, Anaheim, California, United States, 5–9 August 2016.

16.

Duffy

Ganster

Pagon

(2002) Social undermining in the workplace. Academy of Management Journal 45(2): 331–351.

17.

Dulebohn

Bommer

Liden

, et al. (2012) A meta-analysis of antecedents and consequences of leader-member exchange integrating the past with an eye toward the future. Journal of Management 38(6): 1715–1759.

18.

Einarsen

Skogstad

Glasø

(2013) When leaders are bullies: concepts, antecedents, and consequences. In: Lewis

Leonard

Freedman

Passmore

(eds) The Wiley-Blackwell Handbook of the Psychology of Leadership, Change and Organizational Development. Hoboken: Wiley Blackwell, pp. 129–153.

19.

Fernandez

Vecchio

(1997). Situational leadership theory revisited: A test of an across-jobs perspective. The Leadership Quarterly 8(1): 67–84.

20.

Ferris

Treadway

Perrewé

, et al. (2007) Political skill in organizations. Journal of Management 33(3): 290–320.

21.

Fox

Spector

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(3): 291–309.

22.

Gabler

Nagy

Hill

(2014) Causes and consequences of abusive supervision in sales management: a tale of two perspectives. Psychology & Marketing 31(4): 278–293.

23.

Glasø

Einarsen

(2006). Experienced affects in leader-subordinate relationships. Scandinavian Journal of Management 22(1): 49–73.

24.

Gottfredson

Aguinis

(2016) Leadership behaviors and follower performance: deductive and inductive examination of theoretical rationales and underlying mechanisms. Journal of Organizational Behavior 38: 558–591.

25.

Gouldner

(1960) The norm of reciprocity: a preliminary statement. American Sociological Review 25: 161–178.

26.

Graen

Uhl-Bien

(1995) Relationship-based approach to leadership: development of leader-member exchange (LMX) theory of leadership over 25 years: applying a multi-level multi-domain perspective. The Leadership Quarterly 6(2): 219–247.

27.

Hair

Jr Black

Babin

, et al. (2010) Multivariate Data Analysis with Readings. Upper Saddle River: Prentice Hall.

28.

Harris

Harvey

Kacmar

(2011) Abusive supervisory reactions to coworker relationship conflict. The Leadership Quarterly 22(5): 1010–1023.

29.

Harris

Kacmar

Zivnuska

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

30.

Hershcovis

Barling

(2010). Towards a multi‐foci approach to workplace aggression: A meta‐analytic review of outcomes from different perpetrators. Journal of Organizational Behavior 31(1): 24–44.

31.

Fehr

Yam

, et al. (2016) Interactional justice, leader–member exchange, and employee performance: examining the moderating role of justice differentiation. Journal of Organizational Behavior 38: 537–557.

32.

Hofmann

Griffin

Gavin

(2000) The application of hierarchical linear modeling to organizational research. In: Klein

Kozlowski

SWJ

(eds) Multilevel Theory, Research, and Methods in Organizations: Foundations, Extensions, and new Directions. San Francisco: Jossey-Bass, pp. 467–511.

33.

Hoobler

Brass

(2006) Abusive supervision and family undermining as displaced aggression. Journal of Applied Psychology 91(5): 1125.

34.

Karau

Williams

(1993) Social loafing: a meta-analytic review and theoretical integration. Journal of Personality and Social Psychology 65(4): 681.

35.

Kidwell

Jr Robie

(2003) Withholding effort in organizations: toward development and validation of a measure. Journal of Business and Psychology 17(4): 537–561.

36.

Klein

Bliese

Kozlowski

, et al. (2000) Multilevel analytical techniques: commonalities, differences, and continuing questions. In: Klein

Kozlowski

SWJ

(eds) Multilevel Theory, Research, and Methods in Organizations: Foundations, Extensions, and new Directions. San Francisco: Jossey-Bass, pp. 512–553.

37.

Korsgaard

Meglino

Lester

, et al. (2010) Paying you back or paying me forward: understanding rewarded and unrewarded organizational citizenship behavior. Journal of Applied Psychology 95(2): 277–290.

38.

Lian

Brown

Ferris

, et al. (2014) Abusive supervision and retaliation: a self-control framework. Academy of Management Journal 57(1): 116–139.

39.

Liden

Graen

(1980) Generalizability of the vertical dyad linkage model of leadership. Academy of Management Journal 23(3): 451–465.

40.

Liden

Wayne

Jaworski

, et al. (2004) Social loafing: a field investigation. Journal of Management 30: 285–304.

41.

MacKenzie

Podsakoff

(2012) Common method bias in marketing: causes, mechanisms, and procedural remedies. Journal of Retailing 88(4): 542–555.

42.

Mitchell

Ambrose

(2007) Abusive supervision and workplace deviance and the moderating effects of negative reciprocity beliefs. Journal of Applied Psychology 92(4): 1159.

43.

Murphy

Wayne

Liden

, et al. (2003) Understanding social loafing: the role of justice perceptions and exchange relationships. Human Relations 56(1): 61–84.

44.

Nelsen

Quick

(1997) Organizational Behavior. Foundations, Realities, and Challenges. New York: West Publishing Company.

45.

Nunnally

(1978) Psychometric Theory, 2nd ed. New York: McGraw-Hill.

46.

O’Reilly

Aquino

Skarlicki

(2016) The lives of others: third parties’ responses to others’ injustice. Journal of Applied Psychology 101: 171–189.

47.

Olweus

(1978) Aggression in Schools: Bullies and Whipping boys. Washington: Hemisphere.

48.

Opotow

(2001) Social injustice. In: Christie

Wagner

Winter

(eds) Peace, Conflict and Violence: Peace Psychology for the 21st Century. New York: Prentice-Hall, pp. 102–109.

49.

Preacher

Curran

Bauer

(2006) Computational tools for probing interaction effects in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics 31: 437–448.

50.

Preacher

Zyphur

Zhang

(2010) A general multilevel SEM framework for assessing multilevel mediation. Psychological Method 15(3): 209.

51.

Preacher

Zhang

Zyphur

(2011) Alternative methods for assessing mediation in multilevel data: the advantages of multilevel SEM. Structural Equation Modeling 18(2): 161–182.

52.

Rhoades

Eisenberger

(2002) Perceived organizational support: a review of the literature. Journal of Applied Psychology 87(4): 698.

53.

Rousseau

(1995) Psychological Contracts in Organizations. Thousand Oaks: Sage.

54.

Schat

ACH

Desmarais

Kelloway

(2006) Exposure to workplace aggression from multiple sources: validation of a measure and test of a model. Unpublished manuscript, McMaster University, Hamilton, Canada: 331–351.

55.

Schyns

Schilling

(2012) How bad are the effects of bad leaders? A meta-analysis of destructive leadership and its outcomes. Leadership Quarterly 24: 138–158.

56.

Siemsen

Roth

Oliveira

(2010) Common method bias in regression models with linear, quadratic, and interaction effects. Organizational Research Methods 13: 456–476.

57.

Skarlicki

Kulik

(2005) Third-party reactions to employee (mis)treatment: a justice perspective. Research in Organizational Behavior 26: 183–229.

58.

Spector

Fox

(2005) The stressor-emotion model of counterproductive work behavior. In: Fox

Spector

(eds) Counterproductive Behavior. Investigation of Actors and Targets. Washington: American Psychological Association.

59.

Spencer

Rupp

(2009) Angry, guilty, and conflicted: injustice toward coworkers heightens emotional labor through cognitive and emotional mechanisms. Journal of Applied Psychology 94: 429–444.

60.

Tepper

(2000) Consequences of abusive supervision. Academy of Management Journal 43: 178–190.

61.

Tepper

(2007) Abusive supervision in work organizations: review, synthesis, and research agenda. Journal of Management 33(3): 261–289.

62.

Tepper

Duffy

Henle

, et al. (2006) Procedural injustice, victim precipitation, and abusive supervision. Personnel Psychology 59: 101–123.

63.

Tepper

Moss

Duffy

(2011) Predictors of abusive supervision: supervisor perceptions of deep-level dissimilarity, relationship conflict, and subordinate performance. Academy of Management Journal 54: 279–294.

64.

Vecchio

(1995) It’s not easy being green: jealousy and envy in the workplace. Research in Personnel and Human Resources Management 13: 201–244.

65.

Vecchio

(Ed.). (2007). Leadership: Understanding the dynamics of power and influence in organizations. University of Notre Dame Pess.

66.

Watson

Clark

Tellegen

(1988) Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology 54(6): 1063.

67.

Weiss

Cropanzano

(1996) Affective events theory: a theoretical discussion of the structure, causes and consequences of affective experiences at work. In: Staw

Cummings

(eds) Research in Organizational Behavior: An Annual Series of Analytical Essays and Critical Reviews, Vol. 18. Stamford: Elsevier Science/JAI Press, pp. 1–74.

68.

Zellars

Tepper

Duffy

(2002) Abusive supervision and subordinates’ organizational citizenship behavior. Journal of Applied Psychology 87(6): 1068.

69.

Zhao

HAO

Wayne

Glibkowski

, et al. (2007) The impact of psychological contract breach on work-related outcomes: a meta-analysis. Personnel Psychology 60(3): 647–680.