Abstract

Prior research provides various views on subordinates’ workplace deviated behavior as retaliation against supervisory abuse, the effect of abusive supervision (AS) on subordinates as a group and their interpersonal relations gets poor attention. Grounded on the social exchange theory, the present study presents a model where a group of subordinates exhibits seemingly opposite discretionary behaviors in integration to combat supervisory abuse. In particular, this study posits that subordinates who experience abuse from the same supervisor form a group. This group bond provides them with enough power to involve in deviant behavior against their supervisor and supervisor’s favored coworkers. Multiple source data were collected, and linear hierarchal regression in addition to process macro methodology was used for data analysis. Findings support the mediation hypotheses partially.

Key Words

Abusive Supervision, Prosocial Work Behaviors, Deviant Work Behaviors

Introduction

Abusive Supervision (AS) is considered the most widely researched destructive leadership style so far (Pellitier, 2010; Tepper et al., 2017). Researchers have shown a great interest in the area of abusive supervision due to enormous damaging effects and its high pervasiveness (Tepper et al., 2017). This could be due to the involvement of supervisors ?the type of leaders with whom subordinates have direct interaction on a daily basis; such frequent contacts make supervisors potentially more likely to get a chance of abusing subordinates (Purcell & Hutchinson, 2007). Scholars defined AS as “subordinates’ perceptions of the extent to which their supervisors engage in a sustained display of hostile verbal and nonverbal behaviors, excluding physical contact” (Tepper, 2000, p. 178).

Using the theory of social exchange theory, several authors have discussed subordinates retaliate against the hostile treatment of supervisors (Blau, 1964; Tepper, 2007). However, due to unparallel positions, subordinates don’t show their antagonism openly, as this may result in counter-reaction; hence they change their discretional behaviors like Deviant work behaviors (DWBS)  and prosocial behaviors (PSBs), they are considered a safer way for subordinates to react against AS because these actions are not treated as part of their usual job description; therefore they are not assessed through formal procedures, which makes them less observable and hence less punishable ( Liu & Wang, 2013; Dalal, 2005). DWBS are defined as any deliberate behavior contrary to the organization’s interest by an organizational member (Sackett, 2002). On the other hand, PSBS indicates the behaviors that support the psychological and social environment of the organization (Organ, 1997). Dalal et al., 2009 established that AS affects inter-personal discretionary behaviors (PSBS and DWBS) more than any other organizational factor. In this regard, management scholars asserted that subordinates have diverse inter-personal relationships with coworkers, with some considered close friends, some considered threats, and having neutral feelings toward others, depending on the prevailing circumstances (Dalal, 2005; Venkataramani & Dalal, 2007). The present research study focuses on AS as one of the conditions where supervisors exhibit differentiated behaviors towards subordinates (Wu et al., 2010) and become the source of subordinates’ interpersonal discriminated behaviors.

The differentiated behavior of the supervisor refers to the subordinates’ perception that the leader is supporting some members of the group while mistreating others, thereby creating in-group and out-group members (Dalal,2005). Abused subordinates view the supervisor’s actions in the social context, and when they perceive that others are being treated differently by the same supervisor, they change their responses accordingly (Skarlicki & Rupp, 2010). Abused subordinates considering themselves as out-group members; as a result, they get attracted and support other out-group members to exercise greater reaction against the supervisors and their allies (non-abused coworkers) (Arain et al., 2018). Thus, the response to AS leads to a person forming varied behaviors, such as Prosocial behaviors toward other abused coworkers (PSBS-CW) and DWBS toward the supervisor (DWBS-S) and non-abused coworkers (DWBS-NCW) (Scott & Lane, 2000). The objective of this study is to examine the diverse discretionary behaviors of abused subordinates towards different individuals in the organization according to their perceptions; the study captures abused employee behavior a) towards their supervisors, b) towards abused coworkers, and c) towards non- abused coworkers. Hence, this study will assess both the adverse effect of AS on supervisor-subordinate and subordinate-subordinate linkages.

This study makes two significant contributions. First, it is presenting a mechanism through which a group of abused subordinates tends to react against their supervisor by means of group support. Tepper et al. (2017) Asserted that several studies had examined the effect of AS on DWBS without considering causal mechanisms. They claimed that, to date, the cause for subordinates’ reacting to AS with DWBS remains vague. They highlighted a need to study more mechanisms and the practical perspectives that may have an impact on the relationship between AS and DWBS under certain contexts. Second, this study also contributes to the literature by examining the retaliation of abused subordinates as a group, toward distinct coworkers and abusive supervisors, as a considerable amount of work on AS has studied the retaliation against supervisory abuse individually without considering the situational factors, such as group effect (Vidyarthi et al., 2010). The present study explores PSBS-CW (group), which depicts prosocial behaviors among abused coworkers as a mediating variable between AS (group) and reaction against supervisory abuse.

Theory and Hypothesis

Methodology

Sample and Data

Data from 1500 subordinates were collected working in fifteen large to medium scale organizations in Pakistan from multi sectors. According to Gong et al. (2010), the most appropriate respondents for such studies are subordinates having daily contact with their supervisors and working as a group consisting of five people; AS have to be averaged for having the data of AS (group). Out of the 1500 questionnaires, 920 responses were found complete and authentic, belonging to 184 groups, making a response rate of 46 percent. 

Measures

Abusive supervisor. A five-item scale by Priesemuth et al. (2014) was adopted for this study. A 5-point Likert scale was used, ranging from 1 [never] to 5 [almost always]. For AS, the data was aggregated by calculating the average of responses attained from abused subordinates group members. 




PSBS (Abused Coworkers): A six-item scale was used to measure PSBS- ACW given by Priesmuth et al. (2014). PSBS-ACW (group) is assessed by inquiring the focal person (abused subordinate) about the coworkers’ support. We averaged the individual level PSBS (ACW) to attain group PSB (ACW). Data were obtained on 5-point Likert Scale ranging from 1 (never) to 5 (always).




DWBS (Supervisor) and DWBS (Non-Abused Coworker): The scale developed by Dalal (2009) were modified and used in this study. All observations from peer groups were then averaged to get group data. A five-point Likert scale was used ranging from 1 (never) to 5 (always)., each consisting of six items. For DWBS-S (Group), the individual level perceptions were aggregated by calculating the mean value of responses across group members. 




Control Variables: 3 demographic variables were identified from the literature that should be controlled that include gender, age and gender tenure (Robbins et al., 2011). 

Results

Validity and Reliability

For the fitness of the model, a confirmatory test was performed. A model was specified in which all items were loaded on their respective latent constructs. The four-factor model depicts acceptable fit statistics (GFI=.87, CFI=.93, RMSEA=.06, TLI=.92 and AGFI=.86) (Schreiber et al., 2006).

Besides the model fit, the reliability and validity of the constructs were also tested. To test the discriminant and convergent validity, the confirmatory test was run to achieve standardize loading estimates (Gaskin, 2016). Convergent validity was assessed by considering the factor loadings estimates; in addition, average variance extracted (AVE) scores were also calculated for the constructs. The findings (shown in Table I) indicates that all AVE values were greater than 0.50, hence supporting reasonable convergent validity (Fornell & Larcker, 1981).

To check discriminant validity, values of AVE and maximum shared squared variance (MSV) were compared. AVE values for all indicators are higher than MSV. Hence satisfactory results were obtained for discriminatory validity. The square root of AVE and paired correlation coefficients were also compared (shown in Table II). The results establish that the paired correlation coefficients are greater than square roots of AVEs, hence proving discriminant validity. The values of Cronbach alpha were also more than 0.80, which confirms the reliability.

 

Table I. Standardized Loading Estimates, Cronbach’ ?, AVE, MSV and CR

Constructs	Items/Indicators	Standardized loading estimates	CR	AVE	MSV	Cronbach’s ?
ABS	AS 1	.008	0.956	0.656	0.196	.951
	AS 2	.827
	AS 3	.853
	AS 4	.793
PSBS-ACW	AS 5	.840	0.867	0.521	0.186	.809
	PSBS-AP 1	.822
	PSBS-AP 2	.693
	PSBS-AP 3	.741
	PSBS-AP 4	.746
	PSBS-AP 5	.760
DWBS-S	PSBS-AP 6	.743	0.878	0.548	0.233	.917
	DWBS-S 1	.667
	DWBS-S 2	.772
	DWBS-S 3	.805
	DWBS-S 4	.828
	DWBS-S 5	.654
DWBS-NCW	DWBS-S 6	.667	0.907	0.620	0.233	.911
	DWBS-NCW 1	.695
	DWBS-NCW 2	.654
	DWBS-NCW 3	.828
	DWBS-NCW 4	.805
	DWBS-NCW 5	.772
	DWBS-NCW 6	.667

Note. AS=Abusive supervision owns and peer group, PSBS-ACW=Prosocial behaviors towards abused co-workers, DWBS-S=Deviant workplace behavior towards the supervisor, DWBS-NCW=Deviant workplace behavior work behavior towards non-abused co-workers. CR= construct reliability; AVE=average variance extract, MSV= Maximum shared squared variance.

 

Correlations

The results (Table II) showed that experience and age were correlated with the dependent variable, hence needed to be controlled. While gender doesn’t have any correlation with any of the dependent and independent variables, so it is not included in hypotheses testing (Peng et al., 2014).

Table 2. Mean, Standard Deviation and Correlation

Variable	M	SD	1	2	3	4	5	6	7
Age	1.7	.82
Tenure	2.2	1.3	.79*
Gender	1.4	.47	-.14*	.156*
AS	3.2	.74	-.003	.098*	.019	.796
PSBS-ACW	2.9	.63	-.079	-.016	-.034	.432*	0.733
DWBS-S	3.2	.87	.043	.044	.059	.342*	.257*	0.792
DWBS-NCW	3.2	.95	.073*	.095*	.055	.382*	.351*	.481*	0.751

Note. Bold values in diagonals are the square root of AVE scores of latent constructs.

**significance value at 0.01 (Two-tailed)

*Significance value at .05 (Two-tailed)

Hypotheses Results

Direct Effects: We used linear hierarchal regression to test hypotheses 1a and 1b. For hypothesis 1a, no control variable proved to have significant relation with dependent variables (DWBS-S); hence, dependent (DWBS-S) and independent variable (AS) were inserted in model 1. For hypothesis 2, in model 1, two control variables, age and experience, were inserted AS they had a correlation with the independent variable (DWBS-NCW). In model 2, the dependent variable (AS) was inserted. AS reported in Table III, AS positively and significantly influences DWBS toward the abusive supervisor.

Table 3. Effect of AS (group) on DWBS-S and DWBS-NCW

Variables	B	SE	? R²
DWBS-ACW Model 1 Intercept	1.86**	.114	.109**
ABS DWBS-NCW Model 1	.391**	.037	.009
Intercept	2.875**	.071
Age	-.006	.061
Experience Model 2	.082	.044	.138**
Intercept	1.415*	.137
Age	.070	.057
Experience	.012	.041
AS	.489*	.040

** significance at 0.01 level

* Significance at 0.05 level.

 

Mediated Effects: The mediated relationship between AS (group) and DWBS (supervisor) through PSBS (peer) was tested by using a process macro method (Preacher et al., 2007). Model 4 was used to analyze the mediation results (Preacher et al., 2007).

The summary of results in Table IV represents that AS (group) has a significant and positive impact on DWBS-S. The significant indirect effect of AS on DWBS-S through PSBS toward abused coworkers indicated the presence of a mediation effect. The direct effect of AS (group) on DWBS (supervisor) is also significant in the mediated model. The significance of both direct and indirect effect indicates partial mediation of PSBS toward abused coworkers between AS (group) and DWBS toward the supervisor

Table V establishes that DWBS toward non-abused coworkers has a positive and significant association with AS. The significant indirect effect of AS on DWBS-NCW through PSBS-ACW indicates the presence of a mediation effect. The direct effects of AS on DWBS toward the supervisor also remained significant in the mediated model. The significance of both of these effects confirms partial mediation by PSBS toward abused coworkers between AS and DWBS toward a non-abused coworkers.

Table 4. Mediating Effect of PSBS-APS between AS and DWBS-S

	B	SE	95% CI	R²
Constant	1.5229	.1371	1.2539,1.7918
PSBS-ACW (mediator)	.2105	.0478	.1167,.3042
AS (independent variable)	.3143	.0404	.2350,.3935
Model R²				.3574*
The indirect effect of independent
variable on dependent variable	.0764	.0206	.0339,.1145

** significance at the 0.01 level (2-tailed).

* Significance at the 0.05 level (2-tailed).

 

Table 5. Mediating Effect of PSBS-APS between AS and DWBS-NPS

	B	SE	95% CI	R²
Constant	.8165	.1598	.5029,1.1300
PSBS-ACW (mediator)	.3485	.0510	.2485,.4486
ABS (independent variable)	.3629	.0433	.2780,.4479
Age (control variable)	.0948	.0554	-.0139,.2035
Experience (Control	.0076	.0403	-.1715,.0867
variable)
Model R²				.1884**
Direct effect of independent on dependent	.37	.043	.2782,.4480
Indirect effect of independent variable on dependent variable	.1267	.0234	.0825, .1726

** significance at 0.01 level

* Significance at 0.05 level.

Discussion

This study proves that the subordinates facing supervisory abuse form groups and further use this group to retaliate against the supervisor and his/her favored employees. The study proves the positive relationship of AS and DWBS against supervisors (Mitchell & Ambrose, 2007); it also proves subordinates facing supervisory abuse exhibit negative behaviors towards non-abused coworkers as they perceive that these subordinates are supported by supervisors.  Previous studies have also confirmed this relationship by establishing that the subordinates demand equitable behaviors from supervisors, and if they feel that behavior is not equitable, they retaliate against both the supervisor and his/her favored employees (Mitchell and Ambrose, 2007). This study proves that the retaliation against favored coworkers is higher than the supervisor. This phenomenon can be explained by the fact that due to power distance, the retaliation against the supervisor is difficult; rather, abused subordinates choose to retaliate against their coworkers, who are being supported by the supervisor.

The study establishes two opposing relationships. PSBS towards abused coworkers and DWBS towards supervisor’s favored coworkers. Previous studies supported the notion that employees’ inter-personal interaction with each other depends on supervisors’ discriminatory behaviors; a supervisor’s favoritism toward one group and abusive behaviors toward others develops envious sentiments among the favored and non-favoured subordinates (Peng et al., 2014).

The current study also proved PSBS as a mediator between AS and DWBS. This behavior can be explained by the fact that as it is difficult to retaliate against a supervisor so abused subordinates practice group power to show resentment against the higher authority. Apparently, opposite behaviors PSBS and DWBS become cause and outcome; in this regard, literature states that an employee who exercises high-moral behaviors toward one group gets approval to exhibit DWBS towards any other factor of the organization without any image distortion (Merritt et al., 2010).

Managerial Implications

The current study contributed to leadership literature by highlighting an important and missing aspect that is abused supervisor and subordinate exchange process (Tepper et al., 2009). The present study highlights one of the situations where supervisors show discriminatory behaviors towards subordinates, and subordinates respond by supporting each other against the supervisor and his/her allies.

The present study also added to group literature. It proved that abused subordinates work in the group against abusive supervisors. 




Practical Implications

The present study proved the implicit detrimental effect of AS; it has a dual effect, first on the relationship between supervisor and subordinate but also the relationship among subordinates. So, supervisors should be aware of the outcomes of their behaviors. 

Moreover, awareness training programs should be conducted in organizations to give leaders an awareness about the implicit outcomes of their discretionary behaviors. Moreover, organizations should try to create a psychologically safe environment and ethical code of conduct that enables subordinates to fight against any abusive behaviors of supervisors.

Limitations and Future Research Directions

Though we have tried our best to cover all the aspects like every other study, the present study also contains certain limitations. First, it has a limitation of cross-sectional design. Though procedural remedial approaches were used to remove common method bias, Future studies can consider multiple source data and longitudinal design to have a more comprehensive picture.

Secondly, this study considers only one mediator (PSBS); future studies can test more mediators that could have a possible role in the relationship of AS and DWBS.

References

• Aquino, K., & DouglaS, S. (2003). Identity threat and antisocial behavior in organizations: The moderating effects of individual differences, aggressive modeling, and hierarchical status. Organizational Behavior and Human Decision Processes, 90(1), 195-208
• Arain, G. A., Bukhari, S., Khan, A. K., & Hameed, I. (2018). The impact of AS on subordinates' feedback avoidance and subsequent help-seeking behaviour: A moderated mediation model. Journal of Management & Organization, 1-16.
• Blau, P. M. (1964). Exchange and power in social life. Transaction Publishers.
• Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in organizations: A meta-analysis. Organizational behavior and human decision processes, 86(2), 278- 321.
• Dalal, R. S. (2005). A Meta-Analysis of the Relationship Between Prosocial behavior and Counterproductive Work Behavior. Journal of Applied Psychology, 90(6), 1241-1255.
• Fornell, C., & Larcker, D. F. (1981). Structural equation models with unobservable variables and measurement error: Algebra and statistics. Journal of marketing research, 18(3), 382-388.
• Samreen, F., Rashid, M. A., & Hussain, G. (2019). Effect of AS on subordinates' discretionary behaviors. Journal of Management and Organization, 2019(57),1-16
• Gong, Y., Chang, S., & Cheung, S. Y. (2010). High performance work system and collective PSBS: A collective social exchange perspective. Human Resource Management Journal, 20(2), 119- 137.
• Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement. American sociological review, 25(2), 161-178.
• Harris, K. J., Harvey, P., Harris, R. B., & Cast, M. (2013). An investigation of ABS, vicarious ABS, and their joint impacts. The Journal of social psychology, 153(1), 38- 50.
• Hung, T. K., Chi, N. W., & Lu, W. L. (2009). Exploring the relationships between perceived coworker loafing and counterproductive work behaviors: The mediating role of a revenge motive. Journal of Business and Psychology, 24(3), 257-270.
• Kelloway, E. K., Francis, L., Prosser, M., & Cameron, J. E. (2010). Counterproductive work behavior as protest. Human Resource Management Review, 20(1), 18-2
• Liu, X. Y., & Wang, J. (2013). ABS and organizational citizenship behaviour: is supervisor-subordinate guanxi a mediator? The International Journal of Human Resource Management, 24(7), 1471- 1489.
• Martin, R., Guillaume, Y., Thomas, G., Lee, A., & Epitropaki, O. (2016). Leader- Member exchange (LMX) and performance: A Meta-Analytic review. Personnel Psychology, 69(1), 67-121
• Martinko, M. J., Harvey, P., Brees, J. R., & Mackey, J. (2013). A review of ABS research. Journal of Organizational Behavior, 34(S1), S120-S137.