«UNDERSTANDING EMPLOYEES’ BEHAVIORAL REACTIONS TO AGGRESSION IN ORGANIZATIONS by MARIE S. MITCHELL B.S., George Mason University, 1993 M.A.H.R., ...»
Researchers have argued that common method variance may not present a biasing problem when the measures of the variables used have been properly developed (Spector, 1987). Nevertheless, I followed Podsakoff et al.’s (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) recommendation and conducted the Harmon’s single-factor test. Principle components analysis was conducted on all scale items included in this study. If common method variance is present, the analysis should reveal either one single factor or a dominant general factor that accounts for a majority of the variance in the responses. Neither of these emerged; instead, no factor contributed to more than 18.69% of the overall cumulative variance (of 72.14%). Further, Podsakoff et al. recommend other, non-statistical methods for reducing common method variance. In particular, they recommend protecting respondents’ anonymity and ensuring them that the survey contains questions in which there are no “right or wrong answers.” These steps reduce respondents’ evaluation apprehension and allow for more accurate responses. Not only do the instructions on the survey explicitly state that responses will be held in absolute anonymity and that there are no right and wrong answers to the survey questions, but also the researcher discussed these specific issues when describing the nature of this study to the participants.
Third, the use of self-reported measures to assess sensitive behavior (e.g., aggression and withdrawal) may enhance social desirability bias. Indeed, some researchers contend the use of objective data is warranted in workplace aggression research (e.g., Greenberg & Folger, 1988);
however, objective data also suffer from criterion deficiency because organizations generally only report these types of behaviors if the employee has been caught and reprimanded (Fox & Spector, 1999). In contrast, self-reports can potentially offer more accurate reflections of employees enacted behavior, particularly if anonymity is assured and they believe their employers will not punish them for reporting their behavior (Lee, 1993). The sample used in this study provides a situation in which fear of future punishments is not a concern for respondents and anonymity was assured. Further, methodological precautions were taken by controlling for social desirability in the analysis.
A last limitation involves measurement. In particular, a lack of measurement of other dependant variables may explain the lack of results. The measures used in this study strove to best capture the categories of behaviors identified in the typology. Nevertheless, the lack of findings may be explained by a lack of measurement of other types of reactions to perceived supervisor aggression. For example, perhaps the targets of displaced aggression were limited and did not include other potential targets (i.e., strangers, family, friends). Further, this study assessed only task withdrawal. However, other forms of withdrawal also may occur (i.e., job withdrawal—exiting the organization, transferring departments, being absent; Hanisch & Hulin, 1991). Similarly, the constructive problem-solving measure may not have included the broadrange of constructive reactions that an individual may have engaged in as a consequence of perceived supervisor aggression. Thus, the limited forms of assessed dependent variables might explain the non-supported results in the study.
Implications The possible limitations and the nature of some of the unexpected findings highlight a number of directions worth pursuing in future research. Appendix D provides a table that summarizes the overall findings of the study. Findings that were significant and as predicted are highlighted in yellow. Findings that were significant but contrary to predictions are highlighted in green and non-significant findings are white. I use this table as a baseline for my discussions for future research.
One variable that consistently moderated reactions to perceived supervisor aggression is fear of retaliation. Indeed, fear of retaliation influenced the relationship between perceived supervisor aggression and all reactions but withdrawal. Consistent with Dollard et al.’s (1939) frustration-aggression theory, fear of retaliation inhibited retaliatory reactions to aggression.
However, the results do not support Dollard et al.’s contentions about displaced aggression.
Instead, fear of retaliation seemed to weaken the relationship of perceived supervisor aggression and all forms of displaced aggression (toward the organization, coworkers and customers). Thus, individuals victimized by an aggressive supervisor were far less likely to displace aggression when they feared retaliation from the supervisor than those with low fear of retaliation.
These findings do not suggest displaced aggression is as much of an emotional, cathartic reaction as Dollard et al. proposed. Instead, individuals may have chosen not vent their frustration on others either because it would potentially trigger further attacks by their supervisor or promote aggression against them by others. These findings are consistent with both Miller (1948) and Sears (1948), who argued that learned experiences may preclude aggressive reactions. Similarly, in their review of the frustration-aggression theory, Tedeschi and Felson (1994) found some studies suggest individuals may discriminate when to or not to displace aggression. Thus, the question becomes “to whom do employees displace aggression?” Social psychology research has identified factors that influence the target of displaced aggression (e.g., similarity of the target to the source of the harm, triggering events; cf. Marcus-Newhall et al., 2000). The results of this study suggest how “safe” it is to displace aggression on the target may also be an issue. Nevertheless, further investigation is needed to understand displaced aggression as a reaction to perceived supervisor aggression.
The results of aggressive modeling suggest coworkers rather than supervisors provide a strong basis of normative learning effects. One of the basic tenants of social learning theory is that individuals learn from social models (Bandera, 1983), particularly agents of an organization (i.e., supervisors). The results of this study suggest otherwise. Rather, work group members appear to be a better source of vicarious learning. This argument is consistent with O’Reilly and Caldwell (1985), who contend normative expectations of the work group set the tone for how things “ought to be done” rather than how things “should be done” at work. Stated differently, employees learn more about expected and appropriate work behaviors by observing other employees than through organizational rules and policies. According to O’Reilly and Caldwell, normative expectations develop through work group interactions; and when group members exert pressure to conform to central attitudes and behaviors, they create behavioral standards.
The results of this study provide support for these contentions, suggesting that employees learn more about the appropriateness (and inappropriateness) of aggressive behavior from coworkers than supervisors. Overall, the results indicate that frequent exposure to aggressive coworkers strengthens aggressive reactions (both against the aggressive supervisor and coworkers).
Further, when individuals are not frequently exposed to aggressive coworkers, victims of perceived supervisor aggression were less likely to displace aggression on coworkers and more likely to engage in constructive problem-solving.
These findings beg the question of “from whom do employees learn constructive behavior?” Although low coworker aggressive modeling did strengthen constructive problemsolving reactions to perceived supervisor aggression, perhaps frequent exposure to constructive behaviors by work group members might strengthen these types of reactions to aggression even further. In sum, more investigation is needed to fully understand the impact of social learning effects on reactions to perceived aggression.
Further, if constructive or destructive behaviors are norms shared among many organizational members, this is suggestive of an organizational climate. Shared perceptions of aggressive behaviors would suggest the organization has an “aggressive climate,” whereas shared perceptions of constructive behaviors would suggest an organization has a “constructive climate.” If coworker modeling influences reactions to perceived supervisor aggression, then firm-wide shared beliefs of appropriate norms and behaviors might also influence reactions as well. This too is an area for future research.
The lack of findings for the individual personality moderators is also noteworthy. A long stream of research has evolved in social psychology aggression research that demonstrates the effects of various individual characteristics (see Anderson & Bushman, 2002, for a review).
Consistent effects of personality were not captured in this study. However, rather than abandon research on individual characteristics, it may be useful to consider if different individual factors might influence reactions to perceived aggression. For example, perhaps individual factors such as proactive orientation or moral identity would have a stronger influence on reactions to perceived aggression.
Another pattern that emerged from the findings is that the variables seemed to more consistently influence retaliatory reactions to supervisor aggression. Although qualitative research suggests that individuals do not always engage in retaliation (e.g., Tripp & Bies, 1997), the results of this study suggest retaliation might be a more dominant response. Yet, further investigation is necessary to determine whether individuals principally seek to retaliate against the source of harm and whether retaliation is an initial reaction. For example, if an individual retaliates and is thereafter still angry, maybe then he/she will engage in displaced aggression, or withdrawal. Therefore, future research should explore primary and secondary reactions to aggression to see whether an order effect occurs.
Aside from an ordering effect, retaliation might have been a dominant response because other types of reactions were not adequately assessed. In the previous section, I noted that aspects of the model warrant reconsideration. Specifically, the results might also be explained by assessing different forms of the dependent variables (e.g., job withdrawal, different targets of displaced aggression and other forms of constructive problem-solving). One other explanation is that non-behavioral reactions are also important to understand reactions to perceived aggression.
In this study, I purposefully limited the scope of reactions to behavioral reactions. However, research in the stress and aggression literatures suggest emotional and psychological reactions guide behavioral choices to aversive events. For example, Anderson and Bushman (2002) developed an integrated model of aggression, which suggests emotional reactions feed into psychological processes that form the basis of cognitive appraisals of aversive events. They argue that emotions and psychological states, therefore, mediate evaluations of aversive events and behavioral choices. Similarly, the stress literature suggests that individuals may engage in emotion-based coping to buffer the damaging consequences of the situation at hand (see Turner & Roszell, 1994, for a review). For example, individuals may decide to count to ten or talk with others (“venting”; cf. Folkman, 1984; Folkman & Lazarus, 1980). Therefore, future research should be conducted to not only identify the full range of behavioral reactions to perceived aggression but also explore the influence of emotional and psychological states on behavioral reactions.
In sum, the results of this study suggest a great deal of research is yet to be conducted to understand employees’ reactions to perceived supervisor aggression. Future research should consider the broad range of reactions to aggression, as well as the factors that influence those reactions. Further, future research should also consider any underlying processes (e.g., emotions and psychological states) that might influence the behaviors individuals choose to engage in as a consequence of perceived supervisor aggression. Overall, the discussion suggests a more elaborate model may necessary to fully understand reactions to aggression. Although this study specifically limited the scope to behavioral reactions (versus emotional, physical or psychological reactions), understanding these types of processes that guide behavioral choices is also clear next step.
Conclusion Workplace aggression researchers have spent considerable effort investigating the nature and general costs of aggression (see Neuman & Baron, 1998, 2005, for reviews). Consequently, we know a great deal about why individuals engage in aggressive behavior at work (see Baron, 2004). Further, we know a great deal about the sources and types of aggression that individuals engage in at work (see Neuman & Baron, 2005). However, much of this research focuses on explaining why individuals engage in workplace aggression rather than explaining how individuals respond to aggression. This study addresses this issue especially with regard to reactions to perceived supervisor aggression.
The typology and model presented serve as a necessary step in understanding employee reactions to aggression. Given the growing nature and prevalence of aggression in the workplace (Bureau of Justice Statistics, 1999), understanding which factors influence individual responses is an important step to addressing the problem of aggression in the workplace. Unfortunately, this places an enormous burden on organizational decision-makers, who are responsible for creating safe and healthy work conditions. Although this study sought identify various factors that mitigate dysfunctional reactions (i.e., retaliatory aggression, displaced aggression, and withdrawal) and enhance more constructive solutions to perceived aggression (i.e., constructive problem-solving), clearly more work is to be done in the future.
Allen, D., & Erickson, J. 1989. Sexual harassment of faculty at the University of Illinois.
Champaign, IL: Union of Professional Employees, University of Illinois at Urbana
Anderson, C. A. 1997. Effects of violent movies and trait irritability on hostile feelings and aggressive thoughts. Aggressive Behavior, 23: 161-178.