Do the bullies survive? A five-year, three-wave prospective study of indicators of expulsion in working life among perpetrators of workplace bullying.

In recent series of studies, we have shown that targets of workplace bullying are at risk of expulsion in working life, both from current employment (e.g. in terms of changing employer) and from working life itself (e.g. becoming unemployed). The most recent of these, Take It or Leave: A Five-Year Prospective Study of Workplace Bullying and Indicators of Expulsion in Working Life was recently published in Industrial Health, and the present short communication aims to follow up that paper, investigating the possible job "survival" of the perpetrators. A nationally representative sample was employed (n=1,613), and responses were gathered at three time points with a two-year and a five-year time-lag. Outcomes were intention to leave and sickness absence at T1, and sickness absence, change of employer, disability benefit recipiency and unemployment at T2 and T3. The results of regression analyses clearly indicate that the perpetrators' occupational status is largely unchanged, and remains so over time, as opposed to earlier findings regarding the targets of bullying.

Workplace bullying is characterized by repeated subjection of an employee to negative acts that the target has difficulty in defending against due to a formal or informal power-inferiority4). Moreover, bullying is often described as a gradually escalating phenomenon—a process that can last for years5), where ever more intense and frequent negative acts are directed at a subordinate or a peer, leading to victimization of the recipient. One of the first process models of workplace bullying was the four-stage model of Heinz Leymann6). Based on clinical interviews with targets of bullying, he proposed that the typical bullying scenario evolves from interpersonal conflicts and unethical communication in the first stage, to psychological violence in the second stage. At some point, management will have to intervene because the situation, or the targets themselves, demands attention. This intervention, the third stage in the process, makes the case “official”. In some instances, the situation may be resolved at this stage, but more often, management intervention is prejudiced according to Leymann6, 7). In his view, the target has the role of the accusing party at this point, while at the same time being victimized and stigmatized. Thus, the reaction may appear exaggerated, sometimes to a hysteric degree, which increases the bully’s chances of forming an alliance with top management because the bully in contrast to the victim appears reasonable and sensible. In these instances, the bullying process may reach stage four in Leymann’s process model of bullying6, 7), which involves target expulsion.

In our understanding, expulsion in working life in the aftermath of bullying involves illegitimately depriving a person of the opportunity or ability to be a free member of the workforce or participant in the labor market. Thus, the term expulsion is not limited to the act of removing someone from a given position in the organization, but includes events such as being pressured to quit, being discharged, being internally relocated or undergoing severe health impairment leading to reduced work ability. Moreover, as Leymann also claimed, once expelled from the workplace, many targets may have difficulty finding new employment, thus risking expulsion from working life altogether7).

In a recent series of studies1,2,3), we have shown that workplace bullying may in fact be an antecedent to a range of indicators of expulsion in working life. Specifically, throughout these studies, bullying has been demonstrated as an antecedent both to indirect indicators of expulsion in the form of job insecurity and intention to leave2), as well as direct indicators of expulsion, in the form of change of employer, sick-leave, disability benefit recipiency and unemployment1, 3). However, as an extension of Leymann’s process model, one important question still remains unanswered: What happens to the perpetrators? Do they “survive” in their positions while the targets are pressured out? Following up our recent Industrial Health publication3), we will in the present paper investigate if reporting being a perpetrator is a risk factor for later expulsion.

There are arguments both for and against the possibility that perpetrators of bullying face the same risk of expulsion as do targets of bullying. From a legal viewpoint it could be argued that perpetrators should risk expulsion, as bullying in many countries is prohibited by law. In Norway, for instance, the work environment act addresses the issue directly with a regulation stating that no employees shall be exposed to bullying and harassment, or any other kind of inappropriate or disrespectful behavior while at work8). And, in the “European framework agreement on harassment and violence at work” agreed upon by employers and labor-organizations in the forum of the European Social Dialogue in 2007, it is clearly stated that if harassment is confirmed, the perpetrators should be sanctioned with disciplinary action up to and including dismissals9). Moreover, the anti-social nature of bullying is in clear violation of prevailing cultural and social norms of conduct10). Accordingly, it could be argued that witnesses as well as upper management might side with the target, either to empathize or in order to restore a sense of interpersonal justice. On the other hand, witnesses could also choose to take the side of the perpetrator, in fear of becoming a target themselves11). In addition, one of the central premises in the third stage of Leymann’s process model is that the bully manages to establish an alliance with the top management, assuring that the target takes the blame7). In line with this, researchers have found that upper management is reported to be either passive, or even worsening the situation for the target in more than 70% of the cases11). Moreover, in light of previous research showing an increased risk of expulsion among targets1,2,3), an alliance between perpetrator and organization is plausible, and these findings could be understood as empirical arguments to actually expect job “survival” among the perpetrators.

With these different notions in mind, we present the following research question:

Is being a perpetrator of bullying associated with indicators of expulsion in working life, in the form of intention to leave, sickness absence, change of employer, unemployment and disability benefit recipiency?

Via Statistics Norway (SSB), 4,500 employees were randomly selected from the Norwegian Central Employee Register and presented with an opportunity to participate in a survey about psychosocial working conditions12). Participation was voluntary and confidential, and participants could choose to resign from the study at any time. This was informed via an information letter enclosed with the survey questionnaire, and any returned questionnaire was regarded as informed consent. In order to assure that the sample was not characterized by respondents unrepresentative of the “regular” Norwegian working population (such as family businesses employees and students working just a few hours per week), criteria for being invited to participate was being between 18 and 65 yr and working at least 15 h per wk in a company with at least five employees. The study received approval from the Regional Committee for Medical Research Ethics in Western Norway.

Baseline data were collected in 2005 (T1), with a response rate of 56.4% (n=2,539). In 2007 (T2), 70% of the baseline respondents returned the questionnaire (n=1,775), and in 2010 (T3), 72% of the T2-respondents and 11.4% of T1-respondents who had not responded at T2 returned the questionnaire (n=1,613).

Being a Perpetrator of Bullying was measured by presenting the respondents with a definition of workplace bullying, asking them to indicate whether they during the last 6 months had behaved in a manner that could be perceived as bullying towards others in the organization. The definition was presented as follows:

Bullying (for example harassment, torment, freeze-out or hurtful teasing) is a problem in some workplaces and for some employees. To be able to call something bullying, it has to occur repeatedly over a certain period of time, and the bullied person has difficulty in defending him- or herself. It is not bullying when two persons of approximately equal “strength” are in conflict, or if it is a single situation.

The response categories were (1) no, not at all, (2) yes, to some degree, and (3) yes, to a high degree. This variable was dichotomized by coding value 1 as not being a perpetrator of bullying, while values 2 and 3 were coded as being a perpetrator of bullying. This cut-off was chosen because we believe there is a high threshold for admitting to bullying one’s colleagues, and that any affirmative response is likely to reflect actual bullying behavior.

Intention to Leave at T1 was measured using a single item (i.e. It is likely that I during the next 12 months will try to get a new job) which was evaluated by the respondents using a five-point Likert-type scale ranging from “not at all true” to “completely true”. The item was dichotomized, any response above “not at all true” being coded as intention to leave.

More than 24 Sick Days during the last 12 months was used as a measure of high sickness absence. This was measured by asking the respondents to indicate how many sick days they had taken during that period. The cut-off of 24 d was chosen as it is the most common maximum amount of days that Norwegian employees can take per year without medical certification, thus being a representative indicator of high sickness absence in the Norwegian working context.

Change of Employer at T2 was measured by asking whether the respondents had changed employer since the first measurement. At T3, change of employer was measured by asking the respondents which year they had started working for their present employer. Respondents who indicated having changed employer in 2005 but who did not participate at T2 were excluded from the analyses (n=12), because it could not be known if they had changed jobs before or after the T1 measurement in 2005.

Unemployment was measured at T2 and T3 by asking the respondents directly whether or not they were currently unemployed.

Disability Benefit Recipiency was measured at T2 and T3 by asking the respondents to indicate if they were presently a recipient of disability pension or doing vocational rehabilitation. The unemployment and disability benefit recipiency variables were both measured in a common list of present work-status options, and were thus mutually exclusive.

Exposure to Bullying Behaviors was measured using the Negative Acts Questionnaire-Revised (NAQ-R), a questionnaire measuring frequency of exposure to 22 bullying behaviors (e.g. That gossip or rumors have been spread about you; Physical assault or threats of such assault), with response categories being presented using a five-point Likert-type scale ranging from (1) never, through (2) now and then, (3) monthly, (4) weekly to (5) daily. This variable was to be used as a control measure in order to ensure that the perpetrators’ own exposure to bullying behaviors did not affect the results, in case self-labeled bullies were part of an ongoing conflict at T1. Internal reliability was acceptable for this scale with a Cronbach’s alpha of 0.9.

Statistical analyses were performed using the statistical package IBM SPSS Statistics 21. In order to test the relationship between being a perpetrator of bullying and the different indicators of expulsion in working life, logistic regression analyses were conducted. Intention to leave was used as an outcome only at T1. Change of employer, disability benefit recipiency and unemployment were used as outcomes at T2 and T3 in prospective analyses, while high sickness absence was used as an outcome at all three time-points because both the cross-sectional and the prospective associations between bullying and high sickness absence were seen as relevant to the research question. All analyses were conducted twice, first controlling only for age and gender, and secondly controlling also for one’s own exposure to bullying behaviors>. For the outcome measures, the reference categories used were “no intention to leave”, “less than 24 sick days”, “no change of employer” and “full-time employed” (for the disability benefit recipiency and unemployment measures), respectively. The level of significance was set to 0.05 for all analyses.

Descriptive statistics revealed that 2.8% of the respondents answered affirmatively to being a perpetrator of bullying (n=68). Of these, 19.7% reported having responsibilities as a manager (n=13, Missing=3%), and 35.3% were women (n=24).

The results of cross-sectional logistic regression analyses revealed that being a perpetrator of bullying did not significantly increase the probability of reporting intention to leave at T1, neither when controlling only for age and gender (OR=1.29, p>0.05), nor when adding exposure to bullying behaviors to the control measures (OR=0.69, p>0.05). Similarly, being a perpetrator of bullying did not significantly increase the probability of reporting more than 24 sick days at T1 when controlling for age and gender (OR=1.02, p>0.05) exclusively. When adding exposure to bullying behaviors to the control measures, however, being a perpetrator of bullying was associated with a significantly lower probability of reporting more than 24 sick days at T1 (OR=0.27, p<0.05), indicating that perpetrators in fact display a lower probability than non-perpetrators for high levels of sickness absence. The results of the cross-sectional analyses are listed in Table 1.

Table 1.

Logistic regression analyses with being a perpetrator of bullying as a predictor of outcomes associated with expulsion in working life

					Controlling for age and gender				Controlling for age, genderand exposure to bullying behaviors
		N	%		OR	95 % CI	Sig.		OR	95 % CI	Sig.
Cross-sectional analyses
	Intention to leave	1,087	45.9		1.29	0.75–2.21	Ns.		0.69	0.37–1.26	Ns.
	More than 24 sick days	240	10.2		1.02	0.43–2.39	Ns.		0.27	0.08–0.99	*
Prospective analyses
	More than 24 sick days T2	158	10.3		2.45	1.03–6.06	*		2.01	0.75–5.85	Ns.
	More than 24 sick days T3	138	10.6		.56	0.13–2.38	Ns.		0.51	0.11–2.31	Ns.
	Change of employer T2	234	14.9		1.04	0.76–1.38	Ns.		0.86	0.33–2.24	Ns.
	Change of employer T3	455	32.9		0.84	0.41–1.72	Ns.		0.91	0.42–1.97	Ns.
	Unemployment T2	20	0.8		0	-	Ns.		0	-	Ns.
	Unemployment T3	17	1.1		2.06	0.26–16.61	Ns.		1.45	0.16–13.45	Ns.
	Disability benefit recipiency T2	42	1.7		3.11	0.66–14.58	Ns.		0.83	0.08–9.19	Ns.
	Disability benefit recipiency T3	71	4.5		4.16	1.45–11.92	**		2.66	0.66–10.73	Ns.

*p<0.05, **p<0.01, Ns.: Non-significant, OR: odds ratio, 95% CI: 95% confidence interval. In the prospective analyses predicting Sick Days >24 d, T1-levels of the same variable was added to the control variables

The results of the first of the prospective logistic regression analyses revealed that being a perpetrator of bullying significantly increased the probability of reporting more than 24 sick days at T2, when controlling for age, gender and having more than 24 sick days at T1 (OR=2.45, p<0.05). However, when adding exposure to bullying behaviors, the association became non-significant (OR=2.01, p>0.05). Moreover, being a perpetrator of bullying did not significantly increase the probability of reporting more than 24 sick days at T3, neither when controlling for age, gender and having more than 24 sick days at T1 (OR=0.56, p>0.05) exclusively, nor after adding exposure to bullying behaviors to the control measures (OR=0.51, p>0.05).

Concerning change of employer, being a perpetrator of bullying did not significantly increase the probability of change of employer at T2, neither when controlling for age and gender (OR=1.04, p>0.05) exclusively, nor after adding exposure to bullying behaviors to the control measures (OR=0.86, p>0.05). Similarly, being a perpetrator of bullying did not significantly increase the probability of change of employer at T3, neither when controlling for age and gender (OR=0.84, p>0.05) exclusively, nor after adding exposure to bullying behaviors to the control measures (OR=0.91, p>0.05).

Concerning unemployment, the results showed that being a perpetrator of bullying did not significantly increase the probability of unemployment at T2, neither when controlling for age and gender (OR=0, p=1) exclusively, nor after adding exposure to bullying behaviors to the control measures (OR=0, p=1). Similarly, being a perpetrator of bullying did not significantly increase the probability of unemployment at T3, neither when controlling for age and gender (OR=2.06, p>0.05) exclusively, nor after adding exposure to bullying behaviors to the control measures (OR=1.45, p>0.05).

Lastly, concerning disability benefit recipiency, the results showed that perpetrators of bullying did not have a significant increase in the probability of disability benefit recipiency at T2, neither when controlling for age and gender (OR=3.11, p>0.05) exclusively, nor after adding exposure to bullying behaviors to the control measures (OR=0.83, p>0.05). Being a perpetrator of bullying did however significantly increase the probability of disability benefit recipiency at T3 when controlling for age and gender (OR=4.16, p<0.05), but the association became non-significant after adding exposure to bullying behaviors to the control measures (OR=2.66, p>0.05). The results of the prospective analyses are listed in Table 1.

Together, the results of both the cross-sectional and the prospective analyses in the present study strongly support the notion that perpetrators of workplace bullying do not face a significant risk of expulsion in working life. Employing the same dataset and variables, earlier studies have shown that the targets of bullying do have such an elevated risk of expulsion1,2,3). One possible explanation to the null-findings shown here may be found in Leymann’s original notion that the perpetrator in many instances will form an alliance with top management7). However, there are also other possible explanations. For instance, in spite of Leymann’s notion that management tends to intervene at some point7), there may be instances of bullying that the top management remains unaware of, as targets may go on sick-leave, take early retirement or plainly quit their job without ever raising their concern or handing in a formal complaint. In support of this notion, Rayner13) showed that non-targets would have expected a more assertive response if they had been bullied, than the degree of assertiveness displayed by actual targets. Also, researchers have shown that raising your voice against interpersonal mistreatment increases the risk of retaliation, indicating that a fear of retaliation can be an incentive not to speak up about being a target of bullying14). Thus, a potential target expulsion phase may be reached without the bully ever being faced with potential sanctions. Also, it is not known whether the self-reported bullies in the present study differ from bullies in general with respect to their perception and reporting of bullying. For example, previously sanctioned bullies may be less likely to participate in the survey, or, in order to avoid self-inconsistent dissonance15), to be less willing to label their behavior as bullying, for instance after having pleaded their case in an internal investigation or during a legal process. Thus, the original notion of Leymann concerning a potential alliance between the perpetrator and upper management is one of many liable explanations of the present findings.

While the present study has the clear benefit of employing prospective and nationally representative data, a few potential limitations should be mentioned. Firstly, the exclusive use of self-report measures can be problematic, even though the use of time-lags as well as concrete outcomes such as working life statuses can protect against the risk of common-method biases16). In addition, the relatively low prevalence of self-labeled bullies can possibly be problematic with regard to the statistical power of the analyses. Thus, further research efforts might be needed in order to more firmly establish the tendency demonstrated in the present study.

The European framework agreement on harassment and violence at work signed by most EU Countries in 2007 makes it explicit that being a perpetrator of bullying should result in “disciplinary actions up to and including dismissals”9). In many countries, including Norway8), bullying is also illegal, and it is widely held as unjustifiable employee conduct10), but despite the good intentions of legislators, courts of law, organizations and the general public, the present results indicate that there are comparatively few practical mechanisms at play to actually implement sanctions against perpetrators of bullying. Thus, our findings are in contrast to what one might expect from a juridical and an ethical standpoint, and should have implications for practice in organizations. For instance, anti-bullying policies involving a clear declaration of the organizations’ attitudes towards bullying, descriptions of how to notify management if bullying should occur as well as a clear description of potential sanctions could both prevent new cases and facilitate a just process in the aftermath of bullying. The findings should also have implications for research. To the best of our knowledge, no studies have investigated the job-related consequences of being a bully, and further empirical work is needed, for instance in the form of research interviews with managers about their knowledge of and response to bullying in their own organizations, in order to obtain a better understanding of why many bullies appear not to be faced with the same risk of job-related sanctions as targets of bullying. Such an understanding may have yet new practical and theoretical implications, both in terms of laying the groundwork for proper responses to bullying in organizations, and in terms of shedding light on and preventing the process by which targets of bullying experience expulsion in working life.