Workplace mobbing: How the victim’s coping behavior influences bystander responses

ABSTRACT Victims of workplace mobbing show diverse coping behavior. We investigated the impact of this behavior on bystander cognitions, emotions, and helping toward the victim, integrating coping literature with attribution theory. Adult part-time university students (N = 161) working at various organizations participated in a study with a 3(Coping: approach/avoidance/neutral) × 2(Gender Victim: male/female) × 2(Gender Bystander: male/female) design. Victims showing approach (vs. avoidance) coping were considered to be more self-reliant and less responsible for the continuation of the mobbing, and they elicited less anger. Continuation responsibility and self-reliance mediated the relationship between the victim’s coping behavior and bystanders’ helping intentions. Female (vs. male) participants reported more sympathy for the victim and greater willingness to help, and female (vs. male) victims elicited less anger. Theoretical and practical implications of the findings are discussed.

had been subjected to both mobbing and more direct forms of violence, 5% of their sample had been subjected to severe mobbing, and 12% to occasional mobbing.
Given the costs and prevalence associated with workplace mobbing, it is important to investigate factors either contributing to or inhibiting the mobbing process. From a social interactionist perspective, bystanders are actors in the process of workplace mobbing (Namie & Lutgen-Sandvik, 2010), although their role in this process is still unclear (Keashly & Jagatic, 2011). Bystanders can take various positions in the mobbing process, ranging from helping the bully to helping the victim (Paull et al., 2012).
Research that focuses on the antecedents of bystander behavior in workplace mobbing is still scarce and is mainly based on qualitative studies (e.g., D'Cruz & Noronha, 2011;Lutgen-Sandvik & Fletcher, 2013), where experimental studies can enhance theory development in this area (Neall & Tuckey, 2014). One recent study indicates that factors such as bystanders' perceptions of the victim's responsibility for the onset of the mobbing, bystanders' anticipations of stigma by association, and bystanders' gender might all influence their response to workplace mobbing (Mulder, Pouwelse, Lodewijkx, & Bolman, 2014). Yet there is little information about how the victim's response to the mobbing can influence bystander behavior. Therefore, the aim of the present study is to investigate the influence of the victim's coping behavior on bystanders' cognitive, emotional, and behavioral reactions toward the victim. For this purpose, we integrate theory on coping with the attributionemotion model of stigmatization (Dijker & Koomen, 2003;Weiner, Perry, & Magnusson, 1988). Additionally, we examine gender effects based on the social role theory of gender differences (Eagly & Crowley, 1986).

Coping
Coping can be defined as "behaviors used to manage the internal and external demands of situations that are appraised as stressful" (Folkman & Moskowitz, 2004, p. 745). It is often construed as either approachoriented-i.e., consisting of cognitive and behavioral efforts to address the problem-or avoidanceoriented-i.e., consisting of attempts to evade the problem (Carver, 2007;Ebata & Moos, 1991).
Coping has been found to have an impact on social support (Dunkel-Schetter, Folkman, & Lazarus, 1987). It has been shown that when observers believe individuals to be coping poorly, as compared to coping well, they experience more discomfort and anger, and have less sympathy for the individual. Moreover, they intend to provide less social support to the individual (Bos, Dijker, & Koomen, 2007;Schreurs & De Ridder, 1997;Schwarzer & Weiner, 1991;Silver, Wortman, & Crofton, 1990). Furthermore, people estimate the victim's chance for improving the situation as greater for victims who they perceive to be coping well, compared to those they perceive to be coping poorly (Schwarzer & Weiner, 1991). People may therefore fail to recognize the need to support victims who appear to be coping well, as these victims do not signal this need (Silver et al., 1990) and hence appear to be more self-reliant.
It is likely that the victim's coping behavior impacts bystanders' responses to the victim. As the literature shows, victims of workplace mobbing can cope with the mobbing in different ways (Zapf & Gross, 2001). On the one hand, victims can engage in avoidance coping. In doing so, they try to avoid the mobbing and move away from the problem by means of absenteeism, asking for transfer, or by leaving the organization. On the other hand, victims can engage in approach coping, whereby they try to solve the problem by confronting the bully and telling him/her to stop (Olafsson & Johannsdottir, 2004).
Although it has been shown that victims seldom confront the perpetrator (Salin, Tenhiälä, Roberge, & Berdahl, 2014), but instead tend to avoid the situation (Hogh & Dofradottir, 2001), approach coping is often considered the more appropriate coping response (Brodsky, 1976;Salin et al., 2014). Non-victims report that they would engage in approach coping themselves, if they were ever to be bullied (Rayner, 1997). Even victims regret not having been more confrontational (Salin et al., 2014). In a classic study carried out by Brodsky (1976), coworkers perceived victim avoidance behavior (such as asking for support or taking sick leave) as unfair, cheating, or malingering. In summary, people generally view approach coping to be an adequate reaction to workplace mobbing, and one that could lead to a decrease in the amount of mobbing. In contrast, avoidance coping is perceived as an inadequate way of coping that will not stop the mobbing.
This implies that victims are believed to be able to influence the mobbing process and can therefore be held responsible for the outcome. We therefore expect bystanders to react more positively to victims showing approach coping than to victims showing avoidance coping. To explain the underlying mechanism, we use the attribution-emotion model of stigmatization, which will be described in the following paragraph.

Attribution-emotion model of stigmatization
The attribution-emotion model of stigmatization (Dijker & Koomen, 2003;Weiner et al., 1988) proposes that the behavior of a stigmatized person influences the cognitive, emotional, and behavioral responses directed at this person. More specifically, when the stigmatized condition is perceived as controllable and reversible, people attribute responsibility to the stigmatized person. Consequently, high (vs. low) perceived responsibility will evoke more anger and less sympathy, and hence less intention to help. A distinction has been made between onset and offset responsibility (Brickman et al., 1982;Weiner et al., 1988). Onset responsibility refers to responsibility for creating the stigmatized condition, for example, becoming obese as a result of overeating. Offset responsibility refers to responsibility for the resolution or the continuation of the stigmatized condition, for example, dieting or exercising in order to lose weight. Offset responsibility, based on the perceptions of the way a person copes with the problem, is an important predictor of negative evaluations (Black, Sokol, & Vartanian, 2014), and of helping behavior (Karasawa, 1991;Schwarzer & Weiner, 1991). In this study, we are interested in bystanders' perceptions of the victim's responsibility for the continuation of the mobbing; therefore, we use the term continuation responsibility rather than the term offset responsibility.
By integrating the coping literature with the attribution-emotion model of stigmatization, the present study examines how the coping behavior of a victim of workplace mobbing relates to a bystander's cognitive, emotional, and behavioral reactions toward this victim.
In relation to cognition, we expect that victims showing approach (vs. avoidance) coping will be perceived to be less responsible for the continuation of the mobbing (hypothesis 1a). Furthermore, we expect that victims showing approach (vs. avoidance) coping will be perceived to be less dependent on help for resolving the situation-that is, they will be seen as more self-reliant (hypothesis 1b).
In relation to emotions, we expect victims showing approach (vs. avoidance) coping to arouse less anger (hypothesis 2a), and more sympathy (hypothesis 2b), in bystanders.
In relation to helping intention, we expect that continuation responsibility and perceived self-reliance will mediate the relationship between the victim's coping behavior and bystanders' helping intentions. We hypothesize that bystanders will attribute less continuation responsibility toward a victim showing approach (vs. avoidance) coping, and hence will feel more sympathy and less anger toward the victim. As a consequence, we expect these bystanders to show greater willingness to help the victim (hypothesis 3a). At the same time, we expect that approach (vs. avoidance) coping will decrease intended helping behavior as a consequence of the perceived self-reliance of the victim (hypothesis 3b).

Gender
In mobbing situations, gender is an important factor to study because it can affect the mobbing process in several ways (Salin, 2011;Salin & Hoel, 2013). For instance, compared to male victims, female victims of mobbing are more likely to use avoidance coping and to seek help, and are less likely to use assertive strategies (Olafsson & Johannsdottir, 2004). Furthermore, as compared to men, women tend to feel more harassed (Björkqvist, Österman, & Lagerspetz, 1994) and to rate mobbing behavior as more severe (Escartín, Salin, & Rodríguez-Carballeira, 2011). Women also tend to attribute less responsibility to the victim as compared to men (Salin, 2011).
The social role theory of gender differences (Eagly, 1987) may help to explain bystanders' responses to the victim. According to this theory, men are expected to be dominant, controlling and assertive, and women are expected to be subordinate, cooperative, compliant to social influences, and less overtly aggressive (Eagly, Wood, & Johannesen-Schmidt, 2004). Furthermore, as compared to men, women are expected to be more dependent on others (Bornstein, 1995;Lerner, 1983), and are perceived to be more vulnerable (Dijker, 2001). Empirical evidence supports these notions of gender role expectations. When confronted with people with HIV displaying distress, men reported feeling more anger, less sympathy, and were less likely to engage in prosocial behavior as compared to women (Bos et al., 2007). Furthermore, female (vs. male) victims received more sympathy and evoked a higher protective tendency. This effect was stronger for male (vs. female) perceivers (Dijker, 2001).
Based on these theoretical assumptions and empirical findings, we expect bystander gender, as well as victim gender, to affect the cognitions, emotions, and intended helping behaviors of bystanders. First, we theorize that bystanders' reactions toward the victim will depend on their expected gender role-i.e., that the gender of the bystanders will influence their cognitive, emotional, and behavioral reactions toward the victim. We expect that female bystanders will react in a more other-oriented, nurturing way than male bystanders. We hypothesize that they will attribute less continuation responsibility to the victim of workplace mobbing (hypothesis 4a), show more sympathy (hypothesis 4b), less anger (hypothesis 4c), and report a greater willingness to help the victim (hypothesis 4d) toward the victim.
Second, we propose that the victim's gender will influence the cognitive, emotional, and behavioral reactions of the bystanders. We expect female victims to evoke more supportive reactions than male victims. Specifically, as compared to male victims, we expect that female victims will be perceived to be more reliant on help in terms of solving the situation (hypothesis 5a) and evoke more sympathy (hypothesis 5b), less anger (hypothesis 5c), and more willingness to help the victim (hypothesis 5d).

Participants, design, and procedure
Participants were working adults studying part-time at the Open University of the Netherlands (N = 161, 39.8% male, mean age 40.81 years, SD = 10.92). In 2014, 16.888 students were registered at this University, which is organized into regional study centers in various parts of the Netherlands and Belgium. Participants were recruited from classes and through social media groups of the faculties of Psychology and Management. Participants were employed in a wide range of organizations with contract hours ranging from 3 to 40 hours per week.
A 3 (Coping: approach/avoidance/neutral) × 2 (Gender Victim: male/female) × 2 (Gender Bystander: male/female) randomized between-subjects design was used for this vignette study. Gender of participant was randomly assigned to six conditions (χ 2 (5, N = 161) = 6.03, p = .30). By means of an online survey, participants were first presented with one of six vignettes and then requested to answer questions about their cognitive, emotional, and behavioral responses toward the victim. Participants in this study acted as bystanders.

Vignettes
The vignettes first introduced an imaginary coworker (the victim, either female or male) who worked as an administrative assistant in the same department as the participant. The vignettes then described the ongoing negative behavior of another coworker (the perpetrator) toward this first coworker. This behavior included interrupting, belittling, ignoring the victim during informal gatherings, gossiping, and exclusion from conversations and collaborations. The perpetrator's gender was not disclosed. Next, the vignettes presented the victim's coping strategy as either (1) approach coping (telling the bully to stop this behavior, not allowing the bully to interrupt, and asking the perpetrator to tell things to his/her face instead of gossiping); (2) avoidance coping (avoiding the bully, taking sick leave and asking for a transfer); or (3) neutral (the vignette did not provide information on the victim's coping strategy).

Measures
Ratings for all variables ranged from 1 (= "not at all") to 7 (= "very much"). Composite scales were created by aggregating the items and averaging the scores.

Manipulation checks
We measured the effectiveness of the manipulation of approach coping with the item "To what extent does <male name/female name> stand up for her/himself?" and avoidance coping with "To what extent does <male name/female name> try to avoid the situation?"

Dependent variables
Cognitions Two cognitive variables were measured.
Perceived responsibility. The victim's perceived responsibility for the continuation of the mobbing was measured with three items (α = .73): "Is <male/female> responsible for the continuation of the negative treatment by his/her coworker?"; "Does the way <male/female> reacts to the situation influence its continuation?"; and "Do you think that the way <male/female> handles the situation contributes to its continuation?" These items, adapted from those used by Steins and Weiner (1999), were rephrased in order to correspond to the mobbing situation. Self-reliance. Perceived self-reliance of the victim was measured with two items (α = .67): "I expect that <male/female> can look after him/herself in this situation" and "I expect that <male/female> needs help from others" (reversed item).

Emotions
Furthermore, two emotions were assessed based on items used by Struthers, Weiner, and Allred (1998).
Sympathy. Sympathy was measured with three items on how much pity, compassion and concern the participant felt toward the victim (α = .66).
Anger. Anger was measured with three items asking how annoyed, angered and irritated the participant felt toward the victim (α = .84).

Helping Intentions
Finally, helping intention was measured with three items (α = .87), assessing the probability, willingness, and confidence of the participant helping the victim (cf. Greitemeyer & Rudolph, 2003).

Analyses
Hypotheses 1, 2, 4, and 5 were tested with (M)ANOVAs. For hypothesis 3, we used the PROCESS macro v2.13, model 6 for serial multiple mediation testing (Hayes, 2012), applying a method suitable for a multicategorical independent variable (see Hayes & Preacher, 2014) with 1000 bootstrap samples and 95% CI. This method yields unstandardized indirect and direct effects of one group relative to one or more reference groups. For the (M)ANOVAs, the experimental conditions were dummy coded as follows: coping (1 = neutral, 2 = avoidance, and 3 = approach), gender (0 = male, 1 = female). For the mediation analyses, the coping conditions were dummy coded with the approach coping condition serving as the reference group (see Hayes & Preacher, 2014).

Manipulation check
A one-way MANOVA on approach coping and avoidance coping showed a significant effect of the coping conditions (approach, avoidance and neutral), F(4,316) = 95.42, p < .001. Univariate analysis showed that the coping conditions had a significant effect on perceived approach coping, F (2,158) = 337.38, p < .001, η p 2 = .81, and on perceived avoidance coping, F(2,158) = 83.58, p < .001, η p 2 = .51. Participants in the approach condition rated the victim's reaction to the mobbing as more consistent with approach coping than those in the neutral (M = 6.15, SD = .66 and M = 2.60, SD = 1.04, respectively; d = 4.09) or avoidance condition (M = 2.07, SD = .98; d = 4.88). Likewise, participants in the avoidance condition rated the victim's reaction as more consistent with avoidance coping (M = 6.09, SD = 1.35) than those in the neutral (M = 3.66, SD = 1.20; d = 1.90) or approach condition (M = 2.66, SD = 1.53; d = 2.38). LSD posthoc tests showed that the mean differences between the conditions were all significant (p < .01). These results suggest that the manipulations were successful.
There were no other main or interaction effects for either the continuation responsibility variable or the self-reliance variable. This implies that, in contrast to hypothesis 4a, bystander gender did not influence perceived continuation responsibility of the victim. Female participants did not attribute less continuation responsibility to the victim of workplace mobbing. Furthermore, in contrast to hypothesis 5a, victim gender did not affect the perceived self-reliance of the victim. Female victims were not perceived to be more reliant on help in order to resolve the situation.
The mediation analyses showed the two hypothesized indirect effects of the approach coping conditions through the mediators continuation responsibility and self-reliance. Confirming hypothesis 3a, there was an indirect positive effect on helping behavior in the approach coping condition relative to the avoidance coping and neutral conditions (B = 0.07, SE = 0.04, 95% CI = 0.18, 0.12 and B = 0.04, SE = 0.02, 95% CI = 0.12, 0.01, respectively). Compared to participants in the avoidance and neutral conditions, participants in the approach coping condition reported less perceived continuation responsibility, which increased sympathy and hence intended helping behavior toward the victim (see Figure 1). In contrast to hypothesis 3a, anger did not serve as a mediator in this particular sequence.
In line with hypothesis 3b, the approach (vs. avoidance and neutral) conditions had an indirect negative effect on helping intention through self-reliance (B = −0.19, SE = 0.09, 95% CI = −0.04, −0.43 and B = −0.11, SE = 0.06, 95% CI = −0.02, −0.27, respectively). Relative to participants in the avoidance coping and neutral conditions, participants in the approach coping conditions reported more perceived self-reliance and hence, less helping intentions toward the victim (see Figure 2).

Discussion
Workplace mobbing has damaging consequences for the victim. It is therefore important to investigate how workplace mobbing can be tackled and reduced. Bystanders of workplace mobbing can potentially play an important role in terms of stopping bullying and helping victims of workplace mobbing (Leymann, 1996;Lutgen-Sandvik & Fletcher, 2013). As outlined in the literature, however, bystander behavior can range from prosocial to antisocial behavior toward the victim (Paull et al., 2012). To enhance the likelihood of prosocial behavior, it is first necessary to gain insight into the antecedents of this behavior. This insight can, in turn, inform research-based interventions aimed at changing bystander behavior. However, research exploring the antecedents of bystander behavior is scarce. Therefore, the main objective of this study was to examine how the coping behavior of a victim of workplace mobbing influences bystanders' cognitive, emotional, and behavioral responses toward this victim. By demonstrating that the victim's coping behavior influences both bystanders' perceptions of the victim's responsibility and self-reliance, and consequently bystanders' emotions and intended helping behavior toward the victim, our findings show how important it is to take coping behavior into account.
The findings showed two effects of the victim's coping behavior on bystanders' cognitive reactions. First, participants attributed less responsibility for the continuation of the mobbing to a victim showing approach (vs. avoidance) coping behavior (cf. Karasawa, 1991;Schwarzer & Weiner, 1991). This is in line with earlier findings that suggest that non-victims expect victims to react in a more confrontational manner than they actually do (Rayner, 1997). Second, participants perceived the victim showing approach (vs. avoidance) coping behavior to be more self-reliant in handling the situation. This finding is consistent with the suggestion that individuals who cope well at the same time fail to indicate that they need help (cf. Silver et al., 1990, p. 403), and with earlier findings regarding the expectation that individuals who use constructive coping strategies have a higher chance of obtaining a positive outcome (Schwarzer & Weiner, 1991). Together, these findings suggest that bystanders consider approach coping to be more appropriate and effective than avoidance coping.  Figure 1. Model outlining relationships between the victim's coping behavior, bystanders' emotions, and intended helping behavior, with continuation responsibility as mediator. Note. Unstandardized regression weights. D1 = effect of the approach coping condition relative to the avoidance coping condition, with D2 as control variable. D2 = effect of the approach condition relative to the neutral coping condition, with D1 as control variable. Self-reliance served as control variable. *p < .05; **p < .01; ***p < .001; ns = non significance.  Figure 2. Model outlining relationships between the victim's coping behavior, bystanders' emotions, and intended helping, with self-reliance as mediator. Note. Unstandardized regression weights. D1 = effect of the approach coping condition relative to the avoidance coping condition, with D2 as control variable. D2 = effect of the approach condition relative to the neutral coping condition, with D1 as control variable. Continuation responsibility served as control variable. *p < .05; **p < .01; ***p < .001; ns = non significance.
Next, the results regarding the emotional responses of bystanders showed that the victim's coping behavior affected bystanders' feeling of anger toward the victim. More specifically, results indicated that bystanders were less angry with a victim engaging in approach coping behaviors than with a victim engaging in avoiding coping behaviors. This is in line with earlier research suggesting that conveying helplessness results in criticism (Schreurs & De Ridder, 1997) and avoiding a difficult situation and not taking responsibility for it elicits anger (Brodsky, 1976;Schwarzer & Weiner, 1991). Coping behavior did not directly influence the level of sympathy with the victim in the present study; bystanders appear not to feel more sympathy for a victim showing approach coping than for a victim showing avoiding coping. This is in contrast with the findings of Schwarzer and Weiner (1991), who investigated the effect of coping behavior on sympathy. A possible explanation for the lack of effect of the coping conditions on sympathy in the present study might be that victims showing approach coping may not have appeared to be particularly vulnerable or suffering, and did not therefore arouse sympathy (cf. Dijker, 2001), whereas victims showing avoidance coping did not elicit bystanders' sympathy because they were not seen to be coping appropriately with the mobbing (Schwarzer & Weiner, 1991).
Although the victim's coping behavior did not influence helping intention directly, the mediation analyses showed two expected and opposite indirect effects. These two opposing effects may have cancelled each other out, with the net result being that bystander helping intention was not affected. However, these two effects are, in themselves, worth a closer look. First, the findings indicate that bystanders consider a victim showing approach (vs. avoidance) coping to be less responsible for the continuation of the mobbing, and are therefore more inclined to help this victim (Schwarzer & Weiner, 1991;Weiner et al., 1988). This is in line with suggestions that bystander behavior may depend in part on the perceived appropriateness of the victim's coping behavior (Baillien, Neyens, De Witte, & De Cuyper, 2009;Einarsen & Mikkelsen, 2003, p. 137). Second, the findings indicate that bystanders perceive a victim showing approach (vs. avoidance) coping to be more self-reliant, and are therefore less inclined to help this victim. This suggests that bystanders do not provide support to a victim showing approach coping (as operationalized in the present study by confronting the perpetrator) because they expect the victim to be able to handle the situation adequately themselves (cf. Silver et al., 1990).
In sum, the findings suggest that bystanders react more positively to victims of workplace mobbing who are showing approach coping as compared to avoidance coping. Furthermore, they perceive victims showing approach coping to be able to handle the situation themselves. This last perception, however, may be wrong; approach coping does not necessarily resolve the mobbing situation. It may even enhance it and leave the victim showing approach coping in an even worse situation (Zapf & Gross, 2001), because confrontation often leads to retaliation (Cortina & Magley, 2003). As such, bystanders might fail to understand the position of the victim who has to choose between two bad options, namely approach coping, which may lead to the perpetrator retaliating and to scant help owing to bystanders' perception of the victim's self-reliance, or avoidance coping, which may not solve the problem and can lead to negative reactions from bystanders.
Our findings further suggest that bystander as well as victim gender affects bystanders' reactions toward victims of workplace mobbing. In relation to bystander gender, in accordance with social role theory of gender (Eagly, 1987), and findings in stigma research (Bos et al., 2007), female (as compared to male) bystanders reported more sympathy for the victim. Furthermore, the data suggested with a small effect that female bystanders are more willing to help the victim.
Our findings regarding the victim's gender and helping also suggest that male and not female victims elicit more helping intention. This effect was only found in our female participants group. Although the effect was small, we find it interesting enough to pursue. These results may be explained in several ways. First, gender may have operated as a status cue, as it does in organizations (Eagly & Crowley, 1986). This would imply that men are perceived to have a higher status than women by default and are entitled to more help from the gender with the lower status, i.e. women (Eagly & Crowley, 1986). Second, female participants may have felt more capable in situations in which a man needs help in stereotypically feminine situations-i.e., gender atypical situations, and hence tend to help more in these situations (Fiala, Giuliano, Remlinger, & Braithwaite, 1999). This would suggest that female bystanders perceive mobbing victimization to be a typically feminine position. This assumption may also help to explain the finding that male victims elicited more anger that female victims. In line with gender role expectations (Eagly, 1987), male victims should have avoided getting into a negative "feminine" position. However, these conclusions are only speculative, as we did not measure gender labeling in relation to mobbing victimization.
Finally, female victims did not evoke more sympathy than male victims, in contrast to the findings of Dijker (2001). It is possible that participants in the present study perceived the victim's suffering to be low; whereas, according to Dijker (2001), perceived vulnerability needs to be high in order to arouse sympathy. This would imply that the female victim did not appear more vulnerable than the male victim.

Limitations and future research
Although the present study provides more insight into the relationship between the victim's coping behavior and bystanders' helping intention, some limitations need to be acknowledged. First, our sample consisted of highly educated participants, namely employees who are (former) students of a university. Replication of the present study with lower-educated participants would extend the generalizability of the results.
In a similar vein, the use of vignettes might limit the generalizability of our findings. Vignettes can differ from reality in that they may not convey certain "real world" influences, such as participants' experiences with incidents of mobbing at the workplace, including related thoughts and feelings toward a particular victim, or the social climate participants experience in their workplace. These influences may unintentionally conceal effects researchers are interested in. However, vignette studies allow researchers to manipulate and control independent variables. Furthermore, compared to cross-sectional studies, vignette studies have the advantage of providing insight into causal relationships (Aguinis & Bradley, 2014) and are therefore considered crucial for enhancing our knowledge about antecedents of bystanders' behavior. Finally, we took care to enhance the reality of a workplace setting by showing participants a photograph of a workplace before presenting the vignettes (see for instance, Dijker, Koomen, & Kok, 1997).
Furthermore, in the present study, the perpetrator's gender was not disclosed. Perpetrator gender might serve as an additional antecedent of bystanders' behavior toward a victim, either directly or in interaction with the other factors that we studied. For instance, male bystanders might consider the behavior of a male perpetrator as appropriate owing to perceived similarity between himself and the perpetrator (Guèguen, Martin, & Meineri, 2011;Salin, 2011).
Finally, one might argue that a higher level of power could possibly reveal interactions that were not found in the present study (Simmons, Nelson, & Simonsohn, 2011). However a power analysis indicated that the power of this study is sufficient.
The findings of this study provide interesting avenues for further research on antecedents of bystander behavior in workplace mobbing. First, it would be interesting to study whether specific types of victim coping behavior might elicit specific types of bystander support. Earlier research on coping in general indicates that different types of coping behavior may result in different types of support, such as informational support, assistance and emotional support (Dunkel-Schetter et al., 1987). Furthermore, our findings point toward the possibility that victimization of mobbing has a feminine connotation. Further research might investigate how bystanders perceive the position of a victim of mobbing in terms of gender roles (Eagly & Crowley, 1986), and the consequences of this perception for the mobbing process. Also in relation to gender influences, the perpetrator's gender should be taken into account in studies investigating gender effects on bystander behavior toward victims of workplace mobbing.
Finally, we noted that bystanders may expect the victim to use approach coping even though victims generally tend to use avoidance coping (cf. Rayner, 1997). This raises the question of whether bystanders are aware of the general course of the mobbing process, and of the consequences of the victim's behavior, as reported in previous research (e.g., Zapf & Gross, 2001). Consequently, it may be worthwhile studying bystander reactions toward the victim both before and after raising awareness of these consequences (see Strandmark & Rahm, 2014).

Practical implications
The results of the current study led to suggestions for victims on their possible behavior in response to workplace mobbing. First, approach behavior appears to elicit the most positive reaction from bystanders. An implication for victims is that they should reflect on their own response related to the mobbing. A coach, friends, family and even trusted bystanders can be asked for help in this area. Furthermore, even when victims approach the mobbing problem instead of avoid it, they should explicitly express their need for support, this way helping the bystander to become aware of this need. Given this need for support, it is recommended that organizations develop an appropriate policy concerning workplace bullying.
Workplace mobbing is a serious problem for victims and for organizations Hogh et al., 2011), and solutions therefore need to be found. Bystander intervention can be part of that solution. Interventions are preferably theory-and evidence-based, to ensure that they focus on the relevant determinants of behavior (see Bartholomew, Parcel, Kok, Gottlieb, & Fernandez, 2011;Bos, Pryor, Reeder, & Stutterheim, 2013). In line with this, the present study contributes to the surprisingly small body of knowledge on the antecedents of bystander responses to workplace mobbing.

Notes on contributors
Roelie Mulder is a researcher and the Head of the Staff Department at an environmental service for regional governments. Her research interest is workplace mobbing with a focus on bystander behavior. Arjan E. R. Bos is an Associate Professor of Clinical Psychology. His main research interest is social stigma. His research primarily focuses on the psychological impact of stigmatization and on stigma reduction. Mieneke Pouwelse is an Assistant Professor of Work and Organizational Psychology. She has published on school and workplace bullying. Karen van Dam is a Full Professor of Work and Organizational Psychology. Her research interests include adaptation to change, emotion regulation, sustainable careers, and employee well-being.