Abstract
The aim of this article is to examine target responses to workplace mistreatment and to analyze factors that affect the degree of discrepancy between actual and ‘ideal’ (i.e. desired) responses. Two-hundred and seventeen faculty members at a major research university in North America reported their actual and ideal responses to mistreatment. The most common responses involved passive and social support-seeking strategies. Respondents generally wished they could have been more assertive. The size of the discrepancy between actual and ideal responses to mistreatment was predicted by the perceived severity of the behavior, the coping strategy chosen and a difference in organizational status and gender between the perpetrator and the target of mistreatment. While our findings show that status differences were associated with a larger discrepancy regardless of the direction of the status differences, our results indicate that the mechanisms behind the discrepancy differed. Despite being a relatively high status population, faculty at a prestigious university responded more passively to mistreatment than desired, primarily due to situational constraints. Because the reasons for this discrepancy were often structural (i.e. based on organizational or social status structures), this research highlights the need for organizations to address mistreatment proactively, even in the absence of formal complaints.
I would like to have used the [University Faculty Association] as an advocate but [it] had previously warned me that raising the issue and being an untenured faculty member without a PhD would be ‘career suicide’. (Female Faculty Member, 30+)
During the past few decades, we have seen an increased scholarly interest in the phenomenon of workplace mistreatment and how targets cope with it. Research has shown that mistreatment has negative consequences for targets, such as reducing job satisfaction and organizational commitment and increasing psychological distress and physical health problems (Cortina et al., 2001; Fox and Stallworth, 2010; Lim et al., 2008; Pearson et al., 2001; Salin, 2013).
Workplace mistreatment has been studied using a plethora of different terms, such as victimization, incivility, bullying, harassment, emotional abuse and social undermining, just to mention a few (e.g. Hershcovis, 2011). In this article we use the term mistreatment to cover a broad range of behavior, from milder forms of incivility to blatant harassment and violence (Cortina and Magley, 2003). We define mistreatment as ‘offensive behavior that is unsolicited and unwelcome’ and which ‘violates . . . a fundamental right to respectful treatment’(Harlos, 2010: 312). We focus on personalized forms of mistreatment, with a clearly distinguishable perpetrator, rather than more depersonalized forms of injustice, such as policy-related mistreatment (see Boswell and Olson-Buchanan, 2004; Klaas and DeNisi, 1989; Olson-Buchanan and Boswell, 2008).
Research has indicated that many employees stay silent about mistreatment and try to ignore, deny or minimize the experience, that is, use passive strategies rather than active ones that could actually lead to changes in the perpetrator’s behavior (Bies et al., 1997; Cortina et al., 2002). Research on workplace bullying shows a discrepancy between how actual victims of bullying respond to mistreatment and how non-victims believe they would act (e.g. Rayner, 1997). Non-victims typically believe they would respond actively and confront the perpetrator more often than victims actually do (e.g. Rayner, 1997). While we cannot rule out personality differences between targets and non-targets, it is likely that target responses are affected by situational constraints. This is supported by the finding that mistreated employees often face a Catch-22 dilemma (Cortina and Magley, 2003). Employees who voice and confront perpetrators, or who seek social or organizational support, are often subject to retaliation from other organizational members. Consistently, several studies have shown that employees who use grievance systems may be punished for doing so, such as in the form of lower performance ratings (Klaas and DeNisi, 1989; Lewin, 1987). On the other hand, Cortina and Magley (2003) showed that employees who endured mistreatment in silence were the ones who reported the highest levels of psychological and physical damage. Thus, both speaking up and refraining from speaking up are associated with high risks and costs.
The aim of this study is to analyze how targets of workplace mistreatment responded to the mistreatment and, in retrospect, how they wished they would or could have responded. We show that a discrepancy between actual and desired responses typically involved a desire for more assertive responses. Further, we examine factors that might explain a discrepancy between actual and ideal responses: the severity of mistreatment, coping strategy used and differences between the target and perpetrator in social and organizational status. We contribute to the growing literature on how social status affects experiences of, and responses to, mistreatment and victimization (e.g. Berdahl et al., 1996; Cortina, 2008; Harlos, 2010; Lamertz and Aquino, 2004).
Target responses to mistreatment
As scholarly interest in workplace mistreatment has increased, attention has shifted from a focus on physical aggression and sexual harassment toward more generalized forms of mistreatment and incivility (see Berdahl, 2007; Cortina, 2008; Einarsen et al., 2011; Salin, 2003). More generalized forms of mistreatment and incivility are typically of a psychological or verbal nature, and though some examples of individual acts may not seem very severe (e.g. being ‘talked down to’ or ignored), the negative effects of repeated accumulated experiences may be considerable (see Cortina et al., 2001; Pearson et al., 2001; Salin, 2013).
Large surveys indicate that a majority of employees encounter occasional negative acts in their places of work (e.g. Fox and Stallworth, 2005). Several studies point to the importance of social status in the mistreatment process. First of all, social status may affect the risk of being subjected to mistreatment in the first place. For instance, studies show particularly high exposure rates to mistreatment for ethnic minority members (Lewis and Gunn, 2007) and employees of an underrepresented sex within a certain occupation (e.g. Salin and Hoel, 2013).
Moreover, an individual’s location in the social status structure has a significant effect on how that individual appraises a certain act (e.g. Berdahl and Moore, 2006; Berdahl et al., 1996; Carli, 1999; Cortina and Magley, 2009; Lazarus and Folkman, 1984). Having more power will be associated with a greater sense of control over the situation and a high-power individual is more likely than a low-power individual to appraise a stressor as manageable. As such, we can expect different facets of social status – for example, relative organizational status, gender and ethnicity – to influence how mistreatment is construed and appraised. Studies on workplace bullying confirm that gender affects labeling of negative interpersonal behaviors and perceived severity (e.g. Berdahl, 2007; Salin and Hoel, 2013). Similarly, previous research on mistreatment shows that organizational status affects appraisal (Cortina and Magley, 2009). Appraisal, in turn, affects target responses and coping.
Targets can respond to mistreatment in many different ways. In their seminal work on employee responses to sexual harassment, Knapp et al. (1997) proposed four categories of response: (1) conflict avoidance/denial, (2) confrontation, (3) social support and (4) advocacy seeking. Their categorization is based on the idea that responses vary with respect to both focus of response and mode of response. Conflict avoidance, denial and social support are primarily focused on the self, helping the target cope with the situation without addressing the perpetrator or actual event. In contrast, confrontation and advocacy seeking look to actively address (and change) the perpetrator’s behavior. The mode of response reflects the amount of outside support: both social support and advocacy seeking are examples of supported responses, whereas conflict avoidance/denial and confrontation could be carried out by the target. Although empirical studies have generally supported this model, cluster analyses have indicated that denial and avoidance should preferably be treated as separate categories rather than as one (Wasti and Cortina, 2002).
Many factors affect target responses to mistreatment, including organizational status, gender, work-related self-esteem and appraisal (Cortina and Magley, 2009; Harlos, 2010). Research on power shows that being high in power reduces concerns about retaliation, while being low in power promotes anxiety and concern about punishment (Anderson and Galinsky, 2006; Keltner et al., 2003). High power individuals are likely to openly express their feelings and opinions, while low power individuals are likely to keep their feelings and opinions to themselves (Anderson and Berdahl, 2002; Berdahl and Martorana, 2006). Individuals low in power are therefore likely to refrain from choosing responses that are perceived as assertive and consequently risky. In line with the above, we can therefore assume that power or social status also predicts likely responses to mistreatment. Empirical evidence seems to support this assumption: Meares et al. (2004) analyzed interview accounts of mistreatment and found that feeling in control enabled more active and constructive responses. They further showed that feeling in control was intimately tied to a high level of social status. Employees who use active strategies and use voice risk retaliation from superiors and other organizational members (Bergman et al., 2002; Cortina and Magley, 2003; Klaas and DeNisi, 1989). The risk of retaliation increases with lower status and being subject to frequent mistreatment (Cortina and Magley, 2003).
However, some studies question the assumption that high power is always associated with more risk-taking and draw a more complex picture. Jordan et al. (2011) argue that the amount of power and stability of that power interact to affect risk-taking. They showed that both unstable power and stable powerlessness were associated with increased stress and risk-taking. Related to responses to mistreatment, this could mean that low-power targets may, under certain circumstances, contemplate more assertive responses than those with higher levels of power.
Based on all the above, we argue that social status is a critical dimension affecting how organizational members communicate or avoid communicating about mistreatment, and whether employees have confidence that they may effectively resolve the situation (Meares et al., 2004). Major sources of status in organizations include the formal source of organizational position as well as informal sources, such as race/ethnicity, and gender (e.g. Berdahl et al., 1996; Rossides, 1997). We therefore examine the effects of these three sources of status on the discrepancy between how targets actually respond to mistreatment and how they wish they could respond.
Organizational position is one of the most visible dimensions of social status at work. Studies show that targets appraise mistreatment and harassment as more severe when the perpetrator holds a position of higher organizational status than themselves (Langhout et al., 2005). Harlos (2010) reported that relative power vis-a-vis the perpetrator affected propensity to use remedial voice – people with less power were less likely to use voice – and that relative power also interacted with other factors, such as gender and self-esteem. Moreover, studies show that targets face a higher risk of retaliation when speaking up against superiors (Cortina and Magley, 2003). Therefore, we hypothesize that employees will feel more restricted in responding to mistreatment when the perpetrator is a superior compared with a peer and consequently they will report a larger discrepancy between their ideal and actual responses.
However, we argue that previous findings provide an oversimplified picture of the relationship between organizational status and responses to mistreatment. We argue that a gap in power between a perpetrator and target of mistreatment will restrict the target from responding even when the power differential is in the opposite direction. It has been argued that accountability leads high-power individuals to more strongly consider the social consequences of their behavior and thus to inhibit it (Keltner et al., 2003). Given that high-power individuals in organizational contexts often are monitored and held accountable for their behavior toward subordinates, they may feel restricted in their options to respond assertively to mistreatment from a subordinate. High-power individuals may also feel a sense of responsibility toward a subordinate (Fiske and Berdahl, 2007; Overbeck and Park, 2001), as well as the need to model appropriate behavior and guide the subordinate. Finally, powerholders are less likely to engage in risky behavior when their power is stable (Jordan et al., 2011). The larger the gap between a high power target and a low power perpetrator, the more stable the target’s power may be and the less inclined the target may be to take risky, or assertive, action in response to mistreatment from a subordinate.
As such, we argue that power differentials are important predictors of target responses to mistreatment, regardless of the direction of these power differentials. Consequently, we expect employees to feel least restricted to respond to mistreatment when the mistreatment comes from peers, rather than when it comes from organizational superiors or subordinates.
Hypothesis 1: The discrepancy between a target’s actual and ideal responses to mistreatment is larger when there is an organizational status difference between the target and the perpetrator of mistreatment.
In addition to formal organizational power, ethnicity is also linked with social status (Thompson and Subich, 2011). Studies on workplace bullying and mistreatment show that Non-White employees typically report experiencing more negative treatment than White employees in the workplace (e.g. Berdahl and Moore, 2006; Lewis and Gunn, 2007). Selective incivility – low intensity forms of mistreatment targeted at selected individuals – has even been described as a modern and highly subtle form of discrimination (Cortina, 2008). Moreover, Meares et al. (2004) found that, compared with European American respondents, ethnic minority respondents who had been mistreated typically felt less in control and were less likely to believe they could find a constructive solution through voice. Thus, we expect that when Non-White respondents are mistreated, they are more sensitive to perceived threats and retaliation and thus more restricted in their responses than White respondents who are mistreated. Furthermore, this stifling of Non-Whites’ ability to respond assertively to mistreatment will be greater to the degree the perpetrator of this mistreatment is White.
On the other hand, in the context of the growing number of organizations adopting strict anti-discriminatory policies, particularly with respect to ethnicity, some White employees may feel uncomfortable openly confronting or disciplining Non-White employees. Whites may fear that their behavior will be construed as bigoted or prejudiced and may therefore feel threatened in interaction with Non-Whites or even take compensatory actions to demonstrate the opposite (e.g. Harber, 2004; Ruscher et al., 2010). A number of studies show that Whites are hesitant to give critical feedback to Non-Whites, even for substandard performance, out of fear of being perceived as racists (Harber, 2004; Ruscher et al., 2010). Similarly, employees who have been mistreated by a Non-White employee may fear reporting or confronting the perpetrator for fear of being perceived as racist. Thus, White employees may also deliberately refrain from assertively responding to an employee of another ethnicity who mistreats them. If this is the case, we should observe the following:
Hypothesis 2: The discrepancy between a target’s actual and ideal responses to mistreatment is larger when targets and perpetrators differ in ethnicity.
Power and gender are intimately associated with one another (e.g. Carli, 1999), and despite gains made toward gender equality during the past few decades, gender inequalities remain persistent. Studies show that women still have less power and influence in organizations than men and that the glass ceiling remains largely intact (Guinote and Vescio, 2010). We can therefore expect that female targets of mistreatment are generally more likely than male targets to fear threat and retaliation (e.g. Berdahl et al., 1996), particularly when a perpetrator of mistreatment against them is of the other sex (Berdahl, 2007). This is in line with the conclusion of Pinder and Harlos (2001) that women’s lower status fosters their silence about workplace mistreatment.
As with ethnicity and organizational status, however, we can also identify reasons why men who are mistreated by a woman might feel more restricted in their responses than if they are mistreated by a man. Again, the adoption of anti-discriminatory policies may make some men feel uncomfortable about openly confronting or disciplining women out of fear of being viewed as sexist. To the extent that men believe in stereotypes of women as relatively emotional and sensitive (Guinote and Vescio, 2010), men may choose not to take action against a woman if they fear an ‘overly sensitive or emotional’ response. Studies show that male managers are more hesitant to provide critical feedback to female subordinates (e.g. Alimo-Metcalfe, 1994; Corby, 1982), which may partly be explained by a desire to ‘protect vulnerable women’. Finally, boys are typically socialized to not be overtly aggressive against girls and are rewarded for being self-reliant and independent (e.g. Ely and Padavic, 2007; Harlos, 2010). Men may therefore be expected not to show they are bothered by negative acts from women. Considering the arguments above, we hypothesize:
Hypothesis 3: The discrepancy between a target’s actual and ideal responses to mistreatment is larger when the target and the perpetrator differ with respect to gender.
Cortina and Magley (2009) have shown that the type and frequency of mistreatment affects appraisals of its severity. Moreover, appraisals, in turn, predict how targets respond (Olson-Buchanan and Boswell, 2008). Overall, employees appraising mistreatment as more severe tend to choose more active responses (Cortina and Magley, 2009). This makes sense if more severe mistreatment induces a stronger desire to take active measures to combat it, or a greater need to defend oneself against more severe threats. The more severe the mistreatment, the more likely also that the target will have difficulties letting it go and will ruminate about it. Angry rumination has been conceptualized as perseverative thinking about a personally meaningful anger-inducing event (Denson, 2012). Research shows that rumination both increases desire to respond aggressively and reduces forgiveness (Denson, 2012; McCullough et al., 2007).
We propose that the more severe mistreatment is, the greater the gap there is likely to be between actual and ideal responses. While actual responses may become more assertive with more severe mistreatment, ideal responses may become even more assertive, making the gap between actual and ideal widen as ideal responses become ever more confrontational but situational constraints prevent such confrontation. Possible rumination about the perceived injustice may also make it difficult to forgive the offender and leave the incident behind. Instead, the target of a severe case of mistreatment may start harboring thoughts of revenge or at least a desire for taking a more assertive response than prudence allows:
Hypothesis 4: The discrepancy between a target’s actual and ideal responses is larger when the target appraises the mistreatment as more severe.
Finally, we argue that the coping strategy actually taken affects the degree of discrepancy between actual and ideal responses to mistreatment. For this purpose we categorized responses into five categories: (1) conflict avoidance, (2) denial, (3) social support, (4) confrontation and (5) advocacy seeking (see Knapp et al., 1997; Wasti and Cortina, 2002). Of these responses, the two last ones are the only ones that actively address the perpetrator’s behavior, whereas the others involve actions to help the target cope with the behavior.
Previous research has shown that employees who do not use active strategies but endure severe mistreatment in silence are the ones who report the lowest psychological and physical well-being, even lower than those who exercise voice and experience retaliation in return (Cortina and Magley, 2003). Passive responses to interpersonal conflict have also been reported to amplify employee strain (Dijkstra et al., 2009). Restraining from taking direct action when faced with injustice seems to involve self-silencing and emotional suppression, which might lead to rumination, cognitive dysfunction, decreased mental health and physical health (Cortina and Magley, 2003; see Gross and Levenson, 1997, on emotional regulation). In contrast, by confronting or seeking organizational support, the target may possibly reach some level of resolution or justice, which might give the target restored possibilities to again focus on his or her work and regain confidence in the organization. Moreover, an apology from the offender may allow for interpersonal forgiveness among parties and a ‘fresh start’ (Frantz and Bennigson, 2005). Thus, we hypothesize that those actively trying to change the situation when faced with mistreatment will feel happier with their responses and report less discrepancy between actual and ideal responses:
Hypothesis 5: The discrepancy between actual and ideal responses will be (a) smaller when targets have actively tried to change the situation (confronted the target or sought organizational support) and (b) larger when targets have chosen silent strategies (conflict avoidance or denial).
Method
Design and participants
The data of this study were collected at a major North-American research university. Faculty received an email invitation from their head of human resources that informed them of the survey and provided a web link to it. Employees were told that the survey was being conducted by an independent academic researcher and that their responses would remain strictly confidential. Just over one-third (737) of those invited (2100) completed the survey. Consistent with the make-up of the faculty, there were more male (357) than female (255) respondents, with 125 faculty members choosing not to indicate their sex on the survey.
Respondents answered questions about their exposure to a variety of forms of mistreatment and then answered detailed questions about the incident that had bothered them the most, including questions about the perpetrator’s gender, ethnicity and organizational status relative to themselves. Respondents were then asked to indicate how they responded to the mistreatment and, with an open-ended question, how they wish they had responded. Two-hundred and seventeen respondents completed this open-ended question. About half of these respondents were female (105; 107 were male). The age of these respondents varied: only two were under 30 years old, 55 were in their thirties, 72 were in their forties, 58 were in their fifties and 30 were 60 years or older. In the regression analysis we use Listwise deletion and thus due to some random missing information, the final full sample is N = 193.
We performed a missing data analysis by comparing the 217 respondents who had answered the open-ended question and the 520 who had not answered it using a t-test. This analysis revealed no statistically significant differences in ethnic dissimilarity, organizational status difference or gender dissimilarity between target and perpetrator for respondents and non-respondents. However, women were more likely than men to answer the open-ended question about how they wished they had responded to mistreatment. Further, those who answered this question rated their mistreatment as more severe and were more likely to use (any kind of) coping than those who did not answer this question. We discuss implications of this analysis on our results in the discussion section.
Measures
Hierarchical status difference
The perpetrator was a superior in 89 occasions, a peer in 79 occasions and a subordinate in 34 occasions. In 15 of the cases the target did not indicate the perpetrator’s organizational status. An ANOVA showed significant differences in the discrepancy between actual and ideal responses to mistreatment as a function of target-perpetrator difference in organizational status (F = 7.927, d.f. = 2, p < .001). However, the mean difference between subordinates and supervisors was not significant (t = -.471, d.f. = 121, NS), albeit the mean difference between superiors and peers as well as subordinates and peers was significant (t = 2.46, d.f. = 110, p < .05; t = -3.802, d.f. = 165, p < .001). Thus, the level of discrepancy between actual and ideal responses to mistreatment was higher when the perpetrator was either a subordinate or superior. We therefore coded the organizational status difference between respondent and perpetrator as follows: 0 = perpetrator was a peer, 1 = perpetrator was a subordinate or superior.
Gender dissimilarity
We used two variables to analyze the effect of gender on our results. One was the gender of the target; the second was whether the gender of the perpetrator was the same or different from the target’s gender. Gender dissimilarity was coded as follows: 0 = same gender, 1 = different gender. Men were more often mistreated by other men (66.0%) than by women (34.0%), whereas women were almost as likely to be mistreated by women (45.5%) as by men (54.5%). We included gender of the target as a control variable.
Ethnic dissimilarity
Ethnic dissimilarity was coded as follows: 0 = same as mine, 1 = different than mine. In 159 occasions the perpetrator was of the same ethnicity as the target and in 52 cases the perpetrator was of a different ethnicity. We did not have data on the specific ethnicity of targets and perpetrators (e.g. we only knew whether they were similar or different), so we were not able to test different ethnic configurations separately.
Perceived severity of mistreatment
Perceived severity of mistreatment was measured by using a single item: ‘Overall, how much did it bother you?’ Responses were given on a 5-point scale, ranging from 1 = very little to 5 = a lot. On average, respondents reported that their mistreatment bothered them ‘quite a bit’ (3): mean 2.58 (SD .98).
Actual responses
Respondents were asked what they had done in response to the mistreatment. The items were based on the ‘Coping with Harassment Questionnaire (CHQ)’ (Fitzgerald, 1990). The items used capture the different forms of responses identified by Knapp et al. (1997) and Wasti and Cortina (2002): (1) conflict avoidance (e.g. I stayed away from them as much as possible), (2) denial (e.g. I decided it was not really important), (3) confrontation (e.g. I told them I did not like what they were doing), (4) social support seeking (e.g. I talked to a friend or a family member about it) and (5) advocacy seeking (e.g. I talked to a superior about it). We created a binary index for each of the five response categories. Targets who had used at least one of the coping strategies in a category received a 1 for that category; targets who had not used any of the strategies in a category received a 0 for that category. These response styles do not always co-occur (i.e. talking to friend or family member does not always indicate that the person would also talk to a mentor or colleague). Thus, the indicators are not necessarily correlated but convergence is not required from formative indicators (Bollen, 1984; Diamantopoulos and Siguaw, 2006).
Discrepancy between actual and ideal responses
Respondents were asked how they would ideally have responded to the situation, and provided written answers to this open-ended question. These replies (n = 217) were coded by two academic experts who determined the amount of discrepancy between targets’ actual response and the response the target would ideally have taken. The raters were blind to the research hypotheses. Interrater agreement was .79. In cases of disagreement, the coding was discussed until agreement was reached. The discrepancy between respondents’ actual and ideal responses was coded as 1 = no discrepancy (n = 41, 19.0%), 2 = minor discrepancy (n = 96, 44.4%) and 3 = considerable discrepancy (n = 79, 36.6%). An example of a category 1 response was: ‘What I did was appropriate’. Category 2 included responses that indicated that the person wished she/he would have done something slightly more assertive, such as ‘Probably should have pointed out the inappropriateness of the behavior’. Category 3 included responses indicating that the respondent wished they could have done something completely different, such as ‘Contact the sexual harassment officer and bring her in for a mandatory faculty meeting to talk about the gender climate’. Only one respondent expressed remorse about being slightly too assertive. 1 The one exception was removed from further analyses. Discrepancy between actual and desired responses therefore reflected how much more assertive respondents wished they had/could have been.
Qualitative analyses
As described above, the open replies were first coded by two academic coders, who were both blind to the research hypotheses. They coded the replies with respect to the content of the ideal responses. As this part was focused on coding the replies to enable subsequent quantitative analysis, it was important that coders were not affected by the hypotheses put forward and high interrater agreement was of high relevance. Interrated agreement for this part was .79. In addition to analyzing the content of the respondents’ own ideal responses, the coders also noted whether respondents themselves discussed that they would have wished others had intervened (more forcefully). The subsequent part of the qualitative analysis was undertaken by the authors of the article, predominantly the first author. This more in-depth part sought foremost to identify the different kinds of reasons respondents put forward for not taking the responses desired and to increase understanding of the potential mechanisms leading employees to refrain from taking their desired responses. Separate analyses were conducted for those mistreated by superiors and those mistreated by subordinates to identify possible patterns. Similarly, different gender configurations and ethnic configurations were analyzed separately. This part of the analysis did not aim to quantify the qualitative data, but rather use its full richness. In the results section excerpts from the open replies are used for illustrative purposes.
Results
We started our analyses by exploring how targets had actually responded to mistreatment. The results showed that most targets sought social support or used passive forms of response. Seeking some form of social support was reported by the highest number of respondents (63.0%). Conflict avoidance (54.2%) and denial (44.0%) were also reported by a high number of respondents. Less than one-third of the respondents confronted the perpetrator (29.6%), and very few filed formal complaints (3.2%). Thus overall, consistent with prior research, these results suggest that targets talked to friends, family members or colleagues about their experiences of mistreatment and tried to avoid the perpetrator, just put up with it or convinced themselves that it was not really that important.
Next we explored how the level of discrepancy between actual and ideal responses to mistreatment related to our focal variables and hypotheses. Descriptives and intercorrelations between the variables are reported in Table 1.
Descriptives and intercorrelations.
N = 193 for complete data. *p < .05, **p < .01, ***p < .001.
We used linear regression analyses to simultaneously explore the relationships between an actual-ideal response discrepancy and differences in organizational status, ethnicity and gender, as well as perceived severity and coping strategies used. The regression analysis [F (d.f. = 10, 182) = 5.027, p < .001] showed that organizational status difference (β = .162, p < .05), gender dissimilarity (β = .149, p < .05), perceived severity (β = .197, p < .05) and conflict avoidance (β = .179, p < .05) were positively related to the degree of discrepancy between actual and ideal responses to mistreatment (Table 2). This model explained 21.6 percent of the variance in this discrepancy.
Linear regression analysis results predicting amount of discrepancy.
N = 193. *p < .05, **p < .01, ***p < .001.
The regression results offer support for Hypotheses 1, 3 and 4. The results indicate that when there is a difference in hierarchical status and gender and the perceived severity of the incident is more severe, the level of discrepancy between actual and ideal target responses to the mistreatment is higher. In line with Hypothesis 5b, a conflict-avoidant coping strategy was related to a higher level of discrepancy between actual and ideal responses.
Supplemental qualitative analysis
We also analyzed the comments about ideal responses that respondents had provided in their open replies. The open replies were analyzed with respect to two questions: the content of the desired actions and the reasons given by the respondents for why this action had not been taken.
Our results showed that respondents either were generally happy with the responses they had taken, or wished they had been, or could have been, more assertive. Approximately one-fifth (19%) felt that the response they had taken was the ideal one. Approximately 44% wished they had done something slightly more assertive and 37% that they had done something considerably more assertive. Respondents did not express remorse about action they had taken, such as confronting or reporting (with the one exception noted earlier). Respondents wishing they would/could have taken more assertive action typically wished that they could have confronted the perpetrator, or at least had a discussion to let the perpetrator know how the behavior had affected them. More than half of the respondents wishing they had taken more assertive action made comments about confrontation or ‘candid discussions’ with the person in question. Examples included: ‘I wish I had the guts to tell the person off’, ‘Let him and my colleagues know how bad it makes me feel to be ignored’, or ‘I would have liked to confront them more directly in the presence of my colleagues’.
The second most common type of response involved a desire to inform the immediate superior or another superior, such as the dean, and have them take action. Several targets wished they could have made a formal complaint, leading to disciplinary action against the perpetrator – possibly even dismissal. Overall, although respondents were asked about the action they wished they themselves had taken, many respondents discussed the role of their superiors or faculty heads and expressed a wish that they would have been more active. Also, the need for more specific policies was repeatedly mentioned in replies.
Confronting and reporting were the responses discussed most in the open replies. A handful of respondents wished they could have ignored the perpetrator or incident rather than blaming themselves or feeling hurt. For example, one respondent wished he had ‘been confident enough not to let it bother’ him. Other less common responses involved thoughts of ‘revenge’ or quitting the current job. None of the respondents felt they should ideally have sought more social support. No clear differences emerged between different groups of respondents in terms of desired action.
Subsequently, the open replies were analyzed with respect to reasons put forward for not taking the desired response. Respondents were not explicitly asked about this, but several replies included reflections about this. Overall, fear of retaliation was the most common reason mentioned for a discrepancy between actual and ideal responses. For example, hierarchical status differences and power more generally were repeatedly pointed out as the main reason for silence among those mistreated by superiors. Respondents talked about ‘power brokers’, ‘circles of power’ and ‘apparent connections with high administrators’, and expressed concerns about upcoming tenure decisions or future work conditions. A general feeling of hopelessness characterized some of these comments. ‘Nothing can be done . . . You are either “in” the power circle or “out”’. ‘There’s nothing that can be done about it because of our status difference and because there is no structure/mechanism in place’.
Tenure (or rather the lack thereof) was emphasized in several of these replies. It was seen as unwise or even as ‘career suicide’ to speak up against the powerful or to raise a complaint before receiving tenure, particularly if that person was on the tenure committee.
In contrast, respondents in higher positions mistreated by subordinates expressed dissatisfaction with lack of support from higher management, lack of policies for their particular situation and a general unwillingness to intervene in ‘insubordination’. These factors were used to explain why superiors had refrained from taking more assertive action. Sometimes the decision not to take action was also guided by a (potentially misinformed) concern for the subordinate and attempt to protect the employee. For instance, a female professor wrote that she had protected a subordinate who had engaged in repeated inappropriate behavior towards her and further commented: ‘I judged her to be very fragile and hence delayed talking frankly with her’. In this case, the person protected was of the same gender, but different ethnicity.
Ethnicity was discussed by a few respondents, but was not a major theme in the comments. One of the respondents discussed the role of ‘xenophobics’ in the work community, pointing to the perceived importance of ethnicity in her own experiences of mistreatment. On the other hand, a couple of respondents explicitly said that fear of allegations of racism had led them to refrain from taking action against (visible) minority member perpetrators. No systematic differences in terms of reasons emerged for different ethnic configurations.
Gender differences between perpetrator and target were not explicitly raised by the respondents in the open replies. However, it is worth noting that two men who felt mistreated by women reported ideal responses that could be described as containing sexualized content; for example, a desire to ‘spank her until she came across’. It seems clear that societal norms prevented these respondents from carrying out their desired actions. While these examples may be seen as exceptions rather than as typical reflections, it is still worth noting that no such examples emerged with the other gender configurations. Wanting to send the perpetrator to counselling was another reflection that was only mentioned by a man mistreated by a woman. A few women being mistreated by men explicitly discussed the mistreatment itself from a gender perspective, emphasizing that the perpetrator had treated other (young) women in the same way. However, none of the women reflected upon gender as a factor preventing them from taking action against male perpetrators; instead, the explanations involved formal status differences, lack of tenure or organizational politics.
Discussion
Our results highlight the importance of social status discrepancies for responding to mistreatment. Previous studies have shown that social status affects the amount of mistreatment employees are subjected to (e.g. Berdahl and Moore, 2006; Cortina, 2008; Lewis and Gunn, 2007), how employees appraise mistreatment (e.g. Berdahl et al., 1996; Berdahl and Moore, 2006; Cortina and Magley, 2009; Lamertz and Aquino, 2004; Salin and Hoel, 2013) and how employees cope with mistreatment (Cortina and Magley, 2009; Harlos, 2010). This study extends current research by demonstrating how different facets of social status relate to whether targets of mistreatment are able to respond in ideal and desired ways to the mistreatment.
The primary focus and contribution of this study is what it reveals about the gap between how employees actually respond to being mistreated and how they would ideally have liked to respond, and what predicts this gap. Consistent with prior research, this study reveals a high tendency to rely on passive forms of responses or social coping, rather than more active strategies – such as confronting or advocacy seeking – which could actually stop the mistreatment. Employees seldom confronted the perpetrator and very few filed formal complaints.
The preference for avoidant and unassertive response strategies may partly reflect the fact that many of the incidents in this study were rather low-key and thus did not necessarily provoke strong feelings. However, our faculty respondents were asked to reflect upon the episode of mistreatment that had bothered them the most, and average ratings of the severity of the mistreatment episode were moderately high. Similar results are reported in other research on mistreatment – including incivility, bullying and sexual harassment – making it likely that our results reflect a more general trend (see Cortina et al., 2002). Zapf and Gross (2001) point out that because mistreatment and bullying represent situations in which the target has little control and little power, active strategies are often neither useful nor possible. This explains why many targets instead resort to cognitive restructuring, denial, avoidance, or simply ‘doing nothing’. Moreover, targets find it difficult to take more direct action as the mistreatment often is subtle and very hard to prove.
Respondents who reported a discrepancy between their actual and ideal responses, however, wished they would, or could, have been more assertive. This is in line with the finding that having resorted to conflict avoidance was associated with a higher tendency to report a discrepancy – this route seems to have been chosen not because the target wanted to choose it, but because he or she found it too risky to do otherwise. Given that both confrontation and advocacy-seeking may result in retaliation (see Cortina and Magley, 2003), one could expect those who choose such routes to regret their choices afterwards. However, such regrets were expressed only by one individual in our sample. Instead, when those who had taken active responses reported a discrepancy, they, too, wished they had been more, rather than less, assertive.
There may be several explanations for the respondents’ very low tendency to regret assertiveness. First of all, it is possible that only those who felt rather ‘safe’ chose this approach and did so only after judging the probability of success to be very high or the risks to be very low. Second, choosing an active route may have given the targets some general satisfaction, even if they were not successful in ending the mistreatment or even if they faced retaliation. Miller (2001) points out that there are two particularly powerful motives that guide people’s responses to acts of disrespect: restoration of self-esteem and educating the offender. It is thus possible that the restoration of self-esteem and education of the offender outweighed potential costs of confronting the perpetrator or reporting. This is in line with findings showing that those who suffer mistreatment in silence report the worst health outcomes (Cortina and Magley, 2003) and that passive responses to interpersonal conflict amplify the relationship between conflict and strain (Dijkstra et al., 2009).
The open replies in this study revealed that those experiencing a discrepancy between their actual and ideal responses to mistreatment typically refrained from a more assertive response out of fear. Research on voice and silence typically present employee decisions as deliberate choices based on perceived efficacy and perceived safety, that is, likely benefits and risks (Morrison, 2011). This article broadens our understanding of employee decision-making about speaking up and taking assertive action by showing that not only upwards power differences, but also downward power differences affect employee behavior. However, the reasons for refraining from taking assertive action were different when superiors had been mistreated by subordinates. The open replies highlight the fact that refraining from taking action sometimes was motivated by the desire to portray oneself as a ‘just person’, that is, somebody who is not a racist or who protects his or her subordinates. Lack of policies and lack of support were other reasons discussed by those in a higher hierarchical position.
This study examined in more detail the factors that predict a discrepancy between what respondents would have liked to do and what they actually did. The results highlighted the importance of social status and showed that differences between perpetrator and target status were important to predicting the degree of the discrepancy. The study encompasses a broad range of mistreatment, ranging from relatively minor instances of ‘incivility’ to blatant mistreatment and abuse, indicating that our results are relevant for a broad range of different behaviors. An important contribution of this article is that it shows that it matters less whether the perpetrator is of a higher or a lower status than the respondent, as a linear model would predict, and that it matters more whether the perpetrator is simply of a different hierarchical status than the respondent. The difference in discrepancy scores was not significant when comparing responses to mistreatment by superiors or by subordinates, but both of these conditions differed significantly from responses to mistreatment from peers. While status differences mattered regardless of direction, the mechanisms appeared to be different: subordinates refrained from taking action out of fear of retaliation, whereas superiors refrained from taking action primarily because of lack of support/policies or because they wanted to protect the perpetrator or be seen as ‘fair persons’. This clearly demonstrates the very complex relationship between social power, mistreatment and responses.
The study found support for the hypotheses concerning hierarchical status and gender differences. In contrast, our second hypothesis, which concerned ethnicity, had to be rejected as our quantitative findings did not reach statistical significance. The lack of statistical significance in the regression analysis may partly be explained by a relatively low frequency of mixed-ethnicity perpetrator-target dyads, which reflects the fact that the majority of the faculty members at this particular university are White. We thus encourage other researchers to continue investigating this in larger or more racially mixed samples.
The results also indicate that the higher the perceived severity of mistreatment, the higher the discrepancy between actual and desired responses. This may seem counter-intuitive at first, as a higher severity also is likely to lead to more assertive actual responses (see Cortina and Magley, 2009). However, the finding that higher severity is still associated with higher discrepancy appears to highlight the gap between what targets of mistreatment are actually able to do in response to the mistreatment and what they wish they could do, particularly in instances of severe mistreatment. Rumination may also play an important role.
Limitations and future directions
This study is not without its limitations. We asked respondents about their ‘ideal’ responses to mistreatment in retrospect. Respondents are unlikely to have a stable idea of what constitutes an ‘ideal’ response; rather, this may evolve as the result of a long process. This study does not allow us to analyze how respondents’ conceptions of ‘ideal’ responses evolve, nor does it allow us to discriminate between those respondents who, from the start, would have liked to respond differently but felt restricted, and those who originally chose what they thought was the best response but whose conceptions of an ‘ideal’ response changed over time.
Given that answering the open-ended question was voluntary, and given the results from our missing data analysis, most likely we have an underrepresentation of answers in the first category of responses regarding the discrepancy between actual and ideal responses to mistreatment (i.e. no discrepancy). Thus, our regression analysis results might be somewhat suppressed by the fact that those who had a lower severity appraisal and had not tried to cope with the incident were more likely to opt out from answering the question. This provides a more conservative test to the hypotheses as a result of scale truncation. Further, coding responses in three categories introduces scale coarseness, and future research on discrepancy might want to employ scales that are truly continuous, both alleviating concerns of scale truncation and coarseness.
Some other measures are not optimal. For example, our coping measure could be criticized because we use a formative measurement rather than a reflective one. Thus, the validity and reliability of the measure cannot be fully established. Furthermore, we had information about the difference between the target and perpetrator in ethnicity, but not the actual ethnicity of the target (or perpetrator). Additional factors could have been included to predict the discrepancy between how targets would have liked to respond as opposed to how they actually responded when faced with mistreatment. Individual characteristics of targets – for example, locus of control, moral judgment or history of mistreatment – could be investigated in future studies. In terms of contextual factors, we may also hypothesize that people will be more likely to report a smaller discrepancy when they perceive the organization as being a supportive and psychologically safe environment. Understanding the role played by personality and contextual factors are important avenues for further research.
This study was limited to respondents currently employed by a particular university. However, this may have led to a somewhat biased sample. Studies show that one potential risk of workplace mistreatment is unemployment (Salin, 2013). Furthermore, if employees who respond assertively face retaliation, they could face increased risks of losing their jobs. A survey like this, conducted among those who are currently employed, would not capture the experiences of former employees, potentially underestimating the number of individuals regretting assertiveness. Still, as we would expect that job loss is at the extreme end of the retaliation scale, it is notable that we do not find more respondents regretting assertiveness.
Finally, because the data were collected in one organization (and thus also one occupational context), in one country, we can only speculate about its generalizability. For instance, having or lacking tenure is specific to the academic work community and may thus shape responsive action taken more than degree of job (in)security in other occupational contexts. Difficulties with short-term job mobility, as compared with many other occupations, may create particular obstacles for employees wanting to take more assertive measures. Then again, a preference for passive coping strategies has been reported in many previous studies (Bies et al., 1997; Cortina et al., 2002; Zapf and Gross, 2001) and is not unique to our sample. Still, the specific nature of the academic work context calls for replications of our study also in other occupational contexts.
Practical applications
The results demonstrate the difficulties for targets to successfully address mistreatment themselves – even relatively high-status targets, such as these university professors – and highlight the need for organizations to address mistreatment proactively, even if targets do not make formal complaints. Previous results have shown that targets who do not speak about mistreatment report the worst psychological and physical health scores (Cortina and Magley, 2003). This is in line with research on emotional regulation that shows that suppression of emotions may have negative impact not only on health, but also on cognitive functioning (Gross and Levenson, 1997). Impairments in health and cognitive functioning lead to economic risks for organizations, alongside ethical problems. Furthermore, if neither the targets themselves nor organizational representatives confront perpetrators or do anything to actively resolve the situation, the behavior is likely to continue and the perpetrator may further harm both the original target and others. In addition, employees learn by modeling what others in the organizations do and what they get rewarded and punished for. Negative acts may easily spread throughout the organization if they are not tackled and prevented (see Bandura, 1977, on social learning theory).
Thus, a responsible employer needs to identify and address issues of mistreatment proactively, rather than sitting around and waiting for employees to file complaints. Employers need to be especially alert when it comes to mistreatment between employees holding different levels of social status, as targets in such cases may be least able to take action themselves. As has been pointed out by sexual harassment researchers (e.g. Thacker, 1996), it is important that training on harassment and mistreatment also explicitly addresses targets’ inclination to stay silent, and this training needs to provide supervisors with tools to encourage targets to speak up early, without retaliation. From a productivity perspective, and to avoid costly litigation and investigation, it is important that employees can seek recourse internally and have faith in internal procedures. Perceived organizational support and procedural justice have further been argued to reduce the negative effects of workplace mistreatment on employee attitudes and well-being (Parzefall and Salin, 2010).
Overall, respondents wished they would, or could, have been more, rather than less, assertive in their responses to mistreatment, but refrained from doing so out of fear of retaliation, lack of support or a fear to come across as harsh or ‘unfair’. Because employees fear retaliation for assertively responding to mistreatment, organizations must go out of their way to make sure that retaliation does not occur, using both preventative and punitive measures, if necessary (e.g. penalizing employees who retaliate against whistle-blowers).
This study highlighted the complex relationships between social status and responses to mistreatment, showing that status differences matter regardless of direction. Overall, the complex relationships between different forms of social status and mistreatment need to be further studied in order to make sure mistreatment can be managed as effectively as possible when it occurs.
Footnotes
Acknowledgements
We would like to thank Aijana Abdyldaeva and Evelina Petrov for helping us with coding. Also, we would like to thank Birgit Pauksztat and two anonymous reviewers for helpful comments on earlier versions of the manuscript.
Funding
The research received funding from the Social Sciences and Humanities Research Council of Canada (Grant #410-2006-0639).