Sunteți pe pagina 1din 17

Journal of Applied Psychology 2011, Vol. 96, No.

4, 713729

2011 American Psychological Association 0021-9010/11/$12.00 DOI: 10.1037/a0021593

When Distress Hits Home: The Role of Contextual Factors and Psychological Distress in Predicting Employees Responses to Abusive Supervision
Simon Lloyd D. Restubog
The Australian National University

Kristin L. Scott and Thomas J. Zagenczyk


Clemson University

We developed a model of the relationships among aggressive norms, abusive supervision, psychological distress, family undermining, and supervisor-directed deviance. We tested the model in 2 studies using multisource data: a 3-wave investigation of 184 full-time employees (Study 1) and a 2-wave investigation of 188 restaurant workers (Study 2). Results revealed that (a) abusive supervision mediated the relationship between aggressive norms and psychological distress, (b) psychological distress mediated the effects of abusive supervision on spouse undermining, (c) abusive supervision had a direct positive relationship with supervisor-directed deviance, (d) the positive relationship between psychological distress and spouse undermining was stronger for men as opposed to women, and (e) employees engaged in relationship-oriented occupations reported greater levels of abusive supervision and psychological distress. Implications for theory and practice are discussed. Keywords: aggressive norms, abusive supervision, interpersonal conflict, psychological distress, spouse undermining

The past two decades have witnessed a renewed interest in the study of the dark or destructive side of supervisory behavior (Tepper, 2000, 2007). Recent research, in particular, has focused on abusive supervision, defined as subordinates perceptions of the extent to which supervisors engage in the sustained display of hostile verbal and nonverbal behaviors, excluding physical contact (Tepper, 2000, p. 178). This body of research indicates that abusive supervision occurs when supervisors perceive that they have been treated unjustly by their organization (e.g., procedural injustice, psychological contract violation) and/or its agents (e.g., interactional injustice) and that it is negatively related to employee attitudes (e.g., life and job satisfaction, organizational commitment) and behaviors (e.g., organizational citizenship behavior, job performance; see Tepper, 2007, for a review). Though sparse,

This article was published Online First January 31, 2011. Simon Lloyd D. Restubog, Research School of Business and School of Management, Marketing and International Business, The Australian National University, Canberra, Australia; Kristin L. Scott and Thomas J. Zagenczyk, Department of Management, Clemson University. Author names in the bylines are alphabetically arranged; all three authors contributed equally. This research was supported by Australian Research Council Grant DP1094023 awarded to Simon Lloyd D. Restubog. We are grateful to Ray Aldag, Michelle Duffy, Catherine Collins, and Karin Sanders for their valuable suggestions on an earlier version of this article. We also thank Chris Asuncion and Franco Quodala for assistance in data collection and Patrick Garcia and Lemuel Toledano for research assistance. Correspondence concerning this article should be addressed to Simon Lloyd D. Restubog, Research School of Business and School of Management, Marketing, and International Business, 1st Floor, LF Crisp Building 26, College of Business and Economics, The Australian National University, Canberra, ACT 0200, Australia. E-mail: simon.restubog@anu.edu.au 713

research examining potential moderating and mediating variables (e.g., personality, organizational justice, affective commitment) has also ensued (e.g., Tepper, 2000; Tepper, Henle, Lambert, Giacalone, & Duffy, 2008). Despite progress devoted to understanding the antecedents and outcomes of abusive supervision, we still know relatively little about how organizational or contextual factors shape supervisor subordinate interactions. This oversight is critical, as supervisor subordinate interactions are embedded within and influenced by the broader organizational environment that shapes interpersonal relationships and employees work experiences (Erdogan, Liden, & Kraimer, 2006). To address this gap in the literature, we draw on social learning theory (Bandura, 1973, 1986) to argue that supervisor perceptions of aggressive norms within an organization (i.e., negative or hostile attitudes and behaviors toward other employees) is associated with high levels of employee-reported abusive supervision. Importantly, the perspective that we present explores the possibility that contextual factors are potential triggers of abusive supervisory behavior (Tepper, 2007). Furthermore, on the basis of Lazarus and Folkmans (1984) transactional theory of stress, we posit and empirically test the argument that employees who become distressed (i.e., fearful and anxious) as a result of experiencing abusive supervision will displace their aggression by engaging in spousal undermining behaviors (e.g., criticizing, ignoring, demeaning), whereas abused employees who do not experience distress will retaliate directly against the supervisor by adopting subtle yet deviant behaviors (e.g., withholding information, slander, gossip). In doing so, we make several contributions to the literature on abusive supervision. Our foremost contribution is conceptualizing and testing a top-down model that links organizational/contextual factors (aggressive norms) to abusive supervision and both employee significant other and supervisor outcomes. In particular, we

714

RESTUBOG, SCOTT, AND ZAGENCZYK

extend other top-down models (e.g., Hoobler & Brass, 2006) by (a) incorporating the organizational contextual variable of aggressive norms into our model and (b) integrating psychological distress as an explanatory mechanism that determines whether employees will retaliate against supervisors or will instead displace aggression by undermining partners/spouses. Further, our consideration of both gender and occupational differences in the abusive supervision dynamics provides new insights for both scholars and practitioners interested in the abusive supervision phenomenon. Figure 1 shows our proposed theoretical model. We test our hypotheses in two independent, longitudinal multisource investigations. In the sections that follow, we delineate the constructs in our model and discuss the expected relationships among them.

aggressive behavior at work: Aggressive behavior of a focal individual was positively related to the aggregate level of aggression demonstrated by all workgroup members. Finally, employees were more apt to perform deviant behaviors in response to abusive supervision when they felt that deviant behavior was the norm (Tepper et al., 2008). Collectively, these studies suggest that bad behavior in the organization can lead to bad behavior by individuals. In the context of our study, we argue that supervisors who perceive that norms toward aggression are present in their organizations are likely to be perceived as abusive by their subordinates because they will learn and eventually adopt hostile patterns of behavior. Thus, we expect a direct and positive relationship between aggressive norms and perceptions of abusive supervision.1

Theoretical Background and Hypotheses


In this work, we build on the body of research highlighting the influence of the organizational context on supervisor employee relationships (e.g., Cogliser & Schriesheim, 2000; Erdogan et al., 2006; Green, Anderson, & Shivers, 1996). Through the lens of social learning theory (Bandura, 1973), we explicitly make the connection between aggressive norms and abusive supervision. Following this theory, Bandura (1986) has stressed that individuals will observe the behaviors, attitudes, and emotional reactions of others to learn what behaviors and attitudes are acceptable in a given social context. At work, employees model the behaviors of coworkers to fit in or be successful; behaviors that help them are adapted, whereas those that are unsuccessful are rejected (Ibarra, 1999). In support of this notion, studies demonstrate that individuals pay attention to the behavior of coworkers with whom they maintain strong ties and admire to learn organizational citizenship (i.e., helping behavior; Bommer, Miles, & Grover, 2003; Zagenczyk, Gibney, Murrell, & Boss, 2008) and other attitudes or actions that facilitate successful role transitions at work (Ibarra, 1999). With respect to aggressive behavior specifically, Banduras (1973) social learning theory emphasizes the important role of the social context in teaching individuals to engage in such behavior. Through repeated exposure to aggressive behaviors, and then by eventually modeling the negative, though seemingly acceptable behaviors of others in the social context, individuals become likely to adopt aggressive behaviors themselves. For example, prior research has shown that exposure to aggressive norms in ones neighborhoods and household often results in aggression both at home (Berkowitz, 1993; Garcia, Restubog, & Denson, 2010) and in the workplace (Douglas & Martinko, 2001). In the context of the present study, we argue that when aggressive norms are displayed from the top of the organization down through the ranks, aggressive behaviors will also flow downward. To this end, researchers have speculated that work environments that support a degree of hostility may result in aggressive or violent behavior in the workplace because such environments legitimize aggressive or abusive treatment (Hoobler & Brass, 2006) or teach employees to behave in such a manner via processes described in social learning theory (OLeary-Kelly, Griffin, & Grew, 1996). Empirical research supports this notion as well: Aquino, Douglas, and Martinko (2004) demonstrated that employees exposed to aggressive role models were more likely to engage in aggressive behaviors themselves. Similarly, Robinson and OLeary-Kelly (1998) and Glomb and Liao (2003) showed that social learning influenced the adoption of

Employee Responses to Abusive Supervision and Psychological Distress


Once employees experience abuse from a supervisor, how are they likely to respond? Prior work on employee reactions to abusive supervision has yielded seemingly contradictory results with respect to this question. Some evidence suggests that employees retaliate against supervisors perceived as abusive (e.g., Mitchell & Ambrose, 2007; Tepper et al., 2008), whereas other research reveals employees of abusive supervisors undermine spouses instead of retaliating directly against supervisors, presumably to avoid being fired or otherwise disciplined (Hoobler & Brass, 2006). Recently, Tepper et al. (2009) acknowledged that not all employees who experience mistreatment seek revenge (p. 157) and shed some light on this issue by demonstrating that employees of abusive supervisors with greater intentions to leave the organization (i.e., were less threatened by the supervisors power and authority over them) reported high levels of supervisor-directed deviance. In this research, we attempt to clarify how employees are likely to respond to abusive supervision by drawing on the transactional model of stress and coping (Lazarus & Folkman, 1984). We derive predictions concerning both direct (e.g., supervisor-directed deviance) and indirect (e.g., spouse undermining) responses to abusive supervision and argue that employees who experience distress will cope by directing their negative emotions away from the source of their fear and frustration and will undermine spouses, whereas employees who do not experience distress will retaliate directly against their supervisors. According to Lazarus and Folkmans (1984) transactional theory of stress, individuals abilities to cope with stressful situations are influenced by their thoughts and feelings about the stressor. Such appraisals generally lead either to emotion-based or problem-based coping strategies. Problem-based coping serves to alleviate the stressful situation by eliminating the source of stress and involves taking direct action against the stressor often through aggressive or deviant behaviors. In contrast, emotion-based coping involves behavior intended
We recognize that past research (e.g., Aquino et al., 2004; Tepper et al., 2008) has argued for the moderating effects of organizational norms. Our hypothesis is consistent with research on the theory of reasoned action (Fishbein & Azjen, 1975) and the theory of planned behavior (Azjen, 1991), which both suggest that perceived norms directly influence intentions to engage in a behavior. Indeed, a substantial body of research supports the notion that perceived norms directly influence behavioral intentions (see Azjen, 1991).
1

CONTEXTUAL FACTORS AND ABUSIVE SUPERVISION

715

Occupa on

Gender

Aggressive Norms

Abusive Supervision

Psychological Distress

Spouse Undermining

Supervisor-directed deviance

Figure 1. Proposed theoretical model.

to minimize the negative emotional impact of a stressful situation. Such actions are generally focused on changing, or elevating, the focal individuals internal emotional state as opposed to eradicating the source of stress. This type of coping allows the focal individual to avoid the stressor through denial or psychological distancing as opposed to direct confrontation. Subsequently, unpleasant emotions stemming from this situation may be ameliorated (Lazarus, 1991). Research suggests that individuals will generally try to do something about a stressful situation (i.e., eliminate the stressor) unless it is perceived that such actions are unlikely to be successful or the individual is too fearful to take direct action (Aquino & Thau, 2009; Lazarus, 1993). Under the latter circumstances, emotion-based coping (i.e., avoidance) predominates. Abusive supervision and distress. The transactional theory of stress defines a stressor as occurring when external or internal demands (or conflicts between them) are appraised as taxing or exceeding the resources of the person (Lazarus, 1991, p. 112). In line with this definition, we argue that abusive supervision is an external stressor that generates a number of negative thoughts and feelings. Considerable research suggests that recipients of abusive supervision are likely to report greater levels of psychological distress and stressrelated outcomes (e.g., irritability, anxiety, depression, depersonalization, somatic complaints; Duffy, Ganster, & Pagon, 2002; Tepper, 2000). Unlike eustress, or good stress, which is typically associated with positive cognitions and emotions related to self-fulfillment, psychological distress is a mental state characterized by negative thoughts and feelings related to anxiety, fear, or depression (Selye, 1974). Feelings such as these often arise following exposure to stressful and emotionally traumatic experiences such as verbal attacks, demeaning put downs, or angry gestures associated with abusive supervision (Keashly & Harvey, 2005). Thus, abusive supervision can be viewed as a significant source of stress, with profound implications for employees health and well-being. On the basis of this research, we predict that exposure to highly aggressive norms promotes abusive supervision behaviors, which in turn results in psychological distress. In sum, we predict the following: Hypothesis 1: Abusive supervision mediates the relationship between aggressive norms and psychological distress. Psychological distress and spousal undermining. We assert that psychologically distressed employees will seek to alleviate their negative thoughts and emotion by adopting nonviolent, though aggressive behaviors (i.e., undermining) directed at

spouses. Undermining behavior is insidious yet nonviolent, much like abusive supervision, and we believe that distressed employees will seek to reciprocate this type of treatment but will direct it toward family members who pose less of a threat to the abused employee than does an abusive supervisor. According to the transactional model of stress (Lazarus & Folkman, 1984), individuals experiencing fear-based or anxiety-related distress are likely to use emotion-focused coping behavior to help alleviate these negative feelings. Indeed, ancillary research has shown that individuals generally view aggression, such as undermining others, as an outlet for venting their distress or blowing off steam (e.g., Berkowitz, 1993; Lazarus, 1991). However, such individuals are more likely to engage in displaced aggression, which involves redirecting harmful actions away from the source of fear or frustration and onto an alternative target instead (Lazarus & Folkman, 1984; Marcus-Newhall, Pederson, Carlson, & Miller, 2000). By displacing their aggression, individuals perceive that they can improve their mood and reduce their frustration while avoiding any additional unwanted stress associated with the perpetrator (Bushman, Baumeister, & Phillips, 2001). For these reasons, we argue that subordinates experiencing distress as a result of abusive supervision are likely to undermine spouses to combat feelings of distress. This is consistent with theoretical and meta-analytic research investigating antecedent conditions of displaced aggression, which indicates that aggression is likely to be displaced from the original foci to other targets when an individual is fearful of the transgressor (Marcus-Newhall et al., 2000). This is also consistent with studies within the management literature that have shown that employees will avoid aggressing directly against a supervisor for fear of retaliation, disciplinary action, or losing their job (e.g., Aquino, Tripp, & Bies, 2001; Bordia, Restubog, & Tang, 2008) and, instead, are more likely to undermine family members (Duffy, Scott, & OLeary-Kelly, 2004; Hoobler & Brass, 2006). Thus, we propose that employees who experience psychological distress following abusive supervision will engage in spouse undermining behavior to alleviate their negative psychological state resulting from abusive supervision. Hypothesis 2: Psychological distress mediates the effects of abusive supervision on spouse undermining. Abusive supervision and supervisor-directed deviance. We further acknowledge, however, that not all employees become distressed following experiences of abusive supervision.

716

RESTUBOG, SCOTT, AND ZAGENCZYK

Abused employees who do not feel distress (i.e., are not fearful or anxious as a result of abusive supervision) are more likely to use problem-focused coping by taking direct action against the perpetrator to prevent further abuse (Lazarus & Folkman, 1984). In a review of the workplace victimization literature (e.g., bullying, mobbing, abusive supervision), Aquino and Thau (2009) cited several empirical studies in which victims of workplace abuse chose to aggress, or retaliate, against their transgressors and noted that victims often view revenge as an acceptable means of bringing mistreatment to an end. This is consistent with Barling, Dupre, and Kelloways (2009) exten sive review of organizational aggression, which suggests that workers who perceive that they have been a victim or a target of aggression are more likely to engage in aggressive acts as a result. Illustrating this point in terms of abusive supervision, Mitchell and Ambrose (2007) found that employees of abusive supervisors reported engaging in higher levels of aggressive behaviors directed toward the supervisor. This finding is consistent with other studies and meta-analytic work that have shown that employees of abusive supervisors are more likely to retaliate against them (e.g., Bies & Tripp, 1996). However, as Aquino and Thau noted, most supervisor-directed retaliatory behaviors reported in prior studies of abusive supervision including the study noted abovewere more covert such as ignoring a supervisors request, not putting forth much effort on assigned tasks, or withholding citizenship behaviors that might benefit the supervisor (Tepper, 2007). Such findings are in line with studies of workplace incivility that demonstrate that less powerful employees tend to seek revenge against more powerful employees in subtle ways such as tarnishing the perpetrators reputation or reducing their productivity (cf. Pearson, Anderson, & Porath, 2005). It should be noted that none of the studies discussed above directly tested the mediating role of psychological distress between abusive supervision and supervisor-directed deviance. On the basis of Lazarus and Folkmans (1984) transactional model of stress, we believe that abusive supervision is directly related to supervisordirected deviance when employees do not experience psychological distress (i.e., distress does not mediate this relationship), and, therefore, employees are focused on problem-solving coping behaviors intended to prevent further abuse. This is consistent with Tepper et al.s (2009) study, which argued and found that employees with intentions to leave the organization, and who were thus less fearful of the abusive supervisor, were more likely to engage in supervisor-directed deviance. Hypothesis 3: Abusive supervision is positively related to supervisor-directed deviance.

differences in abusive supervision is important (Tepper, 2007). Given the negative consequences of abusive supervision, it is possible and perhaps likely that occupations notorious for the abuse of power (e.g., health care, military) may have higher levels of such consequences than other occupations. Therefore, understanding occupational differences in abusive supervision may inform some industries of problematic practices, which, if rectified, could result in higher levels of productivity and lower levels of deviance and voluntary turnover. Gender as a moderator. We draw on social role theory, which posits that gender differences stem largely from the distinct social positioning of men and women in social roles (Eagly, 1987), to explain differential responses to abusive supervision. Social expectations resulting from these roles have been shown to influence behavior and personality traits. For instance, men learn that aggressive responses are an appropriate part of instrumental behavior that is congruent with the masculine role (Eagly & Steffen, 1986). In contrast, behaviors associated with feminine expressive traits discourage women from enacting aggressive behaviors and instead cause them to utilize actions that reflect concern for others and emotional sensitivity (Eagly & Steffen, 1986). In the context of our study, we argue that men, as opposed to women, who have experienced abusive supervisory treatment are more likely to engage in aggressive behaviors directed toward their supervisors. Indeed, there is evidence to suggest that men tend to be more aggressive than women in general and in the workplace in particular (for a meta-analysis, see Hershcovis et al., 2007). Further, in these studies, the genderworkplace aggression relationship was stronger for aggression directed toward individuals than aggression directed toward the organization. Thus, we predict the following: Hypothesis 4a: Gender moderates the relationship between abusive supervision and supervisor-directed deviance. There will be a stronger positive relationship between abusive supervisor-directed deviance for men as opposed to women. Similarly, feminist theory suggests that men and women will respond to psychological distress differently as well. Feminist theory builds on social role theory and proposes that aggression toward ones spouse is rooted in the patriarchal structure of society (Bograd, 1998). In particular, there exists a hierarchical social structure characterized by an unbalanced power relationship between men and women (i.e., men viewed as more superior than women). Indeed, scholars suggest that the more men have internalized this ideological view, the more likely they are to dominate and control women (Pagelow, 1981, 1984). In the context of our study, we argue that when men, as opposed to women, experience psychological distress, they are more likely to act aggressively toward their spouse. In a meta-analytic study on gender differences in aggression toward heterosexual partners, Archer (2002) reported than men engaged in more self-reported aggression than women and that the same pattern of results was evident in partner-reported aggression. Hypothesis 4b: Gender moderates the relationship between psychological distress and spousal reports of family undermining. There will be a stronger positive relationship between

Examining the Roles of Gender and Occupation


As shown in Figure 1, we also explore the roles of gender and occupation in the abusive supervision dynamics. Understanding the role of gender in employee responses to abusive supervision is important because it speaks to how the exercise of power at work in the form of abusive supervision can yield further mistreatment both within the workplace as well as at home. To this end, Tepper (2007) calls for research exploring gender and abusive supervision. Similarly, an examination of occupational

CONTEXTUAL FACTORS AND ABUSIVE SUPERVISION

717

psychological distress and spousal reports of undermining for men as opposed to women. Occupation as a predictor of abusive supervision and psychological distress. Scholars have also speculated that occupational differences are a contextual factor that could affect the extent to which employees perceive that they experience abusive supervision (Tepper, 2007). To this end, considerable evidence suggests that stress varies across occupations as a result of differences in relationships at work (supervisors, coworkers, etc.; Cooper & Marshall, 1976; Johnson et al., 2005). For instance, in a study of 25,000 individuals encompassing 26 occupations, Johnson et al. (2005) concluded that occupations characterized by frequent face-to-face interaction with customers (e.g., social services and care providers) tend to have higher levels of job stress, along with occupations with highly stressful job content (e.g., fire brigade, ambulance, teacher, etc.). Similarly, research on emotional labor shows that employees with high as opposed to low levels of people contact (e.g., frequent customer interactions) tend to experience greater levels of stress reactions (see Grandey, 2000, for review). Drawing from the workplace bullying literature, Hubert and van Veldhoven (2001) reported that prevalence rates of bullying vary significantly across business sectors. Specifically, they found that employees within the education and public sectors were more vulnerable to supervisor bullying. In another study consisting of 7,787 individuals across 14 occupational groups, Einarsen and Skogstad (1996) reported that industrial, graphical, and hotel and restaurant workers experienced greater frequency of bullying in comparison with other occupational groups. Collectively, these studies suggest that employees in customer service and helping professions are more likely to experience job stress, emotional labor, and burnout. Taken together, we predict that employees involved in relationship-oriented occupations (e.g., sales and marketing, customer service, and public relations) will experience higher levels of abusive supervision and psychological distress relative to employees in other occupational groups because employees in these occupations must interact with both customers and supervisors. As a result, such employees may be more sensitive to unfavorable treatment from supervisors than employees who have less interpersonal interaction with others. Hypothesis 5: Employees in relationship-oriented occupations will tend to report higher levels of abusive supervision and psychological distress than employees in other occupations. We tested these predicted relationships in two studies that made use of multisource data (different operationalizations of key variables, particularly aggressive norms). Study 1 was a three-wave investigation of full-time employees with supervisors in which aggressive norms were operationalized by asking the participants supervisors to report the extent to which they perceive that their organization rewards employees for being aggressive, confrontational, and negative (Douglas & Martinko, 2001). This individuallevel operationalization enabled us to achieve a balance between examining the effects of aggressive norms while minimizing samesource variance. Study 2 was a two-wave investigation of full-time workers with supervisors in which aggressive norms were operationalized as a group level property (i.e., perceptions of aggressive

norms of workers within a work unit/entity), which is consistent with previous research (Robinson & OLeary-Kelly, 1998; Tepper et al., 2008). It was particularly important for us to evaluate aggressive norms from the perspective of individuals other than the supervisor because individuals tend to believe that others in organizations have attitudes similar to their own.

Study 1 Method
Participants and procedure. We tested our predictions among 184 full-time employees enrolled in part-time masters degree in business administration (MBA) programs in Philippine universities. We collected self-report, supervisor-report, and spouse ratings of study variables. All participants completed questionnaires containing identity codes to allow each employees ratings to be matched with those of his/her corresponding supervisor and spouse. We used English because it is the language of business spoken by the vast majority of Filipinos (Bernardo, 2004) and because it is commonly used in Philippine higher education, especially in MBA programs. All participants were assured of confidentiality and informed that they could withdraw from the study at any point in time. This study utilized a longitudinal design with three waves of data collection. At Time 1, supervisor surveys were distributed to 250 part-time MBA students. Participants were instructed to give the survey to their immediate supervisor to complete. The supervisor questionnaire assessed the supervisors demographic variables as well as the supervisors perceptions of the extent to which the culture of the organization as aggressive. A total of 210 completed supervisor reports were received, yielding a response rate of 84%. Time 2 data were collected 1 month following the initial survey; in this phase, participants were asked to report their perceptions of abusive supervision, the extent to which they were psychologically distressed, as well as the degree to which they engaged in supervisor-directed deviance. Of the 250 MBA students, we received 221 completed surveys, for a response rate of 88.4%. At Time 3, 6 months after the Time 2 data collection, we requested the 221 MBA students to give a brief survey to their spouses to complete. Data from unmarried individuals were not included in the analysis. We received 184 completed surveys, yielding a response rate of 83.25%. The three waves of data collection resulted in 184 three-way matches. Of the MBA participants, 55% were men, and the average age was 28.97 years (SD 5.54). The majority of the participants (92.4%) held at least a bachelors degree, whereas the rest had completed postgraduate diplomas and courses; 44% had been working for their organization between 1 and 5 years. Participants worked in a variety of occupational job types, including accounting and finance (21.2%), customer service (17.4%), marketing and sales (10.3%), public relations (7.1%), general management and human resources (10.3%), legal (8.2%), information technology (10.3%), research and development (8.2%), and manufacturing and production (7.1%). Among the supervisor participants, 63% percent were men. Of the participants, 70% were more than 31 years of age; 67% had been working for their organization for at least 6 years. Among the spouses surveyed, 55% were women, and their average age was 27.9 years. We compared the sample of

718

RESTUBOG, SCOTT, AND ZAGENCZYK

subordinates with and without supervisor reports and spouse ratings to assess sampling bias. Univariate tests suggest no significant differences between these groups. Measures. Multi-item scales were used to ensure adequate measurement of each study variable. Unless otherwise specified, the response format for all items, excluding the demographic variables, was a 7-point Likert-type scale, with items coded such that a higher score indicated a greater amount of the focal construct (except for reverse-coded items). Space and time limitations at the Time 2 measurement period forced the shortening of some scales. Aggressive norms. At Time 1, aggressive norms were measured with a four-item scale developed by Douglas and Martinko (2001). We modified the scale slightly, as these authors sought to determine the extent to which norms outside the workplace were aggressive. Supervisors were asked to indicate on a 7-point scale ranging from 1 (absolutely not true) to 7 (absolutely true) the extent to which they believed the following statements to be true about their organization. An example item is In this organization, employees were often engaged in verbal confrontations. In this study, this scale yielded a reliability coefficient of .91. Subordinates perceptions of abusive supervision. At Time 2 (1 month after Time 1 data collection), abusive supervision was assessed with a measure developed by Tepper (2000). Participants were asked to rate the extent to which their immediate supervisor had engaged in an array of undermining behaviors. Example items include My immediate supervisor tells me Im incompetent and My immediate supervisor breaks promises he/she makes. We modified the 5-point scale to a 7-point scale ranging from 1 (I cannot remember her/him using this behavior with me) to 7 (S/he always uses this behavior toward me) for two reasons: First, it provides participants a wide range of response anchors to choose from. Second, research suggests that limited response options may result in loss of power and difficulty in detecting significant effects (Aguinis, Bommer, & Pierce, 1996). Cronbachs alpha for the scale was .97. Subordinates psychological distress. At Time 2 (1 month after Time 1 data collection), psychological distress was measured with a short four-item version of Derogatiss (1993) scale. All the items were prefaced with the following lead-in statement: In the past month, how often have you been . . . and ended with the statements such as feeling fearful, feeling restless, feeling worthless, and feeling in panic. Cronbachs alpha for the scale was .81. To provide evidence that the short and full versions were equivalent, we collected data from an independent sample of 200 full-time bank employees. Participants were instructed to complete both the long and short versions of the distress scale. Bivariate correlations indicate that the short and full versions of the psychological distress scale were highly correlated (r .92, p .001). Supervisor-directed deviance. At Time 2, we measured supervisor-directed deviance with three items adapted from Aquino, Lewis, and Bradfield (1999). We modified the referent by replacing the term coworker with supervisor. Subordinates rated the extent to which they demonstrated these behaviors toward their supervisors. These items include Gossiped about my immediate supervisor, Swore at my immediate supervisor, and Refused to talk to my immediate supervisor. Cronbachs alpha for this scale was .88. Spouse undermining. At Time 3, 6 months after Time 2 data collection, spouse undermining was assessed with a five-item scale

based on previous undermining research (Duffy et al., 2002; Tepper, 2000; Westman & Vinokur, 1998) and was modified by replacing the referent immediate supervisor with partner/ spouse. Spouses were asked to indicate the extent to which their partner (1 I cannot remember her/him using this behavior with me, 7 S/he always uses this behavior toward me) had intentionally engaged in undermining behaviors directed toward them. The items include Acts in an unpleasant or angry manner toward you, gives a critical remark on your ideas, Criticizes you, Insults you, and Gives you the silent treatment. This scale yielded a reliability coefficient of .82. Occupation. Occupation was measured using a single indicator question, What type of occupation are you currently engaged in? On the basis of previous research (Brotheridge & Grandey, 2002), customer service, public relations, and sales and marketing were coded as relationship-oriented occupations because the nature of work involves high levels of client interaction, whereas accounting and finance; legal; human resources and general administration; manufacturing and production; and information technology, research, and development were coded as nonrelationship oriented occupations. Control variables. In line with previous research on abusive supervision, we controlled for the effects of employees gender, age, and tenure. Gender was controlled to account for the gender differences between men and women in responding to angerprovoking situations (Berkowitz, 1993). Age was controlled because there is evidence to suggest that younger individuals are less likely to demonstrate restraint in expressing their emotional reactions than older individuals (Carstensen, 1992). Tenure was controlled because it is associated with aggressive behaviors at work (Robinson & OLeary-Kelly, 1998).

Results
Descriptive statistics and intercorrelations for the study variables are depicted in Table 1. Zero-order correlations were all in the expected direction. Hypotheses 1 and 2 predicted a chain of mediated relationships linking aggressive norms and outcomes. To test this, we used structural equation modeling (Anderson & Gerbing, 1988). A two-step procedure was undertaken. In the first step, we estimated a full measurement model with all the manifest indicators to establish the factor structure of the variables in the model. The second step involved testing the proposed structural model. Measurement model. As a first step, we created item parcels to improve the ratio of N relative to the number of parameters to be modeled (Little, Cunningham, Shahar, & Widaman, 2002). Item parcels refer to aggregate-level indicators that consist of the average of two or more items (Little et al., 2002). On the basis of factor analysis, parcels were created by combining items with the highest and lowest factor loadings, followed by the items with the next second highest and lowest loadings. The hypothesized measurement model had an acceptable fit with the observed data, 2(79, N 184) 201.17, p .001, comparative fit index (CFI) .96, TuckerLewis index (TLI) .95, and root mean square error of approximation (RMSEA) .09. The standardized path estimates of the manifest indicators (ranging between .76 and .97) were all statistically significant. We also compared this measurement

CONTEXTUAL FACTORS AND ABUSIVE SUPERVISION

719

Table 1 Descriptive Statistics and Zero-Order Correlations of Study Variables in Study 1 and Study 2
Study variables M SD 1 2 Study 1 (N 1. 2. 3. 4. 5. 6. 7. 8. Gender Age Tenure Occupation T1 aggressive norms (Sup) T2 abusive supervision (Sb) T2 psychological distress (Sb) T2 supervisor-directed deviance (Sb) 9. T3 spouse undermining (Sp) 0.45 28.97 2.20 0.35 2.90 2.81 3.63 2.65 2.56 0.50 5.54 1.21 0.48 1.14 1.61 1.31 1.42 1.04 .12 .11 .09 .41 .46 .43 .38 .36 3 184) 4 5 6 7 8 9

.80 .16 .16 .04 .18 .06 .02 Study 2 (N

.19 .15 .07 .10 .04 .08 188)

.20 .30 .33 .09 .08

(.91) .54 .45 .30 .19

(.97) .69 .58 .27

(.81) .30 .44 (.88) .27 (.82)

1. 2. 3. 4. 5. 6. 7.

Gender Age Tenure T1 trait anger T1 abusive supervision (Sb) T1 psychological distress (Sb) T1 supervisor-directed deviance (Sb) 8. T2 spouse undermining (Sp)

0.47 29.91 2.46 3.44 2.02 2.77 2.67 2.17

0.50 7.51 1.05 1.28 1.01 1.37 0.87 1.13

.04 .00 .04 .26 .12 .21 .23

.78 .04 .05 .08 .11 .09 Group level

.06 .01 .05 .09 .07

(.87) .16 .14 .20 .28

(.95) .49 .53 .42

(.91) .33 .37 (.67) .31 (.90)

1. Group size 2. T1 aggressive norms

4.48 2.37

1.01 0.97

.21

(.74) subordinate as source; Sp spouse 6 10 years, 4 1115 years, 5

Note. Reliability coefficients are shown on the diagonal in parentheses. T Time; Sup supervisor as source; Sb as source. Gender was coded as 0 male, 1 female. Tenure was coded as 1 less than a year, 2 15 years, 3 16 20 years, 6 2125 years, 7 26 30 years, and 8 more than 30 years. p .05. p .01. p .001.

model with several alternative models. As shown in Table 2, the five factor model had the best fit. Structural model. To test for our proposed structural model, we examined both the fully mediated and partially mediated models. Given that relationship-oriented occupation was measured using a single indicator, we set the path from the indicator variable equal to the square root of the reliability of the observed value (estimated to be .90; Mathieu, Tannenbaum, & Salas, 1992). In addition, the random error variance was set as one minus the

reliability of the observed value times the variance of the observed value. To test the fully mediated model, we specified paths from supervisor-reported aggressive norms to subordinate-reported abusive supervision, from subordinate-reported abusive supervision to subordinate psychological distress, from abusive supervision to supervisor-directed deviance, and from subordinate psychological distress to spouse ratings of undermining. Also, we specified paths from relationship-oriented occupation to both perceptions of abusive supervision and psychological distress. The hypothesized

Table 2 Confirmatory Factor Analyses of the Final Measurement Model and Alternative Models in Study 1
Alternative models Model 1 incorporated all five constructs into one factor. Model 2 combined abusive supervision and psychological distress into F1, aggressive norms into F2, and spouse undermining and supervisor-directed deviance into F3. Model 3 combined constructs based on source of measurement in which abusive supervision, psychological distress, and supervisor-directed deviance were combined into F1, and aggressive norms into F2, and spouse undermining into F3. Model 4 combined abusive supervision and psychological distress into F1, aggressive norms into F2, supervisor-directed deviance into F3, and spouse undermining into F4. Model 5 represented five independent factors. Note. CFI comparative fit index; TLI TuckerLewis index; RMSEA
2

df 88 81

CFI .93 .95

TLI .92 .93

RMSEA .11 .10

291.44 237.17

228.86 212.64 201.17

82 80 79

.95 .96 .96

.94 .95 .95 Factor.

.099 .095 .09

root mean square error of approximation; F

720

RESTUBOG, SCOTT, AND ZAGENCZYK

structural model (Model A) had a good fit, 2(98, N 184) 269.64, p .001, CFI .95, TLI .93, and RMSEA .09. All the predicted paths were significant. To examine the partially mediated structural model (Model B), we added several paths: (a) a path linking supervisor reported aggressive norms and subordinate psychological distress, (b) a path linking supervisor reported aggressive norms and spouse ratings family undermining, and (c) a path linking abusive supervision and spouse undermining. The direct paths were all nonsignificant. Fit statistics of the partially mediated model are as follows: 2(95, N 184) 265.90, p .001, CFI .95, TLI .93, and RMSEA .096. Supplementary analysis provided by the chi-square difference test suggested that Model B did not substan2 tively improve the fit compared with Model A, diff(3) 3.74, ns. We dropped the nonsignificant paths, and we accepted Model A as our final model. We also tested another alternative structural model (Model C) in which we included gender and age in the structural model because our correlation analyses revealed that these demographic characteristics were significantly related to the study variables. For example, we added a direct path linking subordinates gender and aggressive norms, perceptions of abusive supervision, spouse undermining, and supervisor-directed deviance. We also added a direct path linking subordinates age and psychological distress. This resulted in the following fit statistics: 2(126, N 184) 315.42, p .001, CFI .94, TLI .92, and RMSEA .095. The chi-square difference test suggest that Model A (final model) had 2 a significantly better fit than Model C, diff(28) 45.78, p .05. Finally, we conducted a Sobel test (Sobel, 1982) to assess the significance of the indirect effect. Results revealed that the two indirect paths were significant (i.e., path linking supervisor reported aggressive norms and subordinate psychological distress through subordinate abusive supervision, z 6.13, p .001, and the path linking subordinate reported abusive supervision and spouse ratings of undermining through subordinate psychological distress, z 3.94, p .001, were significant). Overall, we accepted Model A as our final model (see Figure 2). Collectively, these results suggest that (a) abusive supervision fully mediated

the relationship between aggressive norms and psychological distress, (b) psychological distress mediated the effects of abusive supervision on spouse ratings of undermining, and (c) abusive supervision was positively related to supervisor-directed deviance. Overall, Hypotheses 1, 2, and 3 were supported. Hypothesis 4a predicted that subordinates gender would moderate the relationship between abusive supervision and supervisordirected deviance. To test this, we conducted a moderated regression analysis. As shown in Table 3, entry of the interactive term did not explain additional variance in predicting supervisordirected deviance, R2 .00, F(1, 180) 1.20, .08, ns. Hypothesis 4a was not supported. Hypothesis 4b predicted that subordinates gender would moderate the relationship between psychological distress and spouse undermining. Entry of the interactive term contributed additional variance in predicting spouse undermining over and above the main effects, R2 .02, F(1, 180) 4.16, .19, p .05. The interaction accounted for an additional 2% of variance in spouse ratings of family undermining, thus falling within the typical range (i.e., .01 R2 .03) reported for moderator effects in nonexperimental studies (Champoux & Peters, 1987). We then examined the regression equations for both male (coded as 0) and female (coded as 1) subordinates representing minus and plus one standard deviation from the mean, respectively, to interpret the nature of the moderating effects. Figure 3 indicates that the pattern of interactions was as predicted in that the relationship between psychological distress and spousal undermining was stronger for male subordinates, t(180) 2.27 p .05, compared with female subordinates, t(180) 0.72, ns. Thus, Hypothesis 4b was supported. Hypothesis 5 predicted employees in relationship-oriented occupations are likely to report higher levels of abusive supervision and psychological distress than employees in other occupations. Results from the structural equation modeling suggest that employees in relationship-oriented occupations (path coefficient .18, p .01) perceived higher levels of abusive supervision compared with those in non-relationship-oriented occupations. Similarly, employees in relationship-oriented occupations (path coefficient .15, p .05) reported greater levels of psychological

Occupation

.15* .18**
T1 Supervisors assessments of aggressive work culture

.51***

T2 Subordinates perceptions of abusive supervision

.68***

T2 Subordinates psychological distress

.33***

T3 Spouse undermining

.62***

T2 Supervisor directed-deviance

Figure 2.

Final structural model in Study 1. T

Time. p

.05.

.01.

.001.

CONTEXTUAL FACTORS AND ABUSIVE SUPERVISION

721

Table 3 Hierarchical Regression Analysis for the Moderating Role of Gender for Study 1
Hypothesis 3a: Supervisor-directed deviance Variables Step 1 Gender Abusive supervision Psychological distress Abusive Supervision Gender Psychological Distress Gender R2 p .05. p .01. p .001. Step 1 .20 .54 Step 2 .19 .51 .08 .37 .00 .23 .19 .02 Hypothesis 3b: Spouse undermining Step 1 .39 .18 Step 2 .38 .15

distress compared with those in non-relationship-oriented occupations. Overall, Hypothesis 5 was supported.

Discussion
Overall, the results provide strong support for the predicted relationships. A potential methodological limitation in Study 1 is that aggressive norms were tested only at the individual level (on the basis of supervisor assessments) and not at the group level (e.g., shared perceptions of employees within a work environment). Indeed, previous researchers have noted that norms are contingent upon social perception processes operating within cohesive work groups (Cogliser & Schriesheim, 2000; Robinson & OLeary-Kelly, 1998; Schein, 1996). Thus, our operationalization of aggressive norms using supervisor-rated perceptions is potentially problematic and may not fully capture this construct. A second limitation is that we did not control for the personality characteristics of the subordinate. Martinko and Zellars (1998) have suggested that individual personality differences influence how individuals interpret situations and form their responses to them. One such class of individual differences may be aggression-

related personality characteristics (Berkowitz, 1993). For example, the extent to which an individual is likely to engage in spouse undermining behaviors may be influenced by ones predisposition to experience strong negative emotional states that may instigate aggressive behavior (Douglas & Martinko, 2001). Individuals generally high in trait anger are believed to interpret a wider variety of situations as anger-inducing than people who are low in trait anger (Spielberger, 1996; Wilkowski, Robinson, Gordon, & TroopGordon, 2007). Furthermore, such high-anger individuals are more likely to respond aggressively in such situations (Spielberger, 1996). In light of these potential limitations, we conducted Study 2 to replicate the findings in Study 1 and to address these methodological issues.

Study 2 Method
Participants and procedure. Data for this study were collected from restaurant employees in fast food stores and restaurant outlets in Manila, the capital city of the Philippines. The employees were crew members involved in service delivery, operations, and food preparation. The participating stores and restaurants had an average of 4.48 crew members, and the teams ranged between 3 and 9 members. Survey packets were distributed to participants by the research team. Each survey packet included a cover letter detailing the research objectives and indicating that participation in the study was confidential and voluntary. Participants were instructed to complete the survey individually and to use the postage-paid envelope to mail the survey directly to the research team. Upon completion of the surveys, participants (both employees and their spouses) received a small gift pack containing chocolates and biscuits. At Time 1, surveys were distributed to 350 fast food and restaurant employees at 42 restaurants. Employees were asked to assess the extent to which the norms in their organizations were aggressive and to report their levels of trait anger, perceptions of abusive supervision, and psychological distress. Surveys were administered only to crew members and not to the restaurant manager/supervisor in charge. We received a total of 291 surveys, yielding a response rate of 83.14%. At Time 2, 6 months after the initial data collection, we instructed participants to give brief

Males 5.0

Females

T3 Spouse undermining

4.0

3.0

2.0 Low T2 Psychological distress High

Figure 3. Interactive effects between subordinates gender and psychological distress in predicting spousal undermining in Study 1. T Time.

722

RESTUBOG, SCOTT, AND ZAGENCZYK

surveys to their spouses. To replicate the Study 1 findings, family undermining ratings from nonspouses or relatives (n 27) with whom participants live with (e.g., brother, sister, son/daughter, nephew or nice who was greater than 18 years of age) were not included in the analysis. We received a total of 195 surveys from spouses, representing a response rate of 67.01%. Thus, complete matching data were 188 employeespouse dyads. Within the employee group, there were 99 men and 89 women. Average age of the employees was 29.91 years. Of the employees, 60% had been working in their restaurants between 1 and 5 years. Among the spouses surveyed, 99 were women, and 89 were men. Mean age of spouses was 29.7 years. Measures. Aggressive norms. Aggressive norms were assessed at Time 1. Crew members within each participating fast food store and restaurant completed the four-item scale developed by Douglas and Martinko (2001) and adapted for use as described in Study 1. The coefficient alpha for this scale was .74. Subordinates perceptions of abusive supervision. Perceptions of abusive supervision were measured at Time 1. Like in Study 1, we used the 15-item scale developed by Tepper (2000). Individual crew members were asked to rate the extent to which their immediate supervisors had engaged in abusive treatment toward them. Cronbachs alpha was .95. Subordinates psychological distress. Psychological distress was measured at Time 1. As in Study 1, psychological distress was measured with the short four-item scale from Derogatis (1993). Cronbachs alpha for the scale was .91. Supervisor-directed deviance. At Time 1, supervisor-directed deviance was measured with the same three-item scale (Aquino et al., 1999) used in Study 1. Cronbachs alpha for this scale was .67. Spouse undermining. Family undermining was measured at Time 2 (6 months after Time 1 data collection). As in Study 1, we used the same scale derived from previous research (Duffy et al., 2002; Westman & Vinokur, 1998). Specifically, spouses reported the extent to which the focal employee had engaged in undermining behaviors. Cronbachs alpha was .90. Control variables. As in Study 1, we controlled for employees demographic characteristics. In addition, we controlled for the subordinates trait anger using the four-item scale developed by Spielberger (1996). For this study, the internal reliability coefficient of this scale was .87. We controlled for the effects of group size at the group level because the number of employees may have

an impact on the stress levels of the immediate supervisor in managing employee performance.

Results
Descriptive statistics, internal consistency reliabilities, and zeroorder correlations are provided in Table 1. Correlations were all in the expected direction. Confirmatory factor analysis. We conducted confirmatory factor analysis to establish the discriminant validity of the study variables. Item parcels were created to improve the ratio of N to the number of parameters to be estimated (Little et al., 2002). As in Study 1, the creation of the parcels was based on the results of the factor analysis. The hypothesized four-factor model fit the data very well, 2(141, N 188) 263.4, p .01, CFI .95, TLI .94, and RMSEA .068. As shown in Table 4, the five factor model had the best fit. Overall, these results support the distinctiveness of aggressive norms, abusive supervision, psychological distress, spouse undermining, and supervisor-directed deviance. Aggregation tests. To provide evidence for the aggregation of aggressive norms to the group level, it was imperative that we demonstrate both between-groups variability as well as withingroup agreement (Hofmann, 1997; Klein, Dansereau, & Hall, 1994). We first conducted a one-way analysis of variance and found between-groups variance for aggressive norms significant at p .001. We then obtained the following values of the intraclass correlations (ICCs): ICC(1) .55, and ICC(2) .86. These values were comparable with the prescribed values of group level constructs reported in the literature (Bliese, 2000). The median within group agreement (rwg) was .90, which is above the .60 threshold (James, Demaree, & Wolf, 1984). Overall, we found good support for the aggregation of aggressive norms at the group level. Hypotheses testing. Hierarchical linear modeling (HLM) was used to examine the predicted relationships in Hypotheses 1 and 2 because it is an appropriate data analytic approach to test for cross-level mediated models (Bryk & Raudenbush, 1992). This procedure partitions individual- and group-level variance in predicting individual outcomes. Two sets of mediation analysis were conducted. For each mediation analysis estimated, we entered the individual (focal study variables and control variables such as age, gender, tenure, and trait anger) and group level variables (aggressive norms and group size) after centering them around the grand

Table 4 Confirmatory Factor Analyses of the Final Measurement Model and Alternative Models in Study 2
Alternative models Model 1 incorporated all five constructs into one factor. Model 2 combined constructs based on source of measurement where aggressive norms, abusive supervision and psychological distress, and supervisor-directed deviance into F1, and spouse undermining into F2. Model 3 combined abusive supervision and psychological distress into F1, aggressive norms into F2, and spouse undermining and supervisor-directed deviance into F3. Model 4 combined abusive supervision and psychological distress into F1, aggressive norms into F2, supervisor-directed deviance into F3, and spouse undermining into F4. Model 5 represented five independent factors. Note. CFI comparative fit index; TLI TuckerLewis index; RMSEA
2

df 151 146 143 142 141

CFI .94 .94 .94 .95 .95

TLI .93 .93 .93 .94 .94 Factor.

RMSEA .077 .077 .075 .069 .068

324.8 309.6 293.79 269.45 263.4

root mean square error of approximation; F

CONTEXTUAL FACTORS AND ABUSIVE SUPERVISION

723

mean (Hofmann & Gavin, 1998). The results are presented in Table 5. In the first set of analysis, we tested the role of abusive supervision as an explanatory mechanism linking aggressive norms and subordinate perceptions of psychological distress. To establish mediation, three conditions have to be met (cf. Baron & Kenny, 1986; Hofmann & Stetzer, 1996). First, there should be a significant relationship between the independent variable (aggressive norms) and the dependent variable (psychological distress). To test this direct relationship, we ran an intercepts-as-outcomes model with aggressive norms entered as a group level variable in predicting psychological distress. Results from the analysis indicated that aggressive norms predicted psychological distress, coefficient .48, t(39) 2.67, p .01. The second condition for mediation requires that the independent variable should be associated with the mediator. Our results provided support for the positive relationship between aggressive norms and perceptions of abusive supervision, coefficient .37, t(39) 2.66, p .01. To provide support for the third condition, the relationship between the independent and dependent variable must disappear when including the mediator in the equation. When our mediating variable, perceptions of abusive supervision, was added to the model, it was positively related to psychological distress, coefficient .47, t(180) 4.25, p .01, and aggressive norms were no longer significant, coefficient .23, t(39) 1.80, ns. We then conducted Sobels (1982) test as a more stringent evaluation of the mediation effects. The results of these analyses showed that the effects of aggressive norms on psychological distress through perceptions of abusive supervision were significant (z 2.26, p .05). Overall, these results suggest that abusive supervision fully mediates the relationship between aggressive norms and psychological distress. Hypothesis 1 was supported. In the second set of analysis, we tested the mediating effect of psychological distress in the relationship between perceptions of abusive supervision and family undermining and supervisordirected deviance. Although our mediation hypothesis was framed at the individual level, we used HLM analysis to take into account the group-level nature of the data (e.g., data were nested in

restaurants/groups; Bryk & Raudenbush, 1992). Following the procedures specified by Baron and Kenny (1986) in testing mediation models, three equations were estimated. In support of the first condition for mediation, we found that abusive supervision at Time 1 significantly predicted spouse ratings of undermining at Time 2, coefficient .38, t(181) 4.31, p .01, as well as Time 1 supervisor-directed deviance, coefficient .41, t(181) 6.95, p .01 . In the second condition, we found that Time 1 subordinate perceptions of abusive supervision were positively associated with Time 1 subordinate perceptions of psychological distress, coefficient .56, t(181) 5.31, p .01. In the third condition, we regressed both subordinates perceptions of abusive supervision and subordinates psychological distress in the same equation for each dependent variable. For spouse undermining as dependent variable, results suggested a drop in the value of abusive supervision from .38 (p .01) to .29 (p .05), t(180) 2.42, p .05, whereas the effects of subordinates psychological distress, coefficient .17, t(180) 2.45, p .05, continued to be significant. As for supervisor-directed deviance, the effects of abusive supervision continued to be significant, coefficient .39, t(180) 6.19, p .01, whereas the effects of subordinates psychological distress were nonsignificant, coefficient .04, t(180) 0.98, ns. Results of Sobels (1982) test further revealed that the indirect path was significant (i.e., the path linking subordinates perceptions of abusive supervision and significant others ratings of family undermining through subordinates psychological distress, z 2.40, p .05). In addition, the indirect effect linking abusive supervision and supervisor-directed deviance through psychological distress was nonsignificant (z 0.98, ns). Overall, the results suggest that (a) perceptions of abusive supervision fully mediated the relationship between aggressive norms and psychological distress, (b) psychological distress mediated the effects of abusive supervision on spouse undermining, and (c) abusive supervision had direct positive effects on supervisor-directed deviance. Thus, Hypotheses 1, 2, and 3 were supported. Because age, tenure, and group size were not related to the dependent variables, we excluded these variables from the analysis. As shown in Table 6, results of the HLM analysis suggest that

Table 5 HLM Analysis of the Mediated Relationships in Predicting Supervisor-Directed Deviance and Spousal Undermining
Hypothesis 1 Model 1 Variables Level 1 Gender Age Tenure Trait anger AS PD Level 2 Intercept Group size AN AN 3 AS .19 .00 .02 .03 Model 2 AN 3 PD .12 .01 .10 .12 Model 3 AN & AS 3 PD .05 .01 .07 .09 .47 2.90 .08 .13 Model 1 AS 3 PD .01 .02 .07 .08 .55 2.91 .13 Model 2 AS 3 SU .27 .00 .04 .15 .38 2.07 .11 Hypothesis 2 Model 3 AS & PD 3 SU .28 .00 .05 .13 .29 .16 2.07 .11 Model 4 AS 3 SU .21 .01 .05 .08 .41 2.67 .02 Model 5 AS & PD 3 SU .20 .00 .05 .10 .39 .04 2.67 .03 spouse undermining.

2.02 .03 .37

2.89 .08 .41

Note. HLM hierarchical linear modeling; AN p .05. p .01.

aggressive norms; AS

abusive supervision; PD

psychological distress; SU

724

RESTUBOG, SCOTT, AND ZAGENCZYK

the interactive term involving abusive supervision and gender in predicting supervisor-directed deviance was not significant over and above the effects of trait anger, coefficient .18, t(183) 1.51, ns. Hypothesis 4a was not supported. In contrast, the interactive term representing psychological distress and gender was significant for spouse ratings of undermining after controlling for trait anger and main effects, coefficient .22, t(183) 2.01, p .05. Figure 4 depicts that the negative relationship between psychological distress and spouse ratings of family undermining was stronger for male subordinates, t(183) 3.17, p .01, as opposed to female subordinates, t(183) 2.16, p .05. Overall, Hypothesis 4b was supported.

Males 4.0

Females

T2 Spouse undermining

3.0

2.0

Discussion
1.0

Once again, the results supported the predicted relationships. Using a different and more rigorous operationalization of aggressive norms, we were able to demonstrate the top-down effects of aggressive norms on abusive supervision and its subsequent consequences for employees (Hypothesis 1). Also, we found corroborating evidence for the differential mediation effects of psychological distress on spouse undermining and supervisor-directed deviance. Specifically, psychological distress mediated the abusive supervision-spouse undermining relationship (Hypothesis 2), whereas abusive supervision had direct positive effects on supervisor-directed deviance (Hypothesis 3). As predicted, gender influenced the positive relationship between psychological distress and spouse undermining (Hypothesis 4b). Men as opposed to women were more likely to engage in spouse undermining as rated by their spouse. These findings add to the growing literature on the spill-over effects of abusive supervision in nonwork contexts. A limitation of Study 2 is that our measure of supervisor-directed deviance had low internal consistency. A plausible explanation for the low reliability is that the behavioral indicators that we have used to capture the construct may have been too broad and may not necessarily have been contextually applicable in the fast food industry. A critical incident technique (Flanagan, 1954) may be beneficial in the development of a supervisordirected deviance measure specifically customized for restaurant workers.

Low T1 Psychological distress

High

Figure 4. Interactive effects between subordinates gender and psychological distress in predicting spousal undermining in Study 2. T Time.

General Discussion Implications for Theory and Practice


Our study contributes to abusive supervision research in several important ways. For one, we offer an explanation for how aggressive norms contribute to the emergence of abusive supervisory behaviors and carry over to affect employees home lives. Researchers studying abusive supervision have focused primarily on individual level predictors and largely ignored group level effects. Recently, Tepper et al. (2008) provided initial evidence to support the role of norms in aggravating responses to abusive supervision. These researchers found that affective commitment mediated the relationship between abusive supervision and organizational devianceand that the link between commitment and deviance was stronger when employees perceived that deviance was a general norm in the organization. In the current article, we take a different perspective: We examined and provided evidence for how aggressive norms influence abusive supervision and ultimately affect subordinate employees and their spouses. Thus, the findings of the

Table 6 HLM Analysis of the Moderating Role of Gender in Study 2


Variables Level 1 Intercept Trait anger Gender Abusive supervision Psychological distress Abusive Supervision Gender Psychological Distress Gender Note. HLM p .05. hierarchical linear modeling. p .001. Hypothesis 3a: Supervisor-directed deviance 2.62 .08 .20 .41 .18 .22 Hypothesis 3b: Spouse undermining 2.14 .13 .37 .37

CONTEXTUAL FACTORS AND ABUSIVE SUPERVISION

725

present study suggest an additional plausible means of theorizing the link between aggressive norms and abusive supervision. That is, aggressive norms, in the long term, may have negative ramifications for employees and their families. In both studies, we found support for the mediating role of psychological distress on the relationship between abusive supervision and spousal undermining. Our results provide an initial explanation for the seemingly contradictory results of previous research, which in some cases suggests that employees are likely to retaliate against supervisors who are perceived as abusive (e.g., Mitchell & Ambrose, 2007; Tepper et al., 2009) and in other cases suggests that employees will displace aggression via family undermining (e.g., Hoobler & Brass, 2006), presumably to avoid the consequences of retaliation against the supervisor. Our research adds to recent work by Tepper et al. (2009), which showed that employees who planned to leave the organization were more apt to retaliate against supervisors than those who planned to maintain membership in the organization. Thus, the second contribution of our research is our use of the transactional model of stress (Lazarus & Folkman, 1984) to clarify the mechanisms underlying employee responses to abusive supervision. The third aspect of our theory building efforts concerned the role of gender. In their recent study, Hoobler and Brass (2006) found initial evidence that abusive supervision may result in family undermining, but they did not elaborate which individuals are likely to engage in such behaviors. Our results suggest that subordinates gender plays an important role in undermining behavior. In addition, we did not find support for the moderating role of gender between abusive supervision and supervisor-directed deviance. These findings should be interpreted within the context of our sample. In the Philippine context, there is a cultural emphasis on male machismo (Church, 1987) or masculinity, which has been associated with strength (Aguiling-Dalisay et al., 2000). As a social expectation for Filipino men, strength is exemplified in the ability to endure pain and project an aggressive masculinity (Aguiling-Dalisay et al., 2000). Given these cultural characteristics, we believe that when Filipino employees become targets of abusive treatment, they are less likely to retaliate against their supervisors because they respect and accept the power asymmetry inherent in the subordinatesupervisor relationship. However, because Filipino men assume a more dominant role in their families, they may feel less constrained from directing their aggression toward their spouse.2 Finally, occupational differences have been theorized as an important contextual factor that is likely to affect the extent to which employees perceive abusive supervisor behavior (Tepper, 2007) or experience victimization (Aquino & Thau, 2009). Although abusive supervision research has been conducted in various work contexts (Tepper, 2007), there has been no systematic attempt to explicitly assess the extent to which abusive supervision and psychological distress vary across groups. Indeed, in a critical appraisal of the victimization literature, Aquino and Thau (2009) noted, Our review of studies comparing various forms of victimization as a function of occupation or work sector does not present a clear conclusion about what types of organizations or job sectors are likely to be associated with higher victimization (p. 726). The results in our study suggest that employees in relationship-oriented occupations such as sales and marketing, customer service, and

public relations reported higher levels of abusive supervision and psychological distress than those in other occupations. This finding is useful for organizations, as it enables them to identify occupational groups in which employees are more likely to be exposed to abusive supervisory treatment and its resulting negative outcomes. We should also note that the current study differs in important ways from other studies that have examined similar models. For example, Haines, Marchand, and Harvey (2006) explored whether workplace aggression experienced by one or both members of a couple accounts for increases in psychological distress in the victims partner. They drew on Westman and Vinokurs (1998) reasoning, which suggests that crossover occurs through (a) common stressors (common stressors may affect both parties), (b) an indirect mediating interaction process (similar to what we propose; reciprocal interactions in the couple dyad; including coping strategies, communication patterns, demands for social support, and social undermining; a spillover model), and (c) a direct crossover through empathic reactions (feeling the partners pain). Haines et al. measured psychological distress of both couples when one member of the couple experienced workplace aggression and questioned whether aggression explains variance for the member of the couple who did not experience aggression when other factors are used as controls. They found that it did. We believe that our article extends this work by more formally specifying the mechanisms leading to workplace aggression (norms and abusive supervision) and by explicitly measuring the mechanisms (psychological distress) through which workplace aggression (specifically, abusive supervision) affects the other member of a couple (displaced aggression). Our research has implications for management practice. First, our findings are consistent with previous research that shows that norms play a powerful role in the occurrence of abusive supervisory behaviors. Hence, organizations should strive to prevent the incidence of abusive supervision by fostering clear guidelines and norms reflecting mutual respect, safety, support, sensitivity, and zero-tolerance for abusive behavior. Another implication of this research is that primary preventive measures must be implemented to curtail abusive supervisory behaviors. In a sample of 278 male university faculty and staff, Dekker and Barling (1998) found that perceived sanctions were associated with reduced harassment behaviors. This research informs us that the implementation of policies and procedures may curtail the enactment of these negative behaviors. Third, organizations should educate supervisors how to interpersonally manage their employees and should make them aware that their treatment of employees shapes subordinate perceptions of abuse and subsequent reactions. Fourth, the results of our research also demonstrate that employees are likely to experience psychological distress as a result of abusive supervision. Employee assistance programs (e.g., counseling) may be developed to help employees cultivate their coping skills in confronting these negative events and to better shield themselves from future hostile supervisory treatment.

We thank an anonymous reviewer for this insightful comment.

726

RESTUBOG, SCOTT, AND ZAGENCZYK

Limitations and Future Research Directions


Our article has several limitations that should be acknowledged. First, although we have attempted to minimize same-source bias by collecting independent assessments of behavior and by using a temporal research design (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003), our data, consistent with prior research on abusive supervision (e.g., Tepper, 2000; Tepper et al., 2008) and psychological distress (e.g., Major, Klein, & Erhart, 2002), were obtained from the same source (i.e., self-report). However, we were able to empirically distinguish these two constructs as separate and distinct across both studies, which bolsters confidence in our results and minimizes concerns about common method effects. Further, it is important to note that only these two variables as well as supervisor-directed deviance were vulnerable to percept-percept inflation. Aggressive norms were collected from different sources (e.g., supervisor assessments in Study 1 and collective assessments in Study 2), as was our family undermining variable, and thus should not be affected by common method variance. In addition, because these two variables were collected during the same time period in the study, inferences about the causal nature of these constructs cannot be assumed. We acknowledge that it is possible that employees experiencing negative emotions may provoke abusive supervisors (i.e., a precipitating victim effect; Tepper, Duffy, Henle, & Lambert, 2006) contrary to our research model. However, and consistent with our results, there is ample evidence to support the notion that stressful working conditions and, in particular, experiences with abusive supervision are likely to generate feelings of distress in employees (e.g., Tepper, 2007; Tepper, Duffy, & Shaw, 2001). Future research in this area may be warranted to further substantiate our findings. A second limitation is that the mean levels of abusive supervision and family undermining are low. However, previous research has reported similarly low levels for abusive supervision (on 5-point scales: 1.49 in Aryee, Sun, Chen, & Debrah, 2007; 1.39 in Tepper et al., 2008) and family undermining (on 5-point scales: 2.2 in Hoobler & Brass, 2006). In addition, an examination of the frequency distribution suggests that participants reported a wide range of responses in which some participants reported moderate to high mean levels of abusive supervision and family undermining. As for family undermining, the objectionable nature of this behavior may raise concerns regarding the likelihood that it is openly reported by the spouse. We should note, however, that although these behaviors may have low incidence or low visibility, and although participants are likely to be reluctant to report them, their impact has important implications for employee well-being and thus requires continued research attention. A third limitation is the possibility that the nature of the constructs explored in this research may have influenced participants to respond in a socially desirable manner. It occurs most often when aspects concerning social norms are questioned (Zerbe & Paulhus, 1987). Indeed, abusive supervision and family undermining reflect socially inappropriate behaviors. As perceptions of abusive supervision tend to be subjective, it is likely to be overstated or exaggerated by subordinates or under certain conditions, underreported. Perhaps future studies could obtain coworker and historical reports, particularly for the measurement of perceptions of abusive supervision.

Finally, that our data were obtained in the Philippines, a collectivist country where smooth interpersonal relationships and feelings of group belongingness are highly salient (Church, 1987; Restubog & Bordia, 2006), may raise questions regarding the generalizability of the study findings. Given the unacceptable nature of abusive supervisory behaviors and negative ramifications resulting from these behaviors, we are cautiously optimistic that the results may also apply in Asian (e.g., China; Aryee et al., 2007), European (e.g., Slovenia; Duffy et al., 2002), and North American (e.g., United States; Tepper, 2000) contexts. Our results offer several avenues for future research. First and foremost, we anticipate that this research will encourage additional theorizing and research on the effects of context on abusive supervision. Future research should examine other contextual conditions as potent facilitators of abusive supervision. For example, there may be value in examining organizational tolerance for aggression or the likelihood of reactions when employees themselves engage in various forms of abusive and hostile behaviors toward coworkers. Second, it is important for future research to unravel the motivational underpinnings that would explain why employees who were abused retaliate against their supervisors (e.g., supervisor-directed deviance). We speculate that employees who experienced abusive treatment are likely to ruminate revenge-seeking thoughts that may drive them to engage in supervisor-directed deviance. Revenge has been defined as an action in response to some perceived harm or wrongdoing by another party that is intended to inflict damage, injury, discomfort, or punishment on the party judged responsible (Aquino et al., 2001, p. 53). Although classified as a behavior, what makes aggressive acts vengeful is the desire to retaliate perceived harm by harming in return. Events associated with a damaged sense of self (e.g., public humiliation or ridicule) are likely to elicit the desire for revenge seeking (Aquino et al., 2001; Bies & Tripp, 1996). Indeed, it has been suggested that revenge may be used to restore equity or remedy perceived harm (Bies & Tripp, 1996). Finally, our operationalization of family undermining only depicts one actor as the sole source of the hostile treatment with the spouse as the recipient. However, research on family violence recognizes that in some cases the victim of the abuse might respond to her/his victimization by targeting other innocent parties or dependents (e.g., children: Appel & Holden, 1998; elderly family members: Wolf, 1997). That is, a spouse who was abused may use punitive child-rearing practices or lash out against elderly family members. An examination of the links between family undermining and its second-order effects is imperative if we are to better understand the disconcerting spill-over effects of abusive supervision in nonwork contexts.

References
Aguiling-Dalisay, G., Mendoza, R., Mirafelix, E., Yacat, J., Sto. Domingo, M., & Bambico, F. (2000). Masculinity: Men in control? Filipino male views on love, sex and women. Quezon City, Philippines: National Association of Filipino Psychology. Aguinis, H., Bommer, W. H., & Pierce, C. A. (1996). Improving the estimation of moderating effects using computer-administered questionnaires. Educational & Psychological Measurement, 56, 10431047. doi:10.1177/0013164496056006010

CONTEXTUAL FACTORS AND ABUSIVE SUPERVISION Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411 423. doi:10.1037/0033-2909.103.3.411 Appel, A., & Holden, G. (1998). The co-occurrence of spouse and physical child abuse: A review and appraisal. Family Psychology, 12, 578 599. doi:10.1037/0893-3200.12.4.578 Aquino, K., Douglas, S., & Martinko, M. J. (2004). Overt anger in response to victimization: Attributional style and organizational norms as moderators. Journal of Occupational Health Psychology, 9, 152164. Aquino, K., Lewis, M. U., & Bradfield, M. (1999). Justice constructs, negative affectivity, and employee deviance: A proposed model and empirical test. Journal of Organizational Behavior, 20, 10731091. Aquino, K., & Thau, S. (2009). Workplace victimization: Aggression from the targets perspective. Annual Review of Psychology, 60, 717741. doi:10.1146/annurev.psych.60.110707.163703 Aquino, K., Tripp, T. M., & Bies, R. J. (2001). How employees respond to personal offense: The effects of blame attribution, victim status, and offender status on revenge and reconciliation in the workplace. Journal of Applied Psychology, 86, 5259. doi:10.1037/0021-9010.86.1.52 Archer, J. (2002). Sex differences in physically aggressive acts between heterosexual partners: A meta-analytic review. Aggression and Violent Behavior: A Review Journal, 7, 313351. Aryee, S., Sun, L., Chen, Z. X., & Debrah, Y. A. (2007). Antecedents and outcomes of abusive supervision: Tests of a trickle-down model. Journal of Applied Psychology, 92, 191201. doi:10.1037/0021-9010.92.1.191 Azjen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179 211. Bandura, A. (1973). Aggression: A social learning analysis. Englewood Cliffs, NJ: Prentice Hall. Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ: Prentice Hall. Barling, J., Dupre, K. E., & Kelloway, E. K. (2009). Predicting workplace aggression and violence. Annual Review of Psychology, 60, 671 692. doi:10.1146/annurev.psych.60.110707.163629 Baron, R. M., & Kenny, D. A. (1986). The moderatormediator distinction in social psychological research: Conceptual, strategic, and statistical considerations, Journal of Personality and Social Psychology, 51, 1173 1182. doi:10.1037/0022-3514.51.6.1173 Berkowitz, L. (1993). Aggression: Its causes, consequences, and control. Philadelphia, PA: Temple University Press. Bernardo, A. B. I. (2004). McKinleys questionable bequest: Over 100 years of English in Philippine education. World Englishes, 23, 1731. doi:10.1111/j.1467-971X.2004.00332.x Bies, R. J., & Tripp, T. M. (1996). Beyond distrust: Getting even and the need for revenge. In R. M. Kramer & T. R. Tyler (Eds.), Trust in organizations (pp. 246 260). Newbury Park, CA: Sage. Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analyses. In K. J. Kelin & S. W. Kozlowski (Eds.), Multilevel theory, research and methods in organizations: Foundations, extensions, and new directions (pp. 349 381). San Francisco, CA: Jossey-Bass. Bograd, M. (1998). Feminist perspectives on woman abuse: An introduction. In K. Yllo & M. Bograd (Eds.), Feminist perspectives on women abuse (pp. 1126). New York, NY: Guilford Press. Bommer, W. H., Miles, E. W., & Grover, S. L. (2003). Does one good turn deserve another? Coworker influences on employee citizenship. Journal of Organizational Behavior, 24, 181196. doi:10.1002/job.187 Bordia, P., Restubog, S. L. D., & Tang, R. L. (2008). When employees strike back: Investigating the mediating mechanisms between psychological contract breach and workplace deviance. Journal of Applied Psychology, 93, 1104 1117. doi:10.1037/0021-9010.93.5.1104 Brotheridge, C. M., & Grandey, A. A. (2002). Emotional labor and burnout: Comparing two perspectives of people work. Journal of Vocational Behavior, 60, 1739. doi:10.1006/jvbe.2001.1815

727

Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models. Newbury Park, CA: Sage. Bushman, B. J., Baumeister, R. F., & Phillips, C. M. (2001). Do people aggress to improve their mood? Catharsis beliefs, affect regulation opportunity, and aggressive responding. Journal of Personality and Social Psychology, 81, 1732. doi:10.1037/0022-3514.81.1.17 Carstensen, L. L. (1992). Social and emotional patterns in adulthood: Support for socio-emotional selectivity theory. Psychology and Aging, 7, 331338. doi:10.1037/0882-7974.7.3.331 Champoux, J. E., & Peters, W. S. (1987). Form, effect size and power in moderated regression analysis. Journal of Occupational Psychology, 60, 243255. Church, A. T. (1987). Personality research in a non-Western culture: The Philippines. Psychological Bulletin, 102, 272292. doi:10.1037/00332909.102.2.272 Cogliser, C. C., & Schriesheim, C. A. (2000). Exploring work unit context and leadermember exchange: A multi-level perspective. Journal of Organizational Behavior, 21, 487511. doi:10.1002/1099-1379(200008) 21:5 487::AID-JOB57 3.0.CO;2-P Cooper, C. L., & Marshall, J. (1976). Occupational sources of stress: A review of the literature relating to coronary heart disease and mental ill health. Journal of Occupational Psychology, 49, 1128. Dekker, I., & Barling, J. (1998). Personal and organizational predictors of workplace sexual harassment of women by men. Journal of Occupational Health Psychology, 3, 718. doi:10.1037/1076-8998.3.1.7 Derogatis, L. R. (1993). BSI Brief Symptom Inventory: Administration, scoring, and procedures manual (4th ed.). Minneapolis, MN: National Computer Systems. Douglas, S. C., & Martinko, M. J. (2001). Exploring the role of individual differences in the prediction of workplace aggression. Journal of Applied Psychology, 86, 547559. doi:10.1037/0021-9010.86.4.547 Duffy, M. K., Ganster, D. C., & Pagon, M. (2002). Social undermining in the workplace. Academy of Management Journal, 45, 331351. doi: 10.2307/3069350 Duffy, M. K., Scott, K. L., & OLeary-Kelly, A. (2004). The radiating effects of intimate partner violence on occupational stress and wellbeing. In P. L. Perrewe & D. C. Ganster (Eds.), Research in occupational stress and well-being: Exploring theoretical mechanisms and perspectives (Vol. 4, pp. 6792). New York, NY: JAI Press/Elsevier Science. Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Erlbaum. Eagly, A. H., & Steffen, V. J. (1986). Gender and aggressive behavior: A meta-analytic review of the social psychological literature. Psychological Bulletin, 100, 309 330. doi:10.1037/0033-2909.100.3.309 Einarsen, S., & Skogstad, A. (1996). Bullying at work: Epidemiological findings in public and private organizations. European Journal of Work & Organizational Psychology, 5, 185201. doi:10.1080/ 13594329608414854 Erdogan, B., Liden, R. C., & Kraimer, M. L. (2006). Justice and leader member exchange: The moderating role of organizational culture. Academy of Management Journal, 49, 395 406. Fishbein, M., & Azjen, I. (1975). Beliefs, attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Flanagan, J. C. (1954). The critical incident technique. Psychological Bulletin, 51, 327358. doi:10.1037/h0061470 Garcia, P. R. J. M., Restubog, S. L. D., & Denson, T. (2010). The moderating role of prior exposure to aggressive home culture in the relationship between negative reciprocity and aggression. Journal of Research in Personality, 44, 380 385. doi:10.1016/j.jrp.2010.04.003 Glomb, T. M., & Liao, H. (2003). Interpersonal aggression in work groups: Social influence, reciprocal, and individual effects. Academy of Management Journal, 46, 486 496. doi:10.2307/30040640 Grandey, A. A. (2000). Emotion regulation in the workplace: A new way

728

RESTUBOG, SCOTT, AND ZAGENCZYK violence and aggression: A cognitive appraisal perspective. In R. W. Griffin, A. OLeary-Kelly, & J. M. Collins (Eds.), Dysfunctional behavior in organizations: Violent and deviant behavior (pp. 1 42). Stamford, CT: JAI Press. Mathieu, J. E., Tannenbaum, S. I., & Salas, E. (1992). Influences of individual and situational characteristics on measures of training effectiveness. Academy of Management Journal, 35, 828 847. doi:10.2307/ 256317 Mitchell, M. S., & Ambrose, M. L. (2007). Abusive supervision and workplace deviance and the moderating effects of negative reciprocity beliefs. Journal of Applied Psychology, 92, 1159 1168. doi:10.1037/ 0021-9010.92.4.1159 OLeary-Kelly, A. M., Griffin, R. W., & Glew, D. J. (1996). Organizationmotivated aggression: A research framework. Academy of Management Review, 21, 225253. doi:10.2307/258635 Pagelow, M. (1981). Women battering: Victims and their experiences. Beverly Hills, CA: Sage. Pagelow, M. (1984). Family violence. New York, NY: Praeger. Pearson, C. M. I., Anderson, L. M., & Porath, P. L. (2005). Workplace incivility. In S. Fox & P. E. Spector (Eds.), Counterproductive work behavior (pp. 177200). Washington, DC: American Psychological Association. doi:10.1037/10893-008 Podsakoff, P. M., MacKenzie, S. B., Lee, J., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879 903. doi:10.1037/0021-9010.88.5.879 Restubog, S. L. D., & Bordia, P. (2006). Workplace familism and psychological contract breach in the Philippines. Applied Psychology, 55, 563585. doi:10.1111/j.1464-0597.2006.00245.x Robinson, S. L., & OLeary-Kelly, A. M. (1998). Monkey see, monkey do: The influence of work groups on the antisocial behavior of employees. Academy of Management Journal, 41, 658 672. doi:10.2307/256963 Schein, E. H. (1996). Culture: The missing concept in organizational studies. Administrative Science Quarterly, 41, 229 240. doi:10.2307/ 2393715 Selye, H. (1974). Stress without distress. Philadelphia, PA: Lippincott. Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. Sociological Methodology, 13, 290 312. doi:10.2307/270723 Spielberger, C. D. (1996). State-Trait Anger Expression Inventory: Professional manual. Odessa, FL: Psychological Assessment Resources. Tepper, B. J. (2000). Consequences of abusive supervision. Academy of Management Journal, 43, 178 190. doi:10.2307/1556375 Tepper, B. J. (2007). Abusive supervision in work organizations: Review, synthesis, and research agenda. Journal of Management, 33, 261289. doi:10.1177/0149206307300812 Tepper, B. J., Carr, J. C., Breaux, D. M., Geider, S., Hu, C., & Hua, W. (2009). Abusive supervision, intentions to quit, and employees workplace deviance: A power/dependence analysis. Organizational Behavior and Human Decision Processes, 109, 156 167. doi:10.1016/ j.obhdp.2009.03.004 Tepper, B. J., Duffy, M. K., Henle, C. A., & Lambert, L. S. (2006). Procedural injustice, victim precipitation, and abusive supervision. Personnel Psychology, 59, 101123. doi:10.1111/j.1744-6570.2006 .00725.x Tepper, B. J., Duffy, M. K., & Shaw, J. D. (2001). Personality moderators of the relationships between abusive supervision and subordinates resistance. Journal of Applied Psychology, 86, 974 983. doi:10.1037/ 0021-9010.86.5.974 Tepper, B. J., Henle, C. A., Lambert, L. S., Giacalone, R. A., & Duffy, M. K. (2008). Abusive supervision and subordinates organizational deviance. Journal of Applied Psychology, 93, 721732. doi:10.1037/ 0021-9010.93.4.721 Westman, M., & Vinokur, A. D. (1998). Unraveling the relationship

to conceptualize emotional labor. Journal of Occupational Health Psychology, 5, 95110. doi:10.1037/1076-8998.5.1.95 Green, S. G., Anderson, S. E., & Shivers, S. L. (1996). Demographic and organizational influences on leadermember exchange and related work attitudes. Organizational Behavior and Human Decision Processes, 66, 203214. doi:10.1006/obhd.1996.0049 Haines, V. Y., Marchand, A., & Harvey, S. (2006). Crossover of workplace aggression experiences in dual-earner couples. Journal of Occupational Health Psychology, 11, 305314. doi:10.1037/1076-8998.11.4.305 Hershcovis, M. S., Turner, N., Barling, J., Arnold, K. A., Dupre, K. E., Inness, M., . . . Sivanathan, N. (2007). Predicting workplace aggression: A meta-analysis. Journal of Applied Psychology, 92, 228 238. doi: 10.1037/0021-9010.92.1.228 Hofmann, D. A. (1997). An overview of the logic and rationale of hierarchical linear models. Journal of Management, 23, 723744. doi: 10.1177/014920639702300602 Hofmann, D. A., & Gavin, M. B. (1998). Centering decisions in hierarchical linear models: Implications for research in organizations. Journal of Management, 24, 623 641. Hofmann, D. A., & Stetzer, A. (1996). A cross-level investigation of factors influencing unsafe behaviors and accidents. Personnel Psychology, 49, 307339. doi:10.1111/j.1744-6570.1996.tb01802.x Hoobler, J. M., & Brass, D. J. (2006). Abusive supervision and family undermining as displaced aggression. Journal of Applied Psychology, 91, 11251133. doi:10.1037/0021-9010.91.5.1125 Hubert, A. B., & van Veldhoven, M. (2001). Risk sectors for undesirable behaviour and mobbing. European Journal of Work & Organizational Psychology, 10, 415 424. doi:10.1080/13594320143000799 Ibarra, H. (1999). Provisional selves: Experimenting with image and identity in professional adaptation. Administrative Science Quarterly, 44, 764 791. James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group inter-rater reliability with and without response bias. Journal of Applied Psychology, 69, 8598. doi:10.1037/0021-9010.69.1.85 Johnson, S., Cooper, C., Cartwright, S., Donald, I., Taylor, P., & Millet, C. (2005). The experience of work-related stress across occupations. Journal of Managerial Psychology, 20, 178 187. doi:10.1108/ 02683940510579803 Keashly, L., & Harvey, S. (2005). Emotional abuse in the workplace. In S. Fox & P. E. Spector (Eds.), Counterproductive work behavior: Investigations of actors and targets (pp. 201235). Washington, DC: American Psychological Association. doi:10.1037/10893-009 Klein, K. J., Dansereau, F., & Hall, R. I. (1994). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19, 195229. doi:10.2307/258703 Lazarus, R. S. (1991). Emotion and adaptation. New York, NY: Oxford University Press. Lazarus, R. S. (1993). From psychological stress to the emotions: A history of changing outlooks. Annual Review of Psychology, 44, 122. doi: 10.1146/annurev.ps.44.020193.000245 Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York, NY: Springer. Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to parcel: Exploring the question, weighing the merits. Structural Equation Modeling, 9, 151173. doi:10.1207/ S15328007SEM0902_1 Major, V. S., Klein, K. J., & Ehrhart, M. G. (2002). Work time, work interference with family, and psychological distress. Journal of Applied Psychology, 87, 427 436. doi:10.1037/0021-9010.87.3.427 Marcus-Newhall, A., Pederson, W. C., Carlson, M., & Miller, N. (2000). Displaced aggression is alive and well: A meta-analytic review. Journal of Personality and Social Psychology, 78, 670 689. doi:10.1037/00223514.78.4.670 Martinko, M. J., & Zellars, K. (1998). Toward a theory of workplace

CONTEXTUAL FACTORS AND ABUSIVE SUPERVISION distress levels within couples: Common stressors, empathic reactions, or crossover via social interaction? Human Relations, 51, 137156. doi: 10.1177/001872679805100202 Wilkowski, B. M., Robinson, M. D., Gordon, R. D., & Troop-Gordon, W. (2007). Tracking the evil eye: Trait anger and selective attention within ambiguously hostile scenes. Journal of Research in Personality, 41, 650 666. doi:10.1016/j.jrp.2006.07.003 Wolf, R. S. (1997). Elder abuse and neglect: Causes and consequences. Journal of Geriatric Psychiatry, 30, 153174. Zagenczyk, T. J., Gibney, R., Murrell, A. J., & Boss, S. R. (2008). Friends

729

dont make friends good citizens, but advisors do. Group and Organization Management, 33, 760 780. doi:10.1177/1059601108326806 Zerbe, W. J., & Paulhus, D. L. (1987). Socially desirable responding in organizational behavior: A reconception. Academy of Management Review, 12, 250 264. doi:10.2307/258533

Received April 17, 2009 Revision received June 15, 2010 Accepted July 21, 2010

Correction to OBoyle and Williams (2011)


In the article Decomposing Model Fit: Measurement vs. Theory in Organizational Research Using Latent Variables, by Ernest H. OBoyle, Jr., and Larry J. Williams (Journal of Applied Psychology, 2011, Vol. 96, No. 1, pp. 112), the sample size of Study 18 in Table 1 (p. 6) is incorrect. The sample size for Study 18 should have been 173, not 161. This changes the RMSEA-P and confidence interval from what were originally reported as .083 [.000, .124] to .079 [.000, .119]. Thus, Study 18 should have been denoted with a superscript a, indicating that it met the .08 threshold of path model fit.
DOI: 10.1037/a0024273

S-ar putea să vă placă și