Articles Review
® Academy of Management Journal 2000, Vol. 43. No. 2. 178-190.
CONSEQUENCES OE ABUSIVE SUPERVISION
BENNETT J. University-of'Keritucky
Drawing on justice theory, I examined the consequences of abusive supervisor behav- ior. As expected, subordinates who perceived their supervisors were more abusive were more likely to quit their jobs. For subordinates who remained with their jobs, abusive supervision was associated witb lower job and life satisfaction, lower norma- tive and affective commitment, and higher continuance commitment, conflict between work and family, and psychological distress. Organizational justice mediated most of these effects, and job mobility moderated some of the deleterious effects of abusive supervision.
What did I tell you the first day? Your thoughts are nothing; you are nothing . . . if you were in my toilet bowl I wouldn't bother flushing it. My bath mat means more to me than you . . . you don't like it here, leave!
You see this watch? That watch costs more than your car. I made $970,000 last year, how much did you make? You see pal, that's who I am, and you're nothing. Nice guy? I don't give a ####. Good father? #### you, go home and play with your kids . . . you think this is abuse, you ####? You don't like it? Leave!
These excerpts from the dialogue in two films, the first George Huang's 1994 Swimming with Sharks, and the second David Mamet's 1984 Glen- garry Glen Ross, vividly illustrate what may be referred to as abusive supervision, a manifestation of dysfunctional workplace behavior that has cap- tured the attention of academic researchers (e.g., Bies & Tripp, 1998; Keashly, Trott, & MacLean, 1994). The notion of abusive supervision evokes images of a tyrannical boss wbo publicly ridicules and undermines tbose reporting to him or her (Ash- forth, 1994). However, despite anecdotal evidence suggesting abusive supervision is ubiquitous and has implications for subordinates' performance, at-
This research was supported by a Summer Research Grant from the Gatton College of Business and Econom- ics of the University of Kentucky. The grant was made possible by a donation of funds to the College by Ashland Oil, Inc. I thank the following individuals for their assis- tance during the data collection and processing phases of the research reported in this article: Rhea Ingram, Daniel Nehring, Robert Nelson, Erica Newton, Edward Taylor, Nicole Terrain, Jackie Thompson, and Annabelle Van't Klooster. I also thank Robert Folger, Chester Schriesheim, and three anonymous AMf reviewers for comments on earlier versions.
titudes, and psychological health (Hornstein, 1996), there has been little theory-based study of the construct. Accordingly, in this research I devel- oped and tested a model of the consequences of abusive supervision. In tbe following sections, I identify the construct's behavioral domain and in- voke the organizational justice literature to explain the effects of abusive supervision on a variety of outcome variables.
ABUSIVE SUPERVISION
For this study, abusive supervision refers to sub- ordinates' perceptions of tbe extent to which super- visors engage in the sustained display of hostile verbal and nonverbal behaviors, excluding physi- cal contact. This definition characterizes abusive supervision as a subjective assessment. The same individual could view a supervisor's behavior as abusive in one context and as nonabusive in an- other context, and two subordinates could differ in their evaluations of the same supervisor's behavior. Like abuse directed toward intimate partners, the elderly, and children, abusive supervision can be characterized as sustained or enduring in the sense that it is likely to continue until (1) the target ter- minates the relationship, (2) the agent terminates the relationship, or (3) the agent modifies his or her behavior (Jezl, Molidor, & Wright, 1996; Shepard & Campbell, 1992). Several features of abusive rela- tionships contribute to their enduring quality. First, targets of abuse may remain in the relation- ships because they feel powerless to take corrective action, are economically dependent on the abusers, or fear the unknown associated with separation more than they fear the abuse, and they may remain because the agents ofren intersperse abusive behav- ior with normal bebavior, in effect intermittently reinforcing the targets' hope the abuse will end (Walker, 1979). Second, because abusers often fail
2000 Tepper 179
to recognize or take responsibility for their abusive behavior, few modify it, and, in many cases, even clinical intervention fails to recast such relation- ships as nonabusive (Wolfe, 1987).
Bies (2000; Bies & Tripp, 1998) identified the following manifestations of abusive supervision: public criticism, loud and angry tantrums, rude- ness, inconsiderate actions, and coercion. Similar themes appear in Ashforth's (1994) description of organizational "petty tyranny," Neuman and Bar- on's (1997) examples of nonphysical workplace ag- gression, and Robinson and Bennett's (1995) typol- ogy of deviant organizational behavior. However, although abusive supervision shares conceptual overlap with these constructs, there are meaningful distinctions that w^arrant treating abusive supervi- sion as a separate construct. For example, Ash- forth's (1994: 757) conceptual and operational def- initions of petty tyranny capture management practices that may or may not involve hostility (for example, "uses authority or position for personal gain," "administers organizational policies unfair- ly," and "discourages initiative"). Second, diverg- ing from Neuman and Baron's (1997) characteriza- tion of aggression as behaviors designed to cause harm, abusive supervision refers to behaviors that refiect indifference (for example, speaking rudely to subordinates in order to elicit desired task per- formance), as well as willful hostility (publicly be- littling subordinates in order to hurt their feelings). Finally, abusive supervision may not be deviant if it conforms with an organization's policies or norms; examples of such conformity are addressed below.
A few systematic studies bave investigated the effects of behavior akin to abusive supervision. Ashforth (1997) found that tyrannical supervision (which included belittling subordinates, displaying little consideration, and using noncontingent pun- ishment) was associated with frustration, helpless- ness, and alienation from work. Keashly and col- leagues (1994) found that nonphysical abuse occurred more frequently than physical violence like throwing things, punching, or threatening with a weapon and that individuals who experienced more supervisory abuse were less satisfied witb their jobs. Studies of medical students and resi- dents suggest abusive supervision is associated with dissatisfaction and elevated levels of psycho- logical distress (Richman, Flaherty, Rospenda, & Christensen, 1992; Sheehan, Sheehan, White, Lei- bowitz, & Baldwin, 1990). Taken together, these . studies suggest nonphysical, abusive supervisor be- baviors negatively infiuence subordinates' work- related attitudes and psychological health (Duffy, Ganster, & Pagon, 1998).
A Justice-Based Model of Subordinates' Responses to Abusive Supervision
According to justice theory, individuals' evalua- tive assessments of fairness draw on perceptions of distributive justice (fairness of outcome allocation), procedural justice (fairness of the procedures used to make allocation decisions), and interactional jus- tice (fairness of the interpersonal treatment indi- viduals receive during the enactment of proce- dures). I expected that the degree to which supervisors engaged in abusive behavior would af- fect subordinates' perceptions of organizational justice, which would, in turn, affect their decisions to quit, job satisfaction, life satisfaction, organiza- tional commitment, conflict between work and family life, and psychological distress. I further expected that perceived mobility—the extent to which employees feel they have attractive employ- ment alternatives—would moderate the effects of abusive supervision. More specifically, I expected the effects of abusive supervision would be more pronounced for individuals who have less mobil- ity.
Interactional Justice. Interactional justice is par- ticularly relevant to this analysis because it refiects the interpersonal dimension of fairness (Bies, 2000). According to Bies and Moag (1986), individ- uals experience interactional injustice when orga- nizational representatives fail to treat them with respect, honesty, propriety, and sensitivity to their personal needs. Contextual factors infiuence whether or not subordinates experience unfairness when they are the targets of behaviors that fit the present definition of abusive supervision; as Bies and Moag stated, "A person may bold a criterion such as personal respect inviolable . . . bowever, if rudeness is seen as an expected part of the proce- dure itself, as in a stress interview, then it may not be perceived as unfair because there is an instru- mental purpose to its occurrence" (1986: 51). Sim- ilarly, a drill instructor may be expected to use verbal battery and degradation as part of the pro- cess of divesting military recruits of the values they held prior to boot camp (Van Maanen & Schein, 1979). Nevertheless, beyond a narrow range of con- texts in which hostility may be tolerated, individ- uals expect others, especially those of higher status, to be aware of communicative acts that constitute face threats, actions that threaten one's social image and self-image (Goffman, 1967). Consistent with this notion is Mikula, Petri, and Tanzer's (1990) inductively derived typology of everyday experi- ences of injustice, which emphasizes interactional concerns and includes several categories of bebav- ior directly relevant to the present focus on abusive
180 Academy of Management Journal April
supervision (such as unfriendliness and impolite- ness on the part of authority figures). Consequently, suhordinates should experience interactional injus- tice when their supervisors are more ahusive.
Procedural justice. Suhordinates who experi- ence long-term ahusive supervision may conclude that their organization has not done an adequate joh of developing or enforcing procedures that disci- pline ahusers or protect targets of ahuse. These conclusions might imply that one or more of the procedural justice rules Leventhal (1980) descrihed has heen violated; these rules are that organiza- tional procedures should he consistent across peo- ple and over time, not he hiased hy self-interest, he based on accurate information, include provisions for appeal, and reflect the concerns and ethical system of those affected. For example, the hias sup- pression and ethicality rules would he violated if suhordinates helieved their organization was un- willing to discipline an ahusive hut high-perform- ing manager out of fear that the hottom line would he negatively affected. The accuracy rule would he violated if suhordinates helieved that organiza- tional representatives had not done a good joh of collecting the information necessary to adequately monitor or document supervisory hehaviors that might warrant disciplinary action. Consequently, suhordinates should experience procedural injus- tice when their supervisors are more ahusive.
Distributive justice. Theories of distributive jus- tice suggest individuals make fairness judgments when they compare their inputs and outcomes with those of a referent (Adams & Freedman, 1976). Sub- ordinates of abusive supervisors may experience what Martin (1981) referred to as relative depriva- tion: the helief that they are getting less than they deserve compared to target referents. For example, subordinates of abusive supervisors may feel dis- advahtaged compared to peers if their supervisors spend more time berating them than providing the mentoring functions that prepare junior colleagues for advancement (Tepper, 1995). Ahusive supervi- sion may also influence suhordinates' perceptions of the inputs that figure in their evaluations of distrihutive justice. For example, suhordinates of abusive supervisors may have to overcome ohsta- cles that increase the time and effort required to fulfill their responsibilities; for instance, gathering needed information from a superior who is more inclined to provide criticism than constructive feedback will take longer. In addition, because peo- ple are likely to see unfair procedures as producing unfair outcomes—a phenomenon known as the fair process effect (Greenberg, 1990)—the procedural injustices suhordinates of ahusive supervisors ex- perience (see the preceding discussion) may result
in the perception that their outcomes are not as favorable as those of suhordinates who do not have abusive supervisors. Given these possihilities, suh- ordinates should experience distributive injustice when their supervisors are more ahusive.
Mediating Effects of Organizational Justice
Workplace injustices cause frustration, threaten employees' self- and social images, and, in some circumstances, produce moral outrage (Greenherg, 1990). Hence, just as the injustices associated with drug-testing programs (Konovsky & Cropanzano, 1991), selection systems (Gilliland, 1994), pay raise decisions (Folger & Konovsky, 1989), and other or- ganizational phenomena foster job dissatisfaction and voluntary turnover (see Aquino, Griffeth, Allen, & Hom, 1997), the perceived injustices re- sulting from abusive supervision are likely to translate into dislike for a job and to prompt a subordinate to seek out and obtain alternative em- ployment. In addition, because for most people, work plays a significant role in terms of time, emo- tional involvement, fulfillment, and self-esteem (Judge, Boudreau, & Bretz, 1994), I expected that the perceived injustices produced hy negative work experiences like abusive supervision would trans- late into dissatisfaction with life. These arguments suggest the following hypotheses:
Hypothesis 1. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their voluntarily leav- ing their jobs.
Hypothesis 2. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their job satisfaction.
Hypothesis 3. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their life satisfaction.
Economic or logistical considerations may make some individuals unwilling to quit their jobs, even if their supervisors' behavior makes those johs un- desirahle. For them, the injustices evoked by abu- sive supervision should translate into what Meyer and Allen (1991) referred to as continuance com- mitment: an attachment to an organization based primarily on need. Moreover, individuals who ex- perience injustice are likely to feel that their em- ploying organizations do not value their contribu- tions or care about their circumstances (Konovsky & Cropanzano, 1991). These suhordinates are un- likely to feel an obligation to remain with the orga- nizations (normative commitment), nor are they likely to develop an emotional attachment to, or a
2000 Tepper 181
sense of identification with, their organizations (af- fective commitment). These arguments suggest the following hypotheses:
Hypothesis 4a. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their continuance commitment.
Hypothesis 4b. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their normative com- mitment.
Hypothesis 4c. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their affective com- mitment.
Conflict between work and family is the degree to which participation in one area (work or family) is made more difficult by participating in the other (Greenhaus & Beutell, 1985). The injustices associ- ated with ahusive supervision may cause subordi- nates to be so preoccupied with work-related matters that the time they spend with their families is ad- versely affected (they experience work-to-family con- flict). In addition, abusive supervisors may try to con- vince subordinates that their familial obligations are interfering with work rather than the other way around. The excerpt from Glengarry Glen Ross at the heginning of this article, in which a malevolent supe- rior characterizes "being a good father" as incompat- ihle with doing the job well, is illustrative. To the extent suhordinates accept such argimients or find that performing routine family duties, such as drop- ping off and picking up their children, leaving work to care for a sick child, or participating in family functions over the weekend, triggers hostile reactions from their supervisors, they are likely to report greater family-to-work conflict. These arguments suggest the following hypotheses:
Hypothesis 5a. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their work-to-family conflict.
Hypothesis 5b. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their family-to-work conflict.
In the justice literature, several processes hy which work-related sources of injustice can under- mine individuals' sense of self-worth and engender anxiety and feelings of helplessness, or psycholog- ical distress, have been identified. For example, in clarifying the nature of inequity distress, Adams
and Freedman (1976) recognized that a person's reaction to inequity could not only involve anger and a desire for restitution (the outcomes ordinarily associated with disadvantaged inequity), but could also involve damaged self-esteem when a target feels an agent's hehavior constitutes a personal and intentional attack. Several more recent justice the- ories are pertinent: according to the self-interest model of procedural justice, individuals will expe- rience psychological distress as a consequence of heing denied voice in decisions that affect them (Lind & Tyler, 1988); in the group value model of procedural justice, such distress would stem from being denied voice when voice would facilitate the feeling of connection with a valued group (Lind & Tyler, 1988); and according to referent cognitions theory, psychological distress stems from individ- uals' believing they would have obtained more fa- vorable outcomes if a decision maker had used different procedures (Folger, 1986). Two further studies support a link hetween the experience of organizational injustice and relatively mild mani- festations of psychological discomfort, one involv- ing employees' reactions to criticism (Baron, 1993) and the other involving employees' responses to hiring outcomes and procedures (Gilliland, 1994). These arguments suggest the following hypothesis:
Hypothesis 6. Subordinates' justice percep- tions will mediate the relationship between abusive supervision and their psychological distress.
Moderating Effects of Perceived Mobility
Empirical research suggests aversive stimuli are less stressful when individuals have greater con- trol, or a helief that they can escape from, avoid, or mitigate the impact of the stimuli—whether or not they actually have an opportunity to do so (Miller, 1980). For example. Bowers (1968) found that ex- perimental subjects who could avoid electric shock were able to tolerate more intense shock than suh- jects who could not avoid the shock. According to Miller, potential control allows targets to set an upper limit on the discomfort an aversive stimulus engenders, therehy alleviating some of their dis- tress. Hence, ahusive supervision should be less distressing when suhordinates have more attractive employment alternatives and feel they can separate themselves from the source of their distress (their supervisors' ahusive behavior). These arguments suggest the final hypothesis:
Hypothesis 7. The consequences of abusive su- pervision will be more pronounced when sub- ordinates have less job mobility.
182 Academy of Management Joumal April
METHODS
Sample and Procedures
I tested the hypotheses with data collected at two points in time. The substantive component of the time 1 survey contained measures of abusive super- vision, perceived job mobility, interactional justice, procedural justice, and distributive justice. At time 2, I measured subordinates' voluntary turnover (quitting their jobs), job satisfaction, life satisfac- tion, organizational commitment, conflict between work and family, and psychological distress.
Time 1 sample €md procedure. At time 1, two research assistants called 2,415 residents of a medium-sized midwestem city using random-digit dialing. Individuals who were employed full-time and had hosses were invited to participate in the study. Of those contacted, 1,073 people were not el- igihle for the study (they were imiemployed or not formally supervised at work), 278 refused to partici- pate, and 1,064 eligible individuals agreed to partic- ipate. Of these, 741 individuals returned surveys via business reply envelopes, 712 of which were com- pleted, producing a usahle response rate of 53 per- cent. The time 1 survey was coded so that the data could be matched with surveys completed at time 2.
Time 2 sample and procedures. The second wave of data collection took place six months after the first wave. Phone calls were made to the indi- viduals who responded to the time 1 survey, and individuals who still had the same supervisors were invited to complete follow-up surveys. Eighty-one individuals could not he located, 177 were not eligible (they or their supervisors had left their johs), 8 were eligihle but refused to partici- pate, and 475 were eligible and agreed to partici- pate. Three-hundred ninety individuals returned surveys via business reply envelopes, 362 of which contained no missing data.
Of the individuals who completed the time 1 and time 2 surveys, 43 percent were women, and 35 percent had female supervisors. The respondents' median age was between 35 and 39 years old, and their supervisors' median age was between 40 and 44 years old. Respondents were employed in the following industries: 42 percent, service; 8 percent, retailing; 12 percent, manufacturing; 11 percent, small husiness; 11 percent, government; and 16 percent, education. In terms of its memhers' back- ground characteristics, the sample was representa- tive of the area from which it was drawn.
ments that capture nonphysical abuse in other kinds of relationships (e.g., Raymond & Bruschi, 1989; Shepard & Gampbell, 1992; Tolman, 1989) and on recent publications in the management lit- erature that describe actions that fit the behavioral domain of nonphysical, ahusive supervision (e.g., Hornstein, 1996; Rohinson & Bennett, 1995). I con- tent-analyzed an initial pool of 20 items using pro- cedures described by Schriesheim and Hinkin (1990). I created a checklist by interspersing these items with 20 items adapted from measures of physical ahuse in spousal and dating relationships. I then administered the checklist to 68 master's of business administration (M.B.A.) degree candi- dates, with instructions to classify the items ac- cording to three categories: nonphysical abuse, physical ahuse, and other (nonahusive hehavior or another kind of abuse). The M.B.A. students put a 1 in the category that best described the item and, if they felt more than one category descrihed the item, put a 2 in the second-hest category. For computa- tional purposes, I recoded the students' ratings, with Is receiving two points and 2s receiving one point. I retained items from the nonphysical abuse scale that received 70 percent or more of their pos- sihle points. These 15 items appear in the Appen- dix. Respondents in my main study indicated the frequency with which their supervisors engaged in each of the 15 behaviors using the following re- sponse scale: 1, "I cannot remember him/her ever using this hehavior with me"; 2, "He/she very sel- dom uses this hehavior with me"; 3, "He/she occa- sionally uses this behavior with me"; 4, "He/she uses this hehavior moderately often with me"; and 5, "He/she uses this behavior very often with me."
Perceived mobility. I measured perceived mobil- ity with the following two items: "If I were to quit my joh, I could find another job that is just as good" and "I would have no problem finding an accept- able job if I quit" (1 = strongly disagree, 5 = strongly agree).
Organizational justice. I measured the three di- mensions of organizational justice using 12 items adapted from Niehoff and Moorman (1993). Respondents reported their level of agreement (1, strongly disagree, to 5, strongly agree) with items such as "My boss treats me fairly" (interac- tional justice, 2 items), "My employer makes deci- sions in an unbiased manner" (procedural justice, 5 items), and "I am fairly rewarded considering my responsibilities" (distributive justice, 5 items).
Time 1 Measures
Abusive supervision. I developed items to mea- sure abusive supervision by drawing on instru-
Time 2 Measures
Voluntary turnover. Information on voluntary turnover was collected when the 741 respondents
2000 Tepper 183
for the first wave of data collection were invited to participate in the second wave. Individuals who did not have the same supervisors at time 2 were asked to descrihe how the relationships had ended. Two independent coders classified the verbatim responses to this question into two categories: in- dividuals who had voluntarily quit their jobs be- tween time 1 and time 2 (n = 119) and individuals who did not have the same supervisors hecause they had been terminated or because their supervi- sors had quit or heen terminated [n = 55). The kappa was .89; classification disputes were re- solved by reaching consensus. The analysis of vol- untary turnover focused on the 119 individuals who had voluntarily quit and the 475 individuals who had the same supervisors at time 1 and time 2 (irrespective of whether or not they agreed to par- ticipate in the time 2 survey). Voluntary turnover was coded as follows: voluntarily quit, 0, and had same supervisor at times 1 and 2 , 1 .
Job satisfaction. I measured joh satisfaction us- ing the three-item overall satisfaction suhscale from the Michigan Organizational Assessment Question- naire (Gammann, Fichman, Jenkins, & Klesh, 1979). The three items were "All in all, I am satisfied with my joh," "In general, I don't like my joh," and "In general, I like working here" (1, strongly disagree, to 7, strongly agree).
Life satisfaction. I measured life satisfaction with Diener, Emmons, Larsen, and Griffin's (1985) five-item satisfaction with life scale. Example items are "I am satisfied with my life" and "If I could live my life over, I would change almost nothing" (1, strongly disagree, to 7, strongly agree).
Organizational commitment. I measured organi- zational commitment using Meyer, Allen, and Smith's (1993) 18-item instrument, which consists of three 6-item measures of continuance commit- ment, normative commitment, and affective com- mitment. Illustrative items are "It would he hard for me to leave my organization right now, even if I wanted to'' (continuance commitment), "I would not leave my organization right now hecause I have a sense of ohligation to the people in it" (normative commitment), and "I really feel as if this organiza- tion's problems are my own" (affective commit- ment). Respondents rated their level of agreement with the items using a scale ranging from 1, strongly disagree, to 7, strongly agree.
Conflict between work and family. I used Nete- meyer, Boles, and McMurrian's (1996) five-item scales to measure work-to-family confiict and fam- ily-to-work confiict. Example items were "The de- mands of my work interfere with my home and family life" (work-to-family confiict) and "Family- related strain interferes with my ahility to perform
job-related duties" (family-to-work confiict). Again, respondents rated their level of agreement with the items using a scale ranging from 1, strongly dis- agree, to 7, strongly agree.
Psychological distress. I measured psychologi- cal distress with six items from the Genter for Epi- demiologic Studies' Depression Scale (Radloff, 1977), six anxiety items from the Diagnostic Inter- view Schedule (Rohins, 1986), and six items from the emotional exhaustion scale of the Maslach Burnout Inventory (Maslach & Jackson, 1986). Va- lidity studies suggest that the depression and anx- iety scales are good predictors of a variety of vari- ahles including drug addiction, family income, and physical health (ohesity, hlood pressure, exercise; see Mirowsky and Ross [1989] for a thorough re- view), and Maslach and her colleagues have pro- duced a great deal of support for the validity of the hurnout scales (see Maslach and Leiter [1997] for a review). All the items were prefaced with the fol- lowing statement: "In the past few months, how often have you . . . " and ended with such state- ments as "wondered if anything is worthwhile" and "heen in low spirits" (depression), "felt irrita- hle, fidgety, or tense" and "felt afraid for no reason" (anxiety), and "felt emotionally drained from your work" and "felt burned out from your work" (emo- tional exhaustion). Respondents indicated the fre- quency with which they experienced the symp- toms captured hy each statement using the following scale: 0, never; 1, seldom; 2, sometimes; and 3, often.
RESULTS
A confirmatory factor analysis (GFA) revealed that the hest-fitting model was a 15-factor one in which the items designed to measure ahusive su- pervision, perceived mohility, interactional justice, procedural justice, distrihutive justice, job satisfac- tion, life satisfaction, continuance commitment, normative commitment, affective commitment, work-to-family confiict, family-to-work confiict, depression, anxiety, and emotional exhaustion loaded on 15 separate factors. All the factor load- ings were strong (greater than .50) and significant [p < .01), supporting the items as indicators of the underlying constructs they were designed to mea- sure.
Tahle 1 shows descriptive statistics, including correlations among the study variahles. When sub- ordinates perceived their supervisors as more ahu- sive, they reported significantly lower joh satisfac- tion, lower life satisfaction, greater continuance commitment, lower normative commitment, lower affective commitment, greater work-to-family con-
184 Academy of Management Journal April
TABLE 1 Descriptive Statistics and Correlations, Time 1 and Time 2"
Variable Mean s.d. 8 9 10 11 12 13 14 15
1.38 0.53 (.90)
3.22 1.16 .02 (.78)
3.69 1.06 - . 5 3 * * - . 0 2 * * (.90)
1. Abusive supervision
2. Perceived
mobility
3. Interactional
justice
4. Procedural justice 3.11 0.97 - . 4 8 * * .04 .48** (.89)
5. Distributive 3.13 1.08 - . 3 9 * * - . 0 6 .36** .62** (.95)
justice
6. Job satisfaction 5.54 1.31 - . 3 5 * * - . 0 2 .56** .43** .38** (.89)
7. Life satisfaction 4.66 1.30 - . 1 9 * * .13* .24** .28** .28** .43** (.89)
8. Continuance
commitment
9. Normative
commitment
10. Affective
commitment
11. Work-to-family 3.32 1.60 .22** - . 0 5 - . 2 7 * * - . 2 8 * * - . 2 3 * * - . 2 5 * * - . 2 2 * * .15** - . 0 9 * * - . 1 1 * (.93)
conflict
12. Family-to-virork
conflict
13. Depression
14. Anxiety
15. Emotional
exhaustion
" JV = 362. Internal consistency reliability coefficients (alphas) appear in parentheses along the main diagonal. * p < .05
** p < .01
4.09 1.09 .14** - . 3 9 * * - . 1 5 * * - . 2 5 * * - . 2 2 * * - . 1 5 * * - . 2 8 * * (.83)
3.70 1.29 - . 2 7 * * .03 .43** • .35** .24** .53** .27** .03 (.67)
4.29 1.29 - . 2 4 * * .01 ' .48** .36** .28** .62** .35** - . 1 5 * * .6?** (.84)
2.32 1.12 .12* - . 0 2 - . 1 3 * - . 1 8 * * - . 1 2 * - . 1 2 * - . 1 9 * * .15** .03 .02 .44** (.87)
1.28 0.59 .18** - . 1 2 * * - . 2 2 * * - . 2 2 * * - . 1 9 * * - . 2 8 * * - . 5 2 * * .26** - . 1 7 * * - . 2 7 * * .25** .22** (.81)
0.90 0.56 . 2 1 * * - . 0 9 - . 1 8 * * - . 1 8 * * - . 1 6 * * - . 2 7 * * - . 4 0 * * .12* - . 1 4 * * - . 2 2 * * .24** .23** .68** (.82)
1.30 0.71 .36** .02 - . 4 0 * * - . 3 8 * * - . 3 9 * * - . 4 6 * * - . 3 8 * * .19** - . 2 0 * * - . 3 4 * * .49** .26** .55** .53** (.86)
flict, greater family-to-work conflict, greater de- pression, greater anxiety, and greater emotional ex- haustion.
Logistic Regression Analysis
I tested Hypothesis 1 using logistic regression analysis, treating abusive supervision, perceived mobility, interactional justice, procedural justice, and distributive justice as predictors of member- ship in two groups: quit job (coded 0) and same supervisor at time 2 (coded 1). A test of the main effects for abusive supervision suggested abusive supervision and perceived mobility were related to voluntary turnover [Ax^2 = 11.71, p < .01). The signs on the standardized coefficients ()3 = -.37, abusive supervision, and j3 = —.22, perceived mo- bility) suggest subordinates were more likely to quit when their supervisors were more abusive and when they had more mobility. To assess whether organizational justice mediated the direct effect of abusive supervision, I evaluated its incremental contribution after exploring the effects of organiza- tional justice. The results of this analysis suggest the justice variables were related to voluntary turn-
over (A;ĉ 3 = 7.98, p < .05). Further, the incremen- tal contribution of abusive supervision, entered at the next step, was not significant, suggesting orga- nizational justice accounted for the effects of abu- sive supervision (Baron & Kenny, 1986). These findings support Hypothesis 1: organizational jus- tice mediated the effects of abusive supervision on voluntary turnover. However, entering the abusive supervision-perceived mobility cross-product at the final step did not improve model fit for either set of analyses, suggesting perceived mobility did not moderate the effects of abusive supervision on voluntary turnover.
Tests of Mediation and Moderation
I tested Hypotheses 2 through 6, which predict mediation effects, by (1) regressing abusive super- vision on each outcome variable, (2) regressing abu- sive supervision on each outcome variable after controlling for the effects of organizational justice, and (3) comparing the multiple squared correlation coefficients [R^s] attributable to abusive supervi- sion. Table 2 shows the hierarchical regression re- sults. For each outcome variable, the R^ attributable
2000 Tepper 185
TABLE 2 Results of Hierarchical Regression Analysis"
Outcome Variable
Job satisfaction Life satisfaction Continuance commitment Normative commitment Affective commitment Work to family conflict Family to work conflict Depression Anxiety Emotional exhaustion
Hypothesis
2 3 4a 4b 4c 5a 5b 6 6 6
Unmediated Effect of Abusive
Supervision
.13*
.04*
.02*
.07*
.06*
.05*
.01*
.03*
.04*
.13**
Effect of Abusive Supervision''
.00
.00
.00
.00
.00
.00
.00
.00
.01*
.01*
Main Effect of Perceived Mobility
.00
.03**
.15**
.00
.00
.00
.00
.01*
. 0 1 '
.00
Interactive Effect of Abuse and
Mobility
.01*
. 0 1 ' :00 .00 .00 .00 . 0 1 ' .01* .00 . 0 1 '
N = 362. Tabled values are H^s. The model includes a control for distributive, procedural, and interactional justice. ' p < . 1 0 * p < .05
**p < .01
to abusive supervision was significantly lower [p < .05) in the mediated model than in the unmediated model. Moreover, for all the outcome variables ex- cept anxiety and emotional exhaustion, the vari- ance accounted for by abusive supervision became nonsignificant after the effects of organizational justice were controlled for. Taken together, these results suggest organizational justice partially me- diated the effects of abusive supervision on anxiety and emotional exhaustion and fully mediated the effects of abusive supervision on the other outcome variables (Baron & Kenny, 1986). Hence, Hypothe- ses 2, 3, 4a, 4b, 4c, 5a, and 5b were supported, and Hypothesis 6 was partially supported,
I tested Hypothesis 7, which predicts a modera- tion effect, by regressing abusive supervision and perceived mobility on each outcome variable and examining the incremental contribution of the abu- sive supervision-perceived mobility cross-product. Table 2 shows that the moderated interaction term accounted for a significant proportion of the vari- ance in job satisfaction and depression [p < .01) and that the effects for life satisfaction, family-to- work conflict, and emotional exhaustion ap- proached significance (p < .10). To assess the shape of these interactions, I compared the sub- group correlations between abusive supervision and each variable for subordinates who fell below the median on perceived mobility (n = 181) and subordinates who were above the median on per- ceived mobility [n = 181). For subordinates with low mobility, correlations with abusive supervi- sion were as follows: —.38 (p < ,01) with job satis- faction, - . 2 7 (p < .01) with life satisfaction, .22 (p < .01) with family-to-work conflict, ,24 (p < .01)
with depression, and .41 (p < .01) with emotional exhaustion. For subordinates with high mobility, correlations with abusive supervision were - . 3 3 (p < .01) with job satisfaction, - . 1 1 (n.s.) with life satisfaction, .02 (n.s.) with family-to-work conflict, ,11 (n.s.) with depression, and .31 (p < .01) with emotional exhaustion. Consistent with Hypothesis 7, these correlations were stronger among subordi- nates with less job mobility. Further analyses sug- gested that the interaction between abusive super- vision and perceived mobility was unrelated to the justice variables and that perceived mobility played no moderating role in the relationships be- tween the justice variables and the outcomes vari- ables. Hence, there was no evidence of mediated moderation or moderated mediation (Baron & Kenny, 1986). That is, the moderating effects of perceived mobility operated independently of or- ganizational justice.
DISCUSSION
The results of this study suggest that (1) abusive supervision produces a number of dysfunctional consequences, (2) subordinates' experiences of in- justice explain many of their reactions to abusive supervision, and (3) some of the consequences of abusive supervision are more pronounced among subordinates who have less job mobility. My dis- cussion reviews the findings' implications for knowledge about abusive supervision in organiza- tions, highlights the implications for justice theory and research, and identifies the study's limitations.
186 Academy of Management Journal April
Abusive Supervision in Organizations
Subordinates whose supervisors were more abu- sive reported higher turnover, less favorable atti- tudes toward job, life, and organization, greater conflict between work and family life, and greater psychological distress. Moreover, the effects for job satisfaction, life satisfaction, family-to-work con- flict, depression, and emotional exhaustion were more pronounced for subordinates who had less job mobility. Hence, abusive supervision had broader effects on indexes of subordinates' atti- tudes and psychological distress when they had no viable means of escape.
The evidence suggesting the moderator operated independently of organizational justice begs con- sideration of the mechanism by which mobility moderates the effects of abusive supervision. One possible explanation stems from Bandura's (1986) contention that individuals' well-being is a func- tion of two components: aversive experiences and the thoughts of inefficacy brought about by inabil- ity to cope with those experiences. With respect to the research reported here, it can be speculated that organizational justice plays a role in explaining the effects of abusive supervision itself (the ayersive experience) and that mobility plays a role in indi- viduals' thought processes regarding their capacity to protect themselves from abusive supervision. In- dividuals with low mobility may dwell upon their coping deficiencies, judge themselves ineffica- cious, and magnify the severity of the distress more than those who can protect themselves from abu- sive supervision by quitting.
The potential harm associated with abusive su- pervision ought to serve as a warning to organiza- tions. Employees regard abusive supervision as a source of injustice that, in turn, has implications for their attitudes and well-being. The conse- quences of poor morale are well documented and include absenteeism, turnover, and reduced perfor- mance of extrarole behaviors. The links between abusive supervision and the various indexes of psy- chological distress are also troubling because even the milder manifestations may engender significant social and financial costs to organizations. For ex- ample, an individual whose distress prevents sleeping well at night may feel run down during the day or have difficulty getting along with coworkers, and distressed people may considerably add to an organization's health care costs (Mirowsky & Ross, 1989).
In future research, this work could be extended through exploration of possible links between abu- sive supervision and other variables that have been associated with the experience of organizational
injustice, such as organizational citizenship behav- ior (discretionary behaviors that in the aggregate promote effective organizational functioning) and dysfunctional organizational behaviors. This liter- ature would predict that subordinates of abusive supervisors repair the feeling that their organiza- tions have treated them unfairly by withholding citizenship behaviors and by performing anti- citizenship behaviors like revenge (Bies & Tripp, 1996) and retaliation (Skarlicki & Folger, 1997) with greater frequency. Moreover, to the extent abusive supervisors damage subordinates' self- esteem, use noncontingent punishment, and fail to provide helpful feedback, abusive supervision may interfere with subordinates' performance of ex- pected or inrole behaviors.
The evidence that abusive supervision has harm- ful effects in formal organizations, together with the possibility that abusive supervision has impli- cations for subordinates' performance of both ex- trarole and inrole behaviors, begs consideration of its prevalence and whether its occurrence is likely to increase or decrease in the future. The evidence regarding the prevalence of abusive supervision is limited to a small body of systematic work suggest- ing most people can recall at least one instance in which they were the targets of nonphysical abuse at work (Keashly et al., 1994). As for future trends, theorists have offered discrepant opinions. Neu- man and Baron (1997) argued that organizational responses to hostile environmental conditions, such as downsizing and increased pressure for pro- duction, cause stress and frustration and increase the frequency of workplace hostility, Ashforth (1997) offered a more optimistic picture, suggesting abusive supervision will become less prevalent as organizations embrace flatter structures, lateral as opposed to vertical decision flows, and knowledge- intensive, nonroutine technologies that require teamwork and self-management. Ashforth argued that "these and other trends will make the manage- rial failings of the petty tyrant more obvious and less tolerable" (1997: 139). In future research, this thesis as well as the competing view that, under some circumstances, employers may accept and even protect high-performing but abusive managers (Homstein, 1996) should be explored.
Implications for Justice Theory and Research
Organizational justice played a mediating role for all the criterion variables in this study. Further- more, although in this study I did not set out to explore the differential effects of the three justice dimensions, careful scrutiny of the regression re- sults suggests that interactional justice was related
2000 Tepper 187
to all the outcomes except continuance commit- ment and family-to-work conflict; procedural jus- tice was only related to the measures of job satis- faction, organizational commitment, and conflict between work and family; and distributive justice was only related to job satisfaction, life satisfaction, and emotional exhaustion. These results extend a growing body of empirical evidence that suggests the various justice dimensions play different pre- dictive roles in organizations. Several researchers have concluded that distributive justice tends to be a better predictor of employees' attitudes toward personal outcomes like pay satisfaction, w^hereas procedural justice tends to be a better predictor of employees' attitudes toward organizations and their representatives, such as organizational com- mitment (Greenberg, 1990; Lind & Tyler, 1988; Mc- Farlin & Sweeney, 1992). The results of the present study extend this line of thinking to suggest that interactional justice may be a better predictor of voluntary turnover than are procedural justice and distributive justice, a better predictor of organiza- tional commitment than is distributive justice, and a better predictor of psychological distress than is procedural justice. Hence, interactional justice's predictive role may be characterized as a hybrid of the roles distributive justice and procedural justice play, having effects on personal outcomes and at- titudes toward the system. In addition, interac- tional justice appears to play an important role in explaining how work-related experiences affect in- dividuals' lives away from work. At the very least, this study provides support for Bies's (2000) con- tention that individuals distinguish between the three kinds of justice and that interactional justice should be treated as a separate fairness dimension rather than as a component of procedural justice.
Limitations
The present study is not without limitations. One limitation is that all the data were collected from the same source. However, several features of the study minimize the likelihood that the relation- ships with abusive supervision are strictly artifac- tual. These features include the use of a research design in which abusive supervision and the crite- ria were measured at different points in time; the CFA results, which suggested a one-factor model did not fit the data well; and the evidence suggest- ing mobility moderated some of the effects of abu- sive supervision (Podsakoff & Organ, 1986). Never- theless, confidence in the instrument's construct validity would be bolstered by evidence suggesting that it correlates with peer ratings or observational
measures of abusive supervision as well as with objective criterion variables.
A second limitation has to do with the measure of voluntary turnover. Campion (1991) argued that tests of "motivational models" require that turn- over measures reflect individual choice. However, many instances of employee-initiated turnover cap- ture nonvoluntariness (for instance, an employee might quit for health reasons or because of knowing she or he will soon be fired). My turnover measure may have also been confounded in the sense that individuals who fully intended to quit but did not (for instance, because their supervisors quit or were fired) would have been left out of the analysis. To the extent my measure captured nonvoluntary turnover or my research design excluded those who would have voluntarily left given the chance, the power of my hypothesis tests is compromised.
A third limitation is that because I did not ad- minister surveys to those who did not have the same supervisors at time 2 as at time 1,1 was unable to determine whether escaping from abusive super- vision brought individuals' attitudes and psycho- logical distress in line with the feelings of those who have not had abusive supervisors. Further lon- gitudinal inquiry will be needed to evaluate how long it takes for an individual's well-being to nor- malize after working for an abusive supervisor and whether working for an abusive supervisor on one job affects a person's work-related behavior and attitudes on subsequent jobs.
A fourth limitation is that this study only ex- plored the possible consequences of abusive super- vision and sheds little light on its causes. One approach to investigating antecedents of abusive supervision would parallel research that has ex- plored abusive behaviors directed toward intimate partners. This work has linked abusive behavior with characteristics of abusers as well as character- istics and behaviors of abused persons (Walker, 1979). A complementary approach would be to ex- plore Neuman and Baron's (1997) proposed link between organizations' reactions to environmental conditions and abusive behavior. Either approach could facilitate the development of theory and could have considerable practical value to the ex- tent it helps organizations and individuals mini- mize and cope with a harmful source of injustice.
REFERENCES
Adams, J. S., & Freedman, S, 1976. Equity theory revis- ited: Comments and an annotated bibliography. In L. Berkowitz & E. Walster (Eds.), Equity theory: Toward a general theory of social interaction: 43-90, New York: Academic Press.
186 Academy of Management Journal April
Aquino, K,, Griffeth, R. W., Allen, D. G,, & Hom, P. W, 1997. Integrating justice constructs into the turnover process: A test of a referent cognitions model. Acad- emy of Management Journal, 40: 1208-1227.
Ashforth, B. 1994. Petty tyranny in organizations. Human Relations, 47: 755-778.
Ashforth, B. 1997. Petty tyranny in organizations: A pre- liminary examination of antecedents and conse- quences. Canadian Journal of Administrative Sci- ences, 14: 126-140.
Bandura, A, 1986. Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-HalL
Baron, R, A. 1993, Criticism (informal negative feedback) as a source of perceived unfairness in organizations: Effects, mechanisms, and countermeasures. In R. Gropanzano (Ed.), Justice in the workplace: Ap- proaching fairness in human resource manage- ment: 155-170. Hillsdale, NJ: Erlbaum.
Baron, R. M., & Kenny, D. A, 1986. The moderator-medi- ator variable distinction in social psychological re- search: Conceptual, strategic, and statistical consid- erations. Journal of Personality and Social Psychology, 51: 1173-1182.
Bies, R, J. 2000, Interactional (in)justice: The sacred and the profane. In J. Greenberg & R, Gropanzano (Eds,), Advances in organizational behavior: Forthcom- ing. Stanford, GA: Stanford University Press.
Bies, R. J., & Moag, J, S. 1986. Interactional justice: Gom- munication criteria of fairness. Research on Nego- tiation in Organizations, 1: 43-55,
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: Frontiers of theory and research: 246-260. Thousand Oaks, GA: Sage.
Bies, R, J., & Tripp, T, M, 1998. Two faces of the power- less: Coping with tyranny. In R, M. Kramer & M. A. Neale (Eds.), Power and influence in organizations: 203-219. Thousand Oaks, CA: Sage.
Bowers, K. S. 1968, Pain, anxiety, and perceived control. Journal of Consulting and Clinical Psychology, 32: 596-602.
Cammann, C , Fichman, M., Jenkins, D., & Klesh, J, 1979, The Michigan Organizational Assessment Ques- tionnaire. Unpublished manuscript. University of Michigan, Ann Arbor.
Campion, M. A, 1991. Meaning and measurement of turnover: Comparison of alternative measures and recommendations for research, journal of Applied Psychology, 76: 199-212.
Diener, E., Emmons, R. A., Larsen, R, J., & Griffin, S, 1985. The Satisfaction with Life Scale, Journal of Personality Assessment, 49: 71-75.
Duffy, M. K,, Ganster, D. G., & Pagon, M, 1998, Social undermining and social support in the workplace.
Paper presented at the annual meeting of the Acad- emy of Management, San Diego,
Folger, R, 1986, Rethinking equity theory: A referent cognitions model. In H. W. Bierhoff, R. L. Gohen, & J, Greenberg (Eds.), Justice in social relations: 145- 162. New York: Plenum,
Folger, R., & Konovsky, M. A. 1989. Effects of procedural and distributive justice on reactions to pay raise decisions. Academy of Management Journal, 32: 115-130.
Gilliland, S. W. 1994, Effects of procedural and distrib- utive justice on reactions to a selection system. Jour- nal of Applied Psychology, 79: 691-701.
Goffman, E. 1967. Interaction ritual. New York: Pan- theon.
Greenberg, J, 1990. Organizational justice: Yesterday, to- day, and tomorrow. Journal of Management, 16: 399-432,
Greenhaus, J. J., & Beutell, N. J. 1985. Sources of conflict between work and family roles. Academy of Man- agement Review, 10: 76-88.
Homstein, H. A. 1996. Brutal bosses and their prey. New York: Riverhead Books.
Jezl, D. R,, Molidor, C. E., & Wright, T, L. 1996, Physical, sexual, and psychological abuse in high school dat- ing relationships: Prevalence rates and self-esteem issues. Child and Adolescent Social Work Journal, 13: 69-87.
Judge, T. A., Boudreau, J, W., & Bretz, R. D. 1994. Job and life attitudes of male executives. Journal of Applied Psychology, 79: 767-782.
Keashly, L., Trott, V., & MacLean, L. M, 1994. Abusive behavior in the workplace: A preliminary investiga- tion. Violence and Victims, 9: 341-357.
Konovsky, M, A., & Cropanzano, R, 1991. Perceived fair- ness of employee drug testing as a predictor of em- ployee attitudes and joh performcince. Journal of Applied Psychology, 76. 698-707.
Leventhal, G. S. 1980. What should be done with equity theory? New approaches to the study of fairness in social relationships. In K. Gergen, M, Greenberg, & R, Willis (Eds.), Social exchange: Advances in theory and research: 27-55, New York: Plenum,
Lind, E. A., & Tyler, T. T. 1988. The social psychology of procedural justice. New York: Plenum.
Martin, J. 1981, Relative deprivation: A theory of distrib- utive justice for an era of shrinking resources. In L. L. Gummings & B. M. Staw (Eds.), Research in orga- nizational behavior, vol. 3: 53-108, Greenwich, GT: JAI Press.
Maslach, C , & Jackson, S. E, 1986. Maslach Burnout Inventory (2nd ed.). Palo Alto, CA: Consulting Psy- chologists Press,
Maslach, C , & Leiter, M. P. 1997. The truth about bum-
2000 Tepper 189
out: How organizations cause personal stress and what to do ahout it. San Francisco: Jossey-Bass.
McFarlin, D. B., & Sweeney, P. D. 1992. Distributive and procedural justice as predictors of satisfaction with personal and organizational outcomes. Academy of Management Journal, 35: 626-637.
Meyer, J. P., & Allen, N. J. 1991. Commitment in the workplace: Theory, research, and application. Thousand Oaks, CA: Sage.
Meyer, J. P., Allen, N. J., & Smith, C. 1993. Commitment to organizations and occupations: Extension and test of a three-component conceptualization. Journal of Applied Psychology, 78: 538-551.
Mikula, G., Petri, B., & Tanzer, N. 1990. What people regard as unjust: Types and structures of everyday experiences of injustice. European Journal of So- cial Psychology, 20: 133-149.
Miller, S. M. 1980. Why having control reduces stress: If I can stop the roller coaster, I don't want to get off. In J. Garber & M. E. P. Seligman (Eds.), Human help- lessness: Theory and applications: 71-95. New York: Academic Press.
Mirowsky, J., & Ross, C. 1989. The social causes of psychological distress. Hawthorne, NY: Aldine de Gruyter.
Netemeyer, R. G., Boles, J. S., & McMurrian, R. 1996. Development and validation of work-family conflict and family-work conflict scales. Journal of Applied Psychology, 81: 400-410.
Neuman, J. H., & Baron, R. A. 1997. Aggression in the workplace. In R. A. Giacalone & J. Greenberg (Eds.), Antisocial behavior in organizations: 37-67. Thou- sand Oaks, CA: Sage.
Niehoff, B. P., & Moorman, R. H. 1993. Justice as a mediator of the relationship between methods of monitoring and organizational citizenship behav- ior. Academy of Management Journal, 36: 527- 556.
Podsakoff, P. M., & Organ, D. 1986. Self-reports in orga- nizational research: Problems and prospects. Jour- nal of Management, 12: 531-544.
Radloff, L. 1979. The CES-D scale: A self-report depres- sion scale for research in the general population. Applied Psychological Measurement, 1: 385-401.
Raymond, R., & Bruschi, I. G. 1989. Psychological abuse among college women in dating relationships. Per- ceptual and Motor Skills, 69: 1283-1297.
Richman, J. A., Flaherty, J. A., Rospenda, K. M., & Chris- tensen, M. L. 1992. Mental health consequences and correlates of reported medical student abuse. Jour- nal of the American Medical Association, 267: 692-694.
Robins, L. 1986. The development and characteristics of the NIMH Diagnostic Interview Schedule. In M. Weissman, J. Myers, & C. Ross (Eds.), Community
surveys of psychiatric disorders: 403-427. New Brunswick, NJ: Rutgers University Press.
Robinson, S. L., & Bennett, R. J. 1995. A typology of deviant workplace behaviors: A multidimensional scaling study. Academy of Management Journal, 38: 555-572.
Schriesheim, C. A., & Hinkin, T. R. 1990. Influence tac- tics used by subordinates: A theoretical and empiri- cal analysis and refinement of the Kipnis, Schmidt, and Wilkinson subscales. Journal of Applied Psy- chology, 75: 246-257.
Sheehan, K. H., Sheehan, D. V., White, K., Leibowitz, A., & Baldwin, D. C. 1990. A pilot study of medical student abuse. Journal of the American Medical Association, 263: 533-537.
Shepard, M. F., & Campbell, J. A. 1992. The abusive behavior inventory: A measure of psychological and physical abuse. Journal of Interpersonal Violence, 7: 291-305.
Skarlicki, D., & Folger, R. 1997. Retaliation in the work- place: The roles of distributive, procedural, and in- teractional justice. Journal of Applied Psychology, 82: 434-443.
Tepper, B. J. 1995. Upward maintenance tactics in supervisory mentoring and nonmentoring relation- ships. Academy of Management Journal, 38: 1191-1205.
Tolman, R. M. 1989. The development of a measure of psychological maltreatment of women by their male partners. Violence and Victims, 4: 159-177.
Van Maanen, J., & Schein, E. H. 1979. Toward a theory of organizational socialization. In B. Staw (Ed.), Research in organizational behavior, vol. 1: 209-264. Greenwich, CT: JAI Press.
Walker, L. 1979. The battered woman. New York: Harper & Row.
Wolfe, D. A. 1987. Child abuse: Implications for child development and psychopathology. Beverly Hills, CA: Sage.
APPENDIX
Abusive Supervision Items
The items were prefaced with tbe statement, "My boss . . . " Respondents used a five-point response scale where 1 was "I cannot remember him/her ever using this behavior with me," 2 was "He/she very seldom uses this behavior with me," 3 was "He/she occasion- ally uses this behavior witb me," 4 was "He/she uses this behavior moderately often with me," and 5 was "He/she uses this behavior very often with me." The items were:
1. Ridicules me 2. Tells me my thoughts or feelings are stupid 3. Gives me the silent treatment 4. Puts me down in front of others
190 Academy of Management Journal April
5. Invades my privacy 6. Reminds me of my past mistakes and failures
• 7. Doesn't give me credit for jobs requiring a lot of effort 8. Blames me to save himself/herself embarrassment 9. Breaks promises he/she makes
10. Expresses anger at me when he/she is mad for an- other reason
11. Makes negative comments about me to others 12. Is rude to me 13. Does not allow me to interact with my coworkers
14. Tells me I'm incompetent 15. Lies to me
Bennett J. Tepper is an associate professor of manage- ment and a research fellow in the Gatton College of Business and Economics at the University of Kentucky. He received his Ph.D. in organizational psychology from the University of Miami. His current research interests focus on organizational justice and prosocial and antiso- cial organizational behavior.