Identity threat and antisocial behavior in organizations: The moderating effects of individual differences, aggressive modeling, and hierarchical status

Abstract

This study examines whether the experience of identity threat predicts antisocial behavior directed towards other employees. A social interactionist model is used as a theoretical framework for predicting that employees who are frequent recipients of actions that challenge or diminish their sense of competence, dignity, or self-worth will engage in higher levels of antisocial behavior. However, it is predicted that the strength of this relation will be moderated by individual (gender, age, and attitudes toward revenge) and situational (aggressive modeling, hierarchical status) factors. Data from 308 employees from three organizations supported moderating effects of age, revenge attitudes, and hierarchical status. A three-way interaction was also found showing that identity threat was more strongly related to antisocial behavior for low as compared to high status employees, but only when they were exposed to low levels of aggressive modeling.

Full text

Identity threat and antisocial behavior in organizations:
The moderating effects of individual differences, aggressive
modeling, and hierarchical status
Karl Aquino
a,*
and Scott Douglas
b
a
Department of Business Administration, College of Business and Economics, University of Delaware, Newark, DE 19716, USA
b
Binghamton University, PO Box 6000, Binghamton, NY 13902, USA
Abstract
This study examines whether the experience of identity threat predicts antisocial behavior directed towards other employees. A
social interactionist model is used as a theoretical framework for predicting that employees who are frequent recipients of actions
that challenge or diminish their sense of competence, dignity, or self-worth will engage in higher levels of antisocial behavior.
However, it is predicted that the strength of this relation will be moderated by individual (gender, age, and attitudes toward revenge)
and situational (aggressive modeling, hierarchical status) factors. Data from 308 employees from three organizations supported
moderating effects of age, revenge attitudes, and hierarchical status. A three-way interaction was also found showing that identity
threat was more strongly related to antisocial behavior for low as compared to high status employees, but only when they were
exposed to low levels of aggressive modeling.
Ó2003 Elsevier Science (USA). All rights reserved.
1. Introduction
Harmful, injurious, and destructive behaviors di-
rected by one employee against another are common
occurrences in todayÕs workplace. One study of Finnish
workers found that 32% had observed one or more co-
workers being exposed to verbally harassing behavior at
work (Bj€
oorkqvist,
€
OOsterman, & Hjelt-B€
aack, 1994). In
another study, a survey of American human resource
managers found that 20% reported that their organiza-
tions had experienced workplace violence since 1990
(Romano, 1994). As a final example, a survey of 327
first-line American workers showed that half reported
acts of mistreatment at work within a three-year time
frame (Ehrlich & Larcom, 1994). Some writers refer to
these acts, and others like them, as antisocial workplace
behavior (e.g., Andersson & Pearson, 1999; Robinson &
OÕLeary-Kelly, 1998).
Antisocial workplace behavior has been defined as
actions directed towards other employees or the orga-
nization that have the potential for producing physical,
economic, psychological, or emotional harm (Robinson
&OÕLeary-Kelly, 1998). The study of these behaviors
has attracted considerable interest in recent years and
several theoretical models have been advanced to ex-
plain their occurrence. One model proposed by OÕLeary-
Kelly and her colleagues (OÕLeary-Kelly, Griffin, &
Glew, 1996) emphasizes situational determinants like
decisions that affect valued outcomes, incentive induce-
ments that reward aggressive behavior, and aversive
physical environments. Other models focus on individ-
ual-level variables like perceptions of injustice (Aquino,
Lewis, & Bradfield, 1999; Greenberg & Alge, 1998;
Skarlicki & Folger, 1997), cognitive appraisal processes
(Martinko & Zellars, 1998), emotional reactivity
(Berkowitz, 1993), and negative affectivity (Skarlicki,
Folger, & Tesluk, 1999). Finally, more dynamic models
describe how antisocial behavior results from an ongo-
ing process of repeated interpersonal exchanges where
one party perceives a threat to his or her self-identity
and retaliates against the perceived source of threat.
This explanation underlies social interactionist theories
of aggression (e.g., Felson, 1992; Felson & Steadman,
1983; Tedeschi & Felson, 1994) and models of conflict
escalation (e.g., Andersson & Pearson, 1999; Folger &
Organizational Behavior and Human Decision Processes 90 (2003) 195–208
www.elsevier.com/locate/obhdp
ORGANIZATIONAL
BEHAVIOR
AND HUMAN
DECISION PROCESSES
*
Corresponding author. Fax: 1-302-831-4196.
E-mail address: aquinok@be.udel.edu (K. Aquino).
0749-5978/03/$ - see front matter Ó2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0749-5978(02)00517-4

Skarlicki, 1998) and revenge in organizations (e.g., Bies,
Tripp, & Kramer, 1997).
This study draws from each of the theoretical per-
spectives cited above to test a model in which identity-
threatening events experienced by an employee are
hypothesized to predict antisocial behaviors performed
by that employee. We view identity-threatening events
as a subclass of the broader category of antisocial be-
havior. However, in contrast to most studies of antiso-
cial behavior, we conceptualize and measure identity
threats as actions directed against an employee by one
or more co-workers. We then assume that these expe-
riences can provoke the threat-recipient to respond by
engaging in other forms of antisocial behavior. The
proposed relationship between identity threat and an
aggressive counter-response follows directly from dy-
namic models of conflict escalation and revenge (e,g.,
Andersson & Pearson, 1999; Bies et al., 1997; Felson &
Steadman, 1983; Folger & Skarlicki, 1998) and is em-
pirically well established (e.g., Felson, 1992; Felson &
Steadman, 1983; Tedeschi & Felson, 1994). Hence, we
treat it as a building block for more complex hypothe-
ses. Rather than simply replicating previous work, our
study makes several unique contributions to the litera-
ture. First, this is the only study to our knowledge that
directly tests the relationship between identity threat-
ening events and antisocial behavior. Second, we go
beyond looking at the main effect of identity threat by
examining individual-level and contextual variables as
moderators of this relationship. Finally, we combine the
individual and contextual variables into a higher-order
interaction that adds new theoretical complexity to ex-
isting models of antisocial behavior.
We investigate two categories of individual-level
variables as possible moderators: demographic charac-
teristics and attitudes toward revenge (Stuckless &
Goranson, 1992). Guided by OÕLeary-Kelly et al.Õs
(1996) model of organization-motivated aggression, we
examine whether the presence of aggressive role models
in the workplace strengthens the relation between
identity threat and antisocial behavior. Lastly, we in-
vestigate the relatively understudied question of whether
this relation is moderated by the threat-recipientÕs hi-
erarchical status.
2. Theoretical background and hypotheses
A personÕs self-identity consists of a combination of
personal attributes (e.g., capable, competent) and the
social identities (e.g., gender, ethnicity) that he or she
seeks to present in a given situation (Erez & Early,
1993). Most people strive to maintain positive self
(Bies, 1999; Brockner, 1988; Steele, 1988) and social
identities (Tajfel & Turner, 1985) and so they are highly
motivated to defend themselves against acts that
threaten these identities (Baumeister, Smart, & Boden,
1996; Felson & Steadman, 1983). The treatment a
person receives from others is an important source of
identity validation (e.g., Darwin, 1872; Goffman, 1967;
Lind & Tyler, 1988; Sennett & Cobb, 1973; Steele,
1988). Consequently, being treated poorly by oneÕs co-
workers can threaten oneÕs personal identity as a moral
being deserving of fairness, consideration, and respect
(Bies, 1999; Lind & Tyler, 1988) as well as his or her
social identity as a valued organizational member
(Lind, 1997).
We define an identity threat as any overt action by
another party that challenges, calls into question, or
diminishes a personÕs sense of competence, dignity, or
self-worth (Bies, 1999; Steele, 1988). Examples include
ethnic or religious slurs, harsh criticism of oneÕs abilities,
deceit, or public humiliation (Bies, 1999). We focus on
these behaviors, because it is the less intense forms of
interpersonal mistreatment, rather than the more highly
publicized forms like physical assault or homicide, that
occur most frequently in organizations (Andersson &
Pearson, 1999; Baron & Neuman, 1996). As mentioned
earlier, identity threats can provoke antisocial behavior.
Oftentimes these behaviors will be retaliatory responses
directed by the threat-recipient (the target) against the
perceived source of threat (the perpetrator). Some of
these responses are likely to qualify as antisocial because
they have the potential to harm the perpetrator.
The motives for retaliation are numerous. They can
include the targetsÕneed to reaffirm a damaged identity,
to restore justice, or to deter future identity threats
(Baumeister et al., 1996; Felson, 1992; Felson & Stead-
man, 1983; Gilligan, 1996; Tedeschi & Felson, 1994).
But the threat-recipient may also direct antisocial be-
havior against non-perpetrators. A recent meta-analysis
showed that displaced aggression in response to various
triggering provocations is a robust and valid phenome-
non (Marcus-Newhall, Pederson, Carlson, & Miller,
2000). In such cases, antisocial behavior in the form of
displaced aggression may serve a value-expressive
function by allowing the threat-recipient to vent nega-
tive emotions (e.g., anger, shame, or resentment) against
a convenient, available, or powerless target (Bies et al.,
1997; Lind & Tyler, 1988; Thomas, 1992). Whether di-
rected against the perpetrator or someone else, there are
theoretical grounds for predicting a relationship be-
tween identity threat and antisocial behavior; hence, we
test the following hypothesis:
Hypothesis 1. Identity threats experienced by an em-
ployee are positively related to antisocial behaviors
performed by that employee.
As noted earlier, Hypothesis 1 is a building block to
our subsequent predictions. Our primary aim in this
study is to test whether individual level variables, social
196 K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208

modeling, and hierarchical status moderate this relation.
We first consider the effect of three individual-level
variables—gender, age, and attitudes toward revenge—
that prior theory and research has identified as likely
predictors of antisocial behavior.
2.1. Demographic differences
Our rationale for hypothesizing that gender and age
moderate the relation between identity threat and anti-
social behavior is based on three well-documented em-
pirical findings. First, anger is one of the primary
emotions generated by an identity threat (Bies, 1999;
Felson & Steadman, 1983; Gilligan, 1996; Steele, 1988).
Second, anger is a known predictor of antisocial be-
havior (Averill, 1982; Felson, 1992; Skarlicki et al.,
1999). And third, peoplesÕobservable reactions to anger-
producing events have been found to differ as a function
of gender (e.g., Kogut, Langely, & OÕNeal, 1992; Tavris,
1982) and age (e.g., Scheiman, 1999). Taken together,
these findings suggest that demographic characteristics
may moderate the relationship between identity threat
and antisocial behavior.
Gender. Several studies have shown that men display
overt anger more frequently than women (Kogut et al.,
1992; Tavris, 1984). Overt expressions of anger often
fall within the same conceptual domain as antisocial
behavior as defined in this study. For example, Speil-
bergerÕs (1996) State-Trait Anger Expression Inventory
(STAXI) identifies behaviors like making sarcastic re-
marks and saying nasty things to others, arguing with
others, and striking out at others to measure the out-
ward expression of anger. Since all of these behaviors
have the potential to harm others, there is reason
to believe that variables predicting overt expressions of
anger, such as gender, may also predict antisocial
behavior.
One explanation for the observed gender differences
in overt expressions of anger is that cultural norms
support a traditional belief that women should suppress
their anger (Tavris, 1984). In an organizational context,
some writers (e.g., Black, 1990) have argued that this
belief leads to a double-standard that gives men greater
latitude in expressing anger and aggression than wo-
men. From this perspective, women often inhibit the
expression of anger for fear of being regarded as in-
appropriately aggressive or domineering (Glomb &
Hulin, 1997). Supporting this view, studies show that
women regard displays of anger as being less socially
appropriate than men (Smith et al., 1989) and as having
greater personal and relationship costs (Davis, LaRosa,
& Foshee, 1992). These differences in cultural expecta-
tions regarding the appropriateness of expressing anger
in response to provocative events provides a theoretical
basis for predicting that women will also exhibit lower
levels of antisocial behavior in response to identity
threats than men. The following hypothesis tests this
prediction:
Hypothesis 2. The relationship between identity threats
experienced by an employee and antisocial behaviors
performed by that employee is stronger for men than
women.
Age. Age is another demographic characteristic that
might moderate the link between identity threat and
antisocial behavior. Studies in non-organizational set-
tings show that older persons experience lower levels of
anger in response to situational stressors than younger
ones (Ross & Van Willigen, 1996; Scheiman, 1999). If
this effect generalizes to the workplace, then we would
expect the strength of the relationship between identity
threat and antisocial behavior to decline with age.
There are several theories supporting this prediction.
Disengagement theory (Cumming & Henry, 1961)
posits that as people become older they withdraw from
society and have fewer social interactions. As a result,
they become less emotionally invested in other people
and objects, less concerned with social expectations,
and more inner focused (Neugarten, 1996). If a disen-
gagement process occurs in the workplace, then older
employees may place less importance on the treatment
they receive from co-workers as a source of self- and
social-validation. As a result, they are less likely to
respond aggressively to identity-threatening events.
Social-emotional selectivity theory (Cartensen, 1992)
makes a similar prediction.
According to social-emotional selectivity theory
(Cartensen, 1992), people become more effective at
conserving their emotions and are less profoundly af-
fected by others as they age. They also grow more tol-
erant of others and show a greater capacity to express
affection even in conflict situations. In sum, this theory
implies that age is accompanied by a state of psycho-
logical and emotional maturity that may reduce the in-
tensity of peoplesÕreactions to identity-threatening
events (Scheiman, 1999). In turn, this should decrease
their motivation to engage in antisocial behavior. Both
disengagement and social-emotional theory suggest the
following hypothesis concerning the moderating effects
of age:
Hypothesis 3. The relationship between identity threats
experienced by an employee and antisocial behavior
performed by that employee is stronger for younger
than older employees.
In addition to demographics, we examine the possible
moderating effect of an attitudinal construct that has
clear conceptual relevance for explaining antisocial be-
havior. This construct is a personÕs attitude towards
revenge (Stuckless & Goranson, 1992).
K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208 197

2.2. Revenge attitudes
Attitudes formed through learning and acculturation
are most likely to influence behavior when the attitudes
are strong, specific, and relevant (Fishbein & Ajzen,
1975). Revenge attitudes fit this description because
most people have had repeated opportunities to con-
template acts of vengeance. Furthermore, the content of
these attitudes is directly associated with the likelihood
of retaliation against perceived mistreatment. We pre-
viously argued that antisocial behavior includes acts
directed against the perceived source of identity threat
as well as ‘‘value-expressive’’ acts that are displaced
against convenient or vulnerable targets. In either case,
there are theoretical reasons to expect antisocial be-
havior to occur more frequently in response to an
identity threat when a person holds more rather than
less favorable attitudes toward revenge.
First, a favorable attitude toward revenge indicates
high levels of a value-expressive orientation regarding
revenge (Stuckless & Goranson, 1992), making the ex-
pression of that value strong enough to encourage dis-
placed aggression (as antisocial behavior) when the
instrumental uses of aggression (acts targeted directly
against the source of identity threat) are unavailable.
Second, antisocial behavior should occur less frequently
among people who disapprove of revenge because (a)
their need to express this value is less strong and (b) the
tendency to aggress against sources of identity threat
would be limited to those cases where it could success-
fully serve an instrumental purpose. Taken together,
these two conditions suggest that the relationship be-
tween identity threat and antisocial behavior strengthens
as peoplesÕattitudes toward revenge become more fa-
vorable. The following hypothesis tests this prediction:
Hypothesis 4. The relationship between identity threats
experienced by an employee and antisocial behavior
performed by that employee is stronger for employees
who hold favorable rather than unfavorable attitudes
toward revenge.
Thus far we have presented arguments supporting the
moderating effects of individual-level variables on the
relation between an identity threat and antisocial be-
havior. In the following section, we introduce a con-
textual variable—aggressive modeling—as another
possible moderator of this relation.
2.3. Aggressive modeling
Social learning theory (Bandura, 1973) treats antiso-
cial behavior as being learned and maintained in much
the same way as any other behavior (Geen, 1990;
Martinko & Zellars, 1998). A basic tenet of the social
learning perspective is that environmental factors can
encourage antisocial behavior. One of these factors is
the presence of aggressive models. As Bandura and his
colleagues (Bandura, Ross, & Ross, 1963) showed in
their famous clown studies, even strangers are influential
role models for antisocial behavior. This observation is
consistent with the predictions of social information
processing theory (Salancik & Pfeffer, 1978). According
to this theory, people use information in their immediate
environment to develop expectations concerning ap-
propriate behavior. Since employees are likely to look to
their co-workers for signals that indicate how they are
expected to behave in the workplace, the extent to which
they choose to engage in antisocial behavior in response
to identity threats should be influenced by their expo-
sure to aggressive role models. Robinson and OÕLeary-
KellyÕs (1998) study of workgroups supports this con-
clusion by showing that the behavior of oneÕs group is a
powerful determinant of an individualÕs antisocial be-
havior. Based on theories of social learning (Bandura,
1973) and social information processing (Salancik &
Pfeffer, 1978), we expect the relation between identity
threat and antisocial behavior to be moderated by the
employeeÕs perception of aggressive models in their im-
mediate work environment. The following hypothesis
tests this prediction:
Hypothesis 5. The relationship between identity threats
experienced by an employee and antisocial behavior
performed by the employee is stronger for employees
exposed to more frequent models of aggressive behavior.
We now turn to social status, another possible mod-
erator of the relation between identity threat and anti-
social behavior that has received scant attention in the
management literature. Although several writers have
argued that status differences have implications for how
people respond to identity threats (e.g., Baumeister
et al., 1996; Gilligan, 1996; Wicklund & Gollwitzer,
1982), this hypothesis has not been empirically tested in
organizations. The clearest indicator of social status in
organizations is hierarchical position; hence, we used
this variable as a basis for testing our predictions re-
garding the moderating effects of status.
2.4. Hierarchical status
The desire to achieve high status is among the most
ubiquitous and powerful of all human motives (Daly &
Wilson, 1988). One reason for this is that in almost all
social groups, those who reach the top of status hier-
archy generally receive a disproportionately large share
of the symbolic and material things toward which peo-
ple strive (e.g., pay, authority, recognition, prestige,
autonomy) whereas those at the lowest levels receive a
disproportionately large share of negative liabilities
(e.g., dangerous working conditions, low pay, harsh
198 K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208

treatment from superiors). Several writers (e.g., Bau-
meister et al., 1996; Gilligan, 1996; Wicklund & Gol-
lwitzer, 1982; Daly & Wilson, 1988) have argued that a
social psychological consequence of status differences in
social hierarchies is that the ‘‘losers’’ become highly
aggressive in defending themselves against identity
threats. Daly and Wilson (1988) argued that this occurs
because people compete not merely to attain high status,
but to also avoid low status. Consequently, people at the
bottom of the hierarchy, particularly if they are men, are
often willing to do whatever it takes to respond com-
petitively to acts that threaten to undermine their al-
ready precarious social positions.
GilliganÕs (1996) interviews with men imprisoned for
violent offenses support Daly and WilsonÕs (1988) claim.
In his interviews, Gilligan found that many of the vio-
lent acts that led to the offendersÕimprisonment were
triggered by what, for most people, would appear to be
rather ‘‘trivial’’ insults (e.g., being given a dirty look,
having oneÕs new shoes stepped on, being shoved by
someone at a bar). However, from the perspective of
men whose lives offered few opportunities for acquiring
status-enhancing resources, aggressive responses to in-
sults became a way of preserving what little status they
had. These findings led Gilligan (1996) to conclude that
the experience of being shamed or humiliated, and the
consequent need to defend oneÕs self-respect, provides a
plausible explanation for societal violence.
Applying the theoretical arguments and empirical
findings cited above to organizations, we propose that a
high status position and the symbolic and material af-
firmations that accompany it provide the role occupant
with a psychological ‘‘buffer’’ against self-invalidating
events. If so, then high status employees should be less
motivated to exhibit antisocial behavior in response to
identity threats than low status employees. This rea-
soning suggests that the relation between identity threat
and antisocial behavior is weaker for high as compared
to low status employees. However, there is a second
theoretical argument supporting this prediction that is
not based on the notion a surfeit of symbolic and ma-
terial affirmations provides a psychological buffer
against identity threat. This explanation posits that high
status employees may perceive greater normative pres-
sures to refrain from exhibiting antisocial behavior be-
cause they occupy highly visible leadership roles.
Leaders who engage in petty revenge, or who punish
others excessively or arbitrarily, can invite resentment
and hostility among subordinates and peers because
such actions may be viewed as an abuse of power
(Freud, 1950; Hogan & Emler, 1981). A study by
Aquino, Tripp, and Bies (2001) supported this argument
by showing that the relationship between the attribution
of blame for an interpersonal offense and revenge
against the offender was weaker when the offender had
high as compared to low hierarchical status. Similarly,
anthropological studies have found that when group
leaders abuse their power, they can provoke reprisals
from their subordinates in what has been described as a
‘‘reverse dominance’’ hierarchy (Boehm, 1993). What
these findings suggest is that it may be to the political
advantage of high status persons to refrain from exces-
sive displays of antisocial behavior. If a norm of social
restraint operates in organizations, then the relationship
between identity threat and antisocial behavior should
be weaker for high as compared to low status employ-
ees. The two theoretical mechanisms described above,
which we will refer to as the ‘‘symbolic affirmations’’
and ‘‘social restraint’’ explanations, lead to the follow-
ing prediction:
Hypothesis 6a. The relationship between identity threats
experienced by an employee and antisocial behavior
performed by the employee is stronger for employees
with low rather than high hierarchical status.
Assuming Hypothesis 6a is supported by data, a
logical follow-up question is which of the two mecha-
nisms best explains the results. We believe it is possible
to provide a tentative answer to this question by testing
a three-way interaction involving identity threat, hier-
archical status, and aggressive modeling. If high status
employees are influenced primarily by a norm of social
restraint that discourages them from engaging in anti-
social behavior in response to identity threat, then the
moderating effect of hierarchical status should be af-
fected by environmental differences in aggressive mod-
eling. Specifically, when the organizational environment
presents few models of aggressive behavior, then high
status persons should feel the strongest normative
pressure to refrain from exhibiting antisocial behavior in
response to identity threats. The reason being that such
actions can generate resentment and hostility among
subordinates and perhaps even peers. As a result, the
high status employeeÕs authority and position of domi-
nance may be undermined (Boehm, 1993). Empirically,
this argument would be supported if we found a stron-
ger relationship between identity threat and antisocial
behavior for low as compared to high status employees
in the absence of aggressive models. However, when the
environment has many aggressive role models, then high
status employee may perceive little or no obligation to
exercise social restraint, since nobody else in the social
environment does so. As a result, in the presence of
aggressive models, the relationship between identity
threat and antisocial behavior should be less influenced
by the targetÕs hierarchical status. The pattern of results
described above predicts a three-way interaction be-
tween identity threat, hierarchical status, and aggressive
modeling consistent with the social restraint argument.
On the other hand, if the symbolic affirmations ar-
gument better accounts for the data we should see a
K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208 199

different pattern of relationships among identity threat,
hierarchical status, and aggressive modeling. Specifi-
cally, if receiving a surplus of symbolic affirmations
explains why high status persons are less likely than low
status persons to respond to identity threats by engaging
in antisocial behavior, then the presence (or absence) of
aggressive role models should not influence the moder-
ating effect of status. This means that we should find
empirical support for a two-way interaction involving
identity threat and hierarchical status (Hypothesis 6a),
but not for a three-way interaction that includes ag-
gressive modeling.
Based on the preceding arguments, we test the fol-
lowing three-way interaction to see which of the theo-
retical mechanisms described above best accounts for
the data, keeping in mind that a lack of evidence for one
mechanism does not necessarily provide conclusive
support for the other:
Hypothesis 6b. The relation between identity threats
experienced by an employee and antisocial behavior
performed by that employee is stronger for employees
who have low rather than high hierarchical status, but
only when the employee is exposed to few rather than
many aggressive role models.
3. Method
3.1. Procedures
Two surveys were distributed to 367 employees from
three organizations located in the northeast United
States. The organizations included a transportation
company, a public school system, and a municipality.
The first survey included the measures of revenge atti-
tudes, aggressive modeling, status, gender, and age.
Prior to completing this survey, employees attended a
meeting with one of the investigators who briefly de-
scribed the studyÕs purpose. Surveys were then admin-
istered on-site and no members of upper management
were present during the administration to lower level
personnel. Completed surveys were placed in a box and
the investigator, who remained in the room during the
administration, removed the contents after all partici-
pants had completed the surveys. The second survey was
completed 5–8 days later and included the measures of
identity threat and antisocial behavior. For this ad-
ministration, employees were again asked to attend a
meeting in which an investigator briefly described the
study. However, rather than describe the second survey
as a continuation of the first, it was presented as an
assessment of workplace behaviors. During both ad-
ministrations the voluntary and confidential nature of
the study was emphasized. Once again, the investigator
remained in the room and removed the contents of the
box after all the participants had completed the second
survey. Respondents provided their motherÕs maiden
name and their date of birth on both surveys so that
their responses could be matched.
3.2. Sample
Three-hundred and eight employees provided usable
data on all study variables (84% response rate). Ninety-
seven respondents were employed by the transportation
company, 95 by the public school system, and 116 by the
municipality. We excluded data from 59 employees who
initially agreed to participate because they either: (a)
failed to complete both surveys; (b) failed to provide the
same identifying information during both administra-
tions; or (c) were not employed by their current em-
ployer for at least six months prior to the first
administration. With the exception of two males who
left during the first administration, eligible employees
elected to participate in the study. Two-hundred and
thirty-six employees were in non-management positions,
47 were line managers, 15 were middle managers, and 10
were in senior management positions. One hundred and
seventy two respondents were males and 136 were fe-
males. Their average age was 35 and their average tenure
was five years. Ninety-one percent of the sample was
White Caucasian, 5% was Native American, and the
remaining 4% was of other ethnic origins.
3.3. Measures
Antisocial behavior. We used six items (‘‘Saying or
doing something to purposely hurt other coworkers
while at work,’’ ‘‘Saying unkind things to purposely
harm other coworkers while at work,’’ ‘‘Doing unkind
things to purposely harm other coworkers while at
work,’’ Criticizing other coworkers while at work,’’
Saying nasty things to other coworkers while at work,’’
‘‘Starting arguments with other coworkers while at
work’’) from a 13-item scale used by Douglas and
Martinko (2001) that describe potentially harmful ac-
tions directed towards co-workers. Their scale was
originally adapted from Robinson and OÕLeary-KellyÕs
(1998) 9-item Individual Anti-Social Behavior measure.
Respondents indicated how often they performed these
behaviors on a 5-point Likert scale format (1 ¼Never,
2¼1–3 times, 3 ¼4–6 times, 4 ¼7–9 times, and
5¼More than 10 times).
Identity threat. Nine items measured this construct.
The items were drawn from previous instruments used
to measure harmful workplace behaviors directed
against the employee by others. We selected items from
these instruments that we believed would best reflect the
concept of an identity threat as we have defined it;
namely, an overt action by another party that chal-
lenges, calls into question, or diminishes a personÕs sense
200 K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208

of competence, dignity or self-worth. We used two items
(‘‘Did something to make you look bad,’’ ‘‘Swore at
you’’) from a study by Aquino, Grover, Bradfield, and
Allen (1999). The remaining seven items (‘‘Made in-
sulting comments about your private live,’’ ‘‘Looked at
you in a negative way,’’ ‘‘Judged your work in an unjust
manner,’’ ‘‘Criticized you unfairly,’’ ‘‘Questioned your
abilities or judgments,’’ ‘‘Embarrassed you in front of
your coworkers,’’ ‘‘Unfairly blamed you for a negative
outcome’’) were adapted from the Workplace Harass-
ment Scale (Bj€
oorkqvist et al., 1994). For each item, re-
spondents reported the number of times that one or
more coworkers displayed the target behavior towards
them within the past 6 months using a 5-point Likert
scale (1 ¼Never, 2 ¼1–3 times, 3 ¼4–6 times, 4 ¼7–9
times, and 5 ¼10 or more times). Furthermore, we asked
them to report only those behaviors that caused them to
experience psychological or emotional discomfort. This
was done so that we measured only those incidents that
were likely to be experienced by respondents as threat-
ening to their self-concepts.
Aggressive modeling. We measured this construct us-
ing nine items adapted from the Individual Anti-Social
Behavior scale developed by Robinson and OÕLeary-
Kelly (1998). The nine adapted items ask the respondents
to indicate the extent they observed their coworkers en-
gaging in antisocial workplace behaviors (e.g., Please
indicate the number of times you have observed your
coworkers: ‘‘saying rude things about other coworkers or
the organization while at work’’ and ‘‘deliberately
bending or breaking rules while at work.’’). The re-
sponses were provided in the first questionnaire using a
5-point Likert format (1 ¼Never, 2 ¼1–3 times, 3 ¼4–6
times, 4 ¼7–9 times, and 5 ¼10 or more times).
The measures of identity threat, antisocial behavior,
and aggressive modeling presumably tap conceptually
distinct constructs because we varied the referent for
each behavior. The items measuring identity threat ask
the respondent to report behaviors directed towards
them, the antisocial behavior items asks them to report
on behaviors they directed against others, and the ag-
gressive modeling items ask about behaviors exhibited
by co-workers in general. However, because many of the
behaviors have a similar negative content, one could
argue that these items tap a common construct. We
examined this possibility by using confirmatory factor
analysis (CFA) to compare a single factor model with
the proposed three-factor model. LISREL 8.3 (Joreskog
& Sorbom, 1999) was used for this analysis. The sample
correlation matrix was used as data input for the CFA.
Av2difference test showed that the three-factor model
fit the data significantly better than a single-factor model
(Dv2¼1015:12, df ¼3, p<:001), indicating that the
items do not tap the same underlying construct.
We wanted to establish that the items underlying the
proposed three-factor model were unidimensional, so
we assessed the fit of the three-factor model to the data.
We used the root-mean-squared error of approximation
(RMSEA) and the comparative fit index (CFI) as in-
dices of fit. The results showed that these fit statistics
failed to reach the recommended levels (Bollen, 1989):
RMSEA ¼:11 and CFI ¼:85. When a model fails to
achieve a good fit, deleting problematic indicators is an
acceptable solution (Anderson & Gerbing, 1988). In-
spection of modification indices and standardized re-
siduals indicated that fit could be improved by deleting
two items from the antisocial behavior measure
(‘‘Criticizing other coworkers while at work,’’ ‘‘Saying
nasty things about other coworkers while at work’’),
three items from the aggressive modeling measure
(‘‘Saying or doing something to purposely hurt other
coworkers while at work,’’ ‘‘Criticizing other coworkers
while at work,’’ ‘‘Starting arguments with other co-
workers while at work’’), and one item from the iden-
tity threat measure (‘‘Swore at you’’). After deleting
these items, the fit statistics fell within acceptable ran-
ges: RMSEA ¼:08 and CFI ¼:93. Furthermore, the
parameter estimates for all the indicators were signifi-
cant. Based on these results, we concluded that the
remaining items showed acceptable unidimensionality
and were combined to form scales. The CronbachÕsa
reliabilities were .88, .93, and .86 for antisocial behav-
ior, identity threat, and aggressive modeling scales, re-
spectively.
Demographics. Gender and age were measured by
self-reports. Gender was dummy coded for the regres-
sion analysis (female ¼0, male ¼1).
Attitude towards revenge. The 20-item Vengeance
Scale (Stuckless & Goranson, 1992) was used to measure
this construct. The scale measures the extent a person
possesses a positive attitude toward seeking vengeance.
Representative items are ‘‘I try to even the score with
anyone who hurts me,’’ ‘‘I believe in the motto an eye
for eye, a tooth for a tooth,’’ and ‘‘If someone causes me
trouble, IÕll find a way to make them regret it.’’ Re-
spondents answered on a seven-point (1 ¼Strongly dis-
agree, 7 ¼Strongly agree) Likert scale. Items were
summed to produce a scale score such that high scores
indicate a more positive attitude towards revenge
(a¼:95).
Hierarchical status. We measured this variable by
asking respondents to indicate whether they were em-
ployed in a non-management, line management, middle
management, or senior/executive management position.
Line-management was described as supervising non-
management personnel and middle management was
described as managing managers. The responses were
coded such that higher values indicated higher status
(1 ¼non-management, 4 ¼senior/executive manage-
ment).
Control variables. We controlled for organizational
factors that might be related to antisocial behavior but
K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208 201

that were not directly measured by creating two dummy
variables to represent the school district and the trans-
portation company. The municipality served as the ref-
erence category.
4. Results
Table 1 shows the means, standard deviations, and
correlations for all study variables.
Hierarchical regression analysis was used to test the
hypotheses. The variables forming the interactions were
centered to minimize multicollinearity among the inter-
actions and their individual components (Aiken & West,
1991). The results of the regression analysis are shown in
Table 2. The table presents the unstandardized beta
weights for the predictor variables, the total R2at each
step, and the DR2for steps 2–4.
The set of main effects accounted for a significant
amount of variance in antisocial behavior (DR2¼:40,
Table 1
Means, standard deviations, and correlations among study variables
Mean SD (1) (2) (3) (4) (5) (6) (7)
(1) Antisocial behavior 5.66 3.28 (.88)
(2) Identity threat 14.82 7.53 .53 (.93)
(3) Attitude toward revenge 55.41 26.23 .52 .47 (.95)
(4) Hierarchical status 1.35 .72 ).07 ).07 ).08 —
(5) Aggressive modeling 17.00 6.75 .26 .37 .20.27 (.86)
(6) Age 3.58 1.10 ).08 .01 ).05 .15 .05 —
(7) Gender .56 .50 .17.12 .33 .12.04 ).03 —
Note.(N¼308). CronbachÕsareliabilities for the scales are shown along the diagonal.
*
p6:05.
**
p6:01.
Table 2
Results of hierarchical regression analysis for antisocial behaviora
Variables Step 1:
Controls
Step 2:
Main effects
Step 3:
Two-way interaction
Step 4:
Three-way interaction
Constant 5.85 .82 1.68 2.36
Controls
School employee ).413 .732 .335 .343
Transportation employee ).195 .461 ).236 ).232
Main effects
Identity threat .157 .131 .090
Gender .511 .470 .478
Age ).220 ).210 ).209
Attitude towards revenge .042 .034 .034
Aggressive modeling .031 .037 .055
Hierarchical status ).085 ).293 ).667
Two-way interaction
Identity threat Gender .035 .038
Identity threat Age ).037
).040
Identity threat Attitude toward revenge .002 .002
Identity threat Aggressive modeling .004 .012
Identity Threat Hierarchical status .012 ).131
Hierarchical status Aggressive modeling .064 .125
Three-way interaction
Identity threat Hierarchical status 
Aggressive modeling
.028
R2.01 .41 .46 .49
DR2.01 .40 .05 .03
a
Entries are unstandardized betas.
*
p6:05.
**
p6:01.
***
p6:001.
202 K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208

p6:001) and, as predicted in Hypothesis 1, identity threat
was positively related to the criterion (B¼:157, p6:001).
Although not hypothesized, the direct effects of attitude
toward revenge (B¼:042, p6:001) was also significant.
However, these main effects are qualified by the presence
of significant interactions (Aiken & West, 1991).
The set of two-way interactions accounted for a sig-
nificant amount of additional variance in antisocial be-
havior (DR2¼:05, p6:001). Inspection of the
individual regression weights for the third step reveals
that while the interactions involving identity threat and
age (B¼:037, p6:05) and identity threat and attitude
towards revenge (B¼:002, p6:01) were significant, the
interactions between identity threat and gender
(B¼:035, p>:05), identity threat and aggressive
modeling (B¼:004, p>:05), and identity threat and
hierarchical status (B¼:012, p>:05) were not. These
results provide initial support for Hypotheses 3 and 4
and fail to support Hypotheses 2, 5 and 6a.
We explored the form of the significant two-way in-
teractions found in step 3 as recommended by Aiken
and West (1991). For each interaction we calculated the
simple slopes of antisocial behavior on identity threat
and their standard errors at three levels (i.e., mean, one
standard deviation above and one standard deviation
below the mean) of the second predictors (i.e., age and
attitudes toward revenge) as suggested by Cohen and
Cohen (1983). Next, we conducted ttests on the values
of the simple slopes divided by their standard errors.
The results of these analyses are shown in Table 3 and
illustrated in Figs. 1 and 2. These results indicate that
the positive relationship between identity threat and
antisocial behavior is stronger for younger employees
than older employees, and for employees who have more
rather less positive attitudes toward revenge. Addition-
ally, since the regression coefficients for the two-way
interactions are significant in the overall regression
analysis, we know that the simple slopes of antisocial
behavior on identity threat differ as a function of the
second predictors (Aiken & West, 1991). These findings
support Hypotheses 3 and 4.
The results for step 4 of the regression analysis offer
conditional support for Hypotheses 5 and 6a. In Step 4,
both the three-way interaction involving identity
threat, aggressive modeling and hierarchical status
(B¼:028, p6:001, DR2¼:03, p6:001) and the two-
way interactions involving identity threat and aggres-
sive modeling (B¼:012, p6:01) and identity threat
Fig. 1.
Table 3
Results of standard errors and ttests for simple slopes of two-way interactions including identity threat and second predictors
Second predictor Simple slope SE t test Intercepta
Age
High .090 .019 4.75 5.92
Mean .131 .020 6.55 6.15
Low .172 .021 8.12 6.38
Attitude towards revenge
High .183 .050 3.67 11.46
Mean .131 .020 6.55 11.21
Low .079 .045 1.74 ns 10.96
ns, Not significant.
a
Predicted value for antisocial behavior when identity threat is at its centered mean (i.e., X¼0).
***
p6:001.
Fig. 2.
K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208 203

and hierarchical status (B¼:131, p6:05) are signif-
icant. These results can be interpreted as showing that
the significance of the two-way interactions depends
upon the value of the omitted third variable. Specifi-
cally, the two-way interaction between identity threat
and aggressive modeling is significant when the value
for hierarchical status is at its mean. Similarly, the two-
way interaction between identity threat and hierarchi-
cal status is significant when the value for aggressive
modeling is at its mean. Further examination of the
data revealed that the two-way interaction between
identity threat and aggressive modeling is significant
when the value for hierarchical status is one standard
deviation below its mean, but is not significant when
the value for hierarchical status is at one standard
deviation above its mean. Also, while the two-way in-
teraction between identity threat and hierarchical status
is significant when the value for aggressive modeling is
one standard deviation above its mean, it is not sig-
nificant when the value for aggressive modeling is one
standard deviation below its mean.
We probed the form of the significant three-way
interaction using procedures recommended by Aiken
and West (1991). First, we calculated the simple slopes
of antisocial behavior on identity threat and their
standard errors at two levels (i.e., one standard devi-
ation above and one standard deviation below the
mean) of the second and third predictors (i.e., hier-
archical status and aggressive modeling). We then
conducted ttests on the values of the simple slopes
divided by their standard errors. The results of these
analyses are shown in Table 4 and illustrated in Fig.
3. These results indicate that the positive relationship
between identity threat and antisocial behavior is
stronger for employees who have low as compared to
high organizational status, but only when the em-
ployee is exposed to few rather than many aggressive
role models. These results support Hypothesis 6b.
These results also show that in the presence of ag-
gressive models, high status persons are more likely to
exhibit antisocial behavior in response to identity
threats than are low status persons.
5. Discussion
This study is the first to directly link identity threat to
antisocial behavior in organizations. There is ample
evidence that people are strongly motivated to maintain
positive self and social identities and that identity-
threatening events can provoke an aggressive behavioral
response. This relationship was the basis for our theo-
retical model and was supported by our data. However,
the unique contribution of this study was to show that
the relationship is more complex because it is moderated
by individual-level and contextual variables.
5.1. Theoretical implications
Past studies support direct relationships between an-
tisocial behavior and specific individual level predictors,
but these variables typically explain relatively little
variance (Robinson & Greenberg, 1998). Our findings
suggest that if individual level variables play a role, it is
more likely as moderators of the relationship between
provocative events and responses to those events. Con-
temporary research on various forms of antisocial be-
havior support this view. For example, the personality
trait of negative affectivity has been found to moderate
the relationship between perceived unfairness and re-
taliatory behavior (Skarlicki et al., 1999). Other writers
argue that cognitive tendencies such as hostile attribu-
Table 4
Results of standard error and ttest for simple slopes of three-way interaction including identity threat, hierarchical status and aggressive modeling
Second and third predictors Simple slope SE t test Intercepta
High status High aggressive modeling .212 .103 2.067.50
Low status High aggressive modeling .130 .106 1.22 7.24
High status Low aggressive modeling ).222 .155 1.43 5.54
Low status Low aggressive modeling .240 .092 2.61 7.72
ns, Not significant.
a
Predicted value for antisocial behavior when identity threat is at its centered mean (i.e., X¼0).
*
p6:05.
***
p6:001.
Fig. 3.
204 K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208

tion bias intensify a personÕs reactions to perceived of-
fenses (Greenberg & Alge, 1998).
We investigated an attitude that has a logical con-
nection to a personÕs responses to identity threat. As
expected, the data showed that having favorable atti-
tudes toward revenge strengthened the relation between
identity threat and antisocial behavior. The interpreta-
tion of these findings is straightforward. People who
hold favorable attitudes toward revenge are more likely
to endorse retaliation against perceived mistreatment as
an acceptable and appropriate moral response. How-
ever, we also suggested that favorable attitudes toward
revenge might motivate antisocial behavior performed
for value-expressive rather than purely instrumental
purposes. Unfortunately, our data do not allow us to
test this argument, leaving it to future research to dis-
cover whether revenge attitudes moderate relations be-
tween identity threat and acts directed against only the
perpetrator or against convenient and powerless targets.
Apart from attitudes, we found that age moderates
the relationship between identity threat and antisocial
behavior. In support of disengagement and socioemo-
tional selectivity theories, our data suggest that older
persons may react less strongly to identity threats than
younger ones. Furthermore, it is noteworthy that age
alone did not predict antisocial behavior, but only in
combination with a provocation like identity threat.
This finding corroborates recent articles in the popular
press suggesting that younger employees are more likely
to blame and criticize others when dealing with negative
workplace events (Neusner, Basso, Brenna, & Lobet,
2001). Clearly, the influence of age and social maturity
on antisocial behavior is a fruitful and relatively ne-
glected area for further investigation.
We did not find the expected moderating effect of
gender, nor did we find a direct relationship between
gender and antisocial behavior. A possible explanation
for this null-result is suggested by the findings of a recent
meta-analysis by Bettencourt and Miller (1996). Al-
though many studies have shown that men express anger
more readily than women, and are generally more prone
to display aggressive behavior, Bettencourt and Miller
found that when provocation (e.g., identity threat) is
considered, these gender differences are greatly attenu-
ated. Consequently, their findings suggest that taking
identity threat into account might have diminished the
strength of both the direct and moderating effects of
gender on antisocial behavior. Notwithstanding this ex-
planation, we urge caution in concluding from data that
gender plays no moderating role because it may be that its
effect is more pronounced for certain kinds of antisocial
behavior. For example, Bj€
oorkqvist (1994) argues that
while men are more likely to engage in direct expressions
of aggression or anger, women engage in more indirect
expressions (e.g., social manipulation and gossip).
Therefore, had we measured different types of antisocial
behaviors, such as those that were less direct, we may have
found significant direct and moderating effects of gender.
Some studies have examined social norms as predic-
tors of antisocial behavior (e.g., Greenberg, 1998;
Robinson & OÕLeary-Kelly, 1998). These studies rec-
ognize that local workgroup norms may support certain
forms of antisocial behavior even when such behavior
violates norms espoused by the dominant administrative
coalition, OÕLeary-Kelly and her colleagues (OÕLeary-
Kelly et al., 1996) suggest these norms influence anti-
social behavior through social modeling (Bandura,
1973). Robinson and OÕLeary-Kelly (1998) provide
empirical evidence supporting this argument by showing
that aggressive modeling in workgroups affects individ-
ual group membersÕantisocial behavior. Our results
extend these findings by showing that the effect of ag-
gressive modeling on responses to identity threat depend
on the level of a third variable: the threat-recipientÕs
hierarchical status.
The question of how an employeeÕs status within a
organizational hierarchy might influence his or her will-
ingness to exhibit antisocial behavior in response to
provocation has been relatively unexplored. Our data
show that this question warrants further investigation as
we found that the relation between identity threat and
antisocial behavior was stronger among low than high
status employees when the social environment provides
few aggressive role models. This result parallels Aquino
et al.Õs (2001) finding that the relationship between blame
attribution and revenge is stronger for low than high
status employees. However, our study extends Aquino
et al.Õs (2001) work by showing that this difference only
occurred in the absence of aggressive role models. We
interpret this pattern as providing indirect support for
our social constraint argument, which posits that high
status employees are more sensitive to differences in
workplace norms that discourage aggression compared
to those with low status. As a result, in the absence of
these norms the relation between identity-threat and
antisocial behavior is less influenced by status differences.
Interestingly, we also found that the positive rela-
tionship between identity threat and antisocial behavior
was stronger for high as compared to low status em-
ployees in the presence of aggressive models. Although
we did not predict this result, it is consistent with the
underlying premise of the social restraint argument;
namely, that high status employees are more sensitive to
workplace norms as compared to those with low status.
What this particular finding suggests is that research on
antisocial behaviors committed by high-ranking em-
ployees may want to consider how the presence of ag-
gressive role models might influence their willingness to
engage in these acts. For example, it is plausible that
antisocial behaviors like abusive supervision (Tepper,
2000) or petty tyranny (Ashforth, 1994) committed by
high status employees may be strongly influenced by the
K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208 205

aggressive behaviors displayed by peers, superiors, or
perhaps even subordinates.
5.2. Managerial implications
A wave of younger employees has entered the work-
force to replace the aging Baby Boom generation (Ful-
lerton & Toossi, 2001). If younger employees react more
strongly to self-invalidating events, as our findings sug-
gest, then this demographic trend portends the possi-
bility that workplace conflict and antisocial behavior
may increase in the near future. Given current concerns
about antisocial workplace behavior, practicing man-
agers should place additional emphasis on establishing
conflict resolution mechanisms such as grievance sys-
tems and peer mediation programs that enable younger
employees to vent their frustrations while avoiding po-
tentially harmful and disruptive behaviors. However,
the social attitudes people bring with them to work (e.g.,
attitudes toward revenge) may detract from the effec-
tiveness of such programs. But because attitudes are the
product of learning and experience, the proclivity to
exhibit antisocial behavior in response to identity threats
may be modified by changing peopleÕs beliefs about the
acceptability of revenge.
Organizational environments are replete with cues
indicating the norms for acceptable workplace behavior.
One of the strongest is the behavior exhibited by co-
workers. Our results show that aggressive modeling may
influence whether an employee engages in antisocial be-
havior in response to identity threats. If so, then practi-
tioners can discourage antisocial acts and reinforce more
suitable forms of behavior by fostering ethical climates
that encourage dignified and respectful conduct and by
instituting and reinforcing zero-tolerance policies. Our
results further suggest that in the absence of environ-
ments that reinforce antisocial behavior, employees of
lower organizational status may be more likely to react
antisocially to identity threats. Given the inevitability of
status differences in organizations, managers might fur-
ther enhance the development of ethical climates and
reduce antisocial workplace behavior by implementing
stress and anger management programs specifically tar-
geted toward lower status employees.
5.3. Limitations
Like all studies, ours has limitations that deserve
comment. One is the cross-sectional nature of the de-
sign. Since we did not collect data over time, we cannot
establish the direction of causality. It may be that an-
tisocial behavior leads to identity threat, rather than
the other way around. However, much of the theoriz-
ing on antisocial behavior presumes, as we do, that
these acts occur in response to some triggering event or
series of events (e.g., the self-reported frequency of
identity threats experienced by the employee). What we
can say more definitively based on our results is that
the relationship between identity threat and antisocial
behavior appears to be more complex than a simple
direct effect.
All of our data were collected by questionnaires and
the measures of identity threat and antisocial behavior
were collected at the same time. Consequently, the
threat of common method variance is a second limita-
tion of the study. We note, however, that the data for
moderating variables and the dependent variable was
collected on two separate administrations several days
apart. This procedure should presumably reduce con-
cerns about common method variance inflating corre-
lations among the attitudinal and behavioral measures
(Podsakoff & Organ, 1986). We would also point out
that the concern over common method bias is further
mitigated in our study because we found several signif-
icant two- and three-way interactions. Percept-percept
inflation is a linear confound that appears as an additive
bias in correlation and regression analyses (Podsakoff &
Organ, 1986), therefore the presence of significant in-
teractions helps to strengthen our argument that the
observed relations are a function of the constructs being
studied rather than a methodological artifact.
A third limitation is the self-report nature of the data.
Reports of antisocial behavior may be particularly sus-
ceptible to demand characteristics that could result in
underreporting of actual behavior. But other studies
have shown that people are surprisingly willing to report
having engaged in deviant, even illegal, behavior (Ben-
nett & Robinson, 2000). Further, if respondents under-
reported their antisocial behavior, it would make it more
difficult to detect significant relations due to a restriction
of range in the variance for the dependent variable.
Thus, given that our model accounted for 49% of the
variance in the dependent measure and that we found
significant relations, including two- and three-way in-
teractions, it is not obvious that respondents under-re-
ported antisocial behaviors. It also seems unlikely that
the complex interactions we found in our study are
better explained by a differential tendency among certain
groups of people (e.g., those who are exposed to low
levels of aggressive behavior and who also occupy high
rather than low status positions) to systematically un-
derreport engaging in antisocial behavior rather than
the theoretical mechanisms we have proposed.
In short, despite the limitations stated above, we be-
lieve our study makes a unique contribution to the lit-
erature by showing that both individual level and
contextual variables matter in predicting antisocial be-
havior as a function of identity-threatening events. So
while most people feel compelled to defend themselves
against threatened self-identities, some may experience
and act upon this need more often and perhaps with
greater intensity than others.
206 K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208

References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and
interpreting interactions. Thousand Oaks, CA: Sage.
Anderson, J. G., & Gerbing, D. W. (1988). Structural equation
modeling in practice: A review and recommended two-step
approach. Psychological Bulletin, 103, 411–423.
Andersson, L. M., & Pearson, C. M. (1999). Tit for tat? The spiraling
effect of incivility in the workplace. Academy of Management
Review, 24, 452–471.
Aquino, K., Grover, S. L., Bradfield, M., & Allen, D. G. (1999). The
effects of negative affectivity, hierarchical status, and self-determi-
nation on workplace victimization. Academy of Management
Journal, 42, 260–272.
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, 1073–1091.
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, 52–59.
Ashforth, B. (1994). Petty tyranny in organizations. Human Relations,
47, 755–778.
Averill, J. (1982). Anger and aggression: An essay on emotion. New
York: Springer.
Bandura, A. (1973). Aggression: A social learning analysis. Englewood
Cliffs, NJ: Prentice-Hall.
Bandura, A., Ross, D., & Ross, S. (1963). Imitation of film mediated
aggressive models. Journal of Abnormal Psychology, 66, 3–11.
Baron, R. A., & Neuman, J. H. (1996). Workplace violence and
workplace aggression: Evidence of their relative frequency and
potential causes. Aggressive Behavior, 22, 161–173.
Baumeister, R. F., Smart, L., & Boden, J. M. (1996). Relation of
threatened egotism to violence and aggression: The dark side of
high self-esteem. Psychological Review, 103, 5–33.
Bennett, R. J., & Robinson, S. L. (2000). The development of a measure
of workplace deviance. Journal of Applied Psychology, 85, 349–360.
Berkowitz, L. (1993). Aggression: Its causes, consequences, and control.
Philadelphia: Temple University Press.
Bettencourt, B. A., & Miller, N. (1996). Gender differences in
aggression as a function of provocation: A meta-analysis. Psycho-
logical Bulletin, 119, 422–447.
Bies, R. J. (1999). Interactional (in)justice: The sacred and the profane.
In J. Greenberg & R. Cropanzano (Eds.), Advances in organiza-
tional behavior. San Francisco: The New Lexington Press.
Bies, R. J., Tripp, T. M., & Kramer, R. M. (1997). At the breaking
point: Cognitive and social dynamics of revenge in organizations.
In R. Giacalone & J. Greenberg (Eds.), Anti-social behavior in
organizations (pp. 18–36). Thousand Oaks, CA: Sage.
Bj€
oorkqvist, K. (1994). Sex differences in physical, verbal, and indirect
aggression: A review of recent research. Sex Roles, 30, 177–188.
Bj€
oorkqvist, K.,
€
OOsterman, K., & Hjelt-B€
aack, M. (1994). Aggression
among university employees. Aggressive Behavior, 20, 173–184.
Black, K. S. (1990). Can getting mad get the job done? Working
Woman, 86–90.
Boehm, C. (1993). Egalitarian behavior and reverse dominance
hierarchy. Current Anthroplogy, 34, 227–254.
Bollen, K. A. (1989). Structural equations with latent variables. New
York: Wiley.
Brockner, J. (1988). The effects of work layoffs on survivors: Research,
theory, and practice. In B. M. Staw & L. L. Cummings (Eds.),
Research in organizational behavior (Vol. 10, pp. 213–255). Green-
wich, CT: JAI Press.
Cartensen, L. L. (1992). Social and emotional support patterns in
adulthood: Support for socioemotional selectivity theory. Psychol-
ogy and Aging, 7, 331–338.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation
analyses for the behavioral sciences (2nd ed.). Hillsdale, NJ:
Erlbaum.
Cumming, E., & Henry, W. (1961). Growing old: The process of
disengagement. New York: Basic Books.
Daly, M., & Wilson, M. (1988). Homicide. New York: Aldine De
Gruyter.
Darwin, C. (1872). The expression of emotion in men and animals. New
York: Scribners.
Davis, M. A., LaRosa, P. A., & Foshee, D. P. (1992). Emotion work in
supervisor–subordinate relations: Gender differences in the percep-
tions of anger displays. Sex Roles, 26, 513–531.
Douglas, S. C., & Martinko, M. J. (2001). Exploring the role of
individual differences in the prediction of workplace aggression.
Journal of Applied Psychology, 86, 547–559.
Ehrlich, H. J., & Larcom, B. E. K. (1994). Ethnoviolence in the
workplace. Baltimore: Center for the Applied Study of Ethnovio-
lence.
Erez, M., & Early, P. (1993). Culture, self-identity, and work. New
York: Oxford University Press.
Felson, R. (1992). ‘‘KickÕem when theyÕre down’’: Explanations of the
relationship between stress and interpersonal aggression and
violence. Sociological Quarterly, 33, 1–16.
Felson, R., & Steadman, H. (1983). Situational factors in disputes
leading to criminal violence. Criminology, 21, 59–74.
Fishbein, M., & Ajzen, I. (1975). Belief, intention, and behavior.
Reading, MA: Addison-Wesley.
Folger, R., & Skarlicki, D. P. (1998). A popcorn metaphor for
employee aggression. In R. W. Griffin, A. OÕLeary-Kelly, & J. M.
Collins (Eds.), Dysfunctional behavior in organizations (pp. 43–81).
Stamford, CT: JAI Press.
Freud, S. (1950). Totem and taboo. New York: Norton.
Fullerton, H. N., & Toossi, M. (2001). Labor force projections to 2010:
Steady growth and changing composition. Monthly Labor Review,
November.
Geen, R. G. (1990). Human aggression. Pacific Grove, CA: Brooks-
Cole.
Gilligan, J. (1996). Violence: Our deadly epidemic and its causes. G.P.
Putnam.
Glomb, T. M., & Hulin, C. L. (1997). Anger and gender effects in
observed supervisor–subordinate dyadic interactions. Organiza-
tional Behavior and Human Decision Processes, 27, 281–307.
Goffman, E. (1967). Interaction ritual. New York: Pantheon.
Greenberg, J. (1998). The cognitive geometry of employee theft:
Negotiating ‘‘The Line’’ between taking and stealing. Dysfunc-
tional behavior in organizations: Non-violent dysfunctional be-
havior. In R. W. Griffin, A. OÕLeary-Kelly, & J. M. Collins (Eds.),
Monographs in organizational behavior and industrial relations (Vol.
23(B), pp. 147–193). Stanford, CT: JAI Press.
Greenberg, J., & Alge, B. J. (1998). Aggressive reactions to workplace
injustice. In R. W. Griffin, A. OÕLeary-Kelly, & J. M. Collins
(Eds.), Dysfunctional behavior in organizations: Violent and deviant
behavior (pp. 83–117). Stanford, CT: JAI Press.
Hogan, R., & Emler, N. P. (1981). Retributive justice. In M. J. Lerner
& S. C. Lerner (Eds.), The justice motive in social behavior (pp. 125–
143). New York: Plenum.
Joreskog, K. G., & Sorbom, D. (1999). LISREL 8.3: Structural
equation modeling with the SIMPLIS command language. Hillsdale,
NJ: Scientific Software.
Kogut, D., Langely, T., & OÕNeal, E. C. (1992). Gender role masculinity
and angry aggression in women. Sex Roles, 26, 355–368.
Lind, E. A. (1997). Litigation and claiming in organizations. In R.
Giacalone & J. Greenberg (Eds.), Anti-social behavior in organiza-
tions (pp. 150–171). Thousand Oaks, CA: Sage.
Lind, E. A., & Tyler, T. R. (1988). The social psychology of procedural
justice. New York: Plenum.
K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208 207

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.
Martinko, M. J., & Zellars, K. L. (1998). Toward a theory of
workplace violence: A cognitive appraisal perspective. In R.
Griffin, A. OÕLeary-Kelly, & J. Collins (Eds.), Dysfunctional
behavior in organizations: Violent and deviant behavior (pp. 1–42).
Stanford, CT: JAI Press.
Neugarten, B. L. (1996). Time, age, and the life cycle. In D. A.
Neugarten (Ed.), The meanings of age (pp. 114–127). Chicago: The
University of Chicago Press.
Neusner, N., Basso, P., Brenna, S., & Lobet, I. (2001, September). The
boomersÕkids get a job. US News and World Report, 3, 28.
OÕLeary-Kelly, A., Griffin, R. W., & Glew, D. J. (1996). Organization-
motivated aggression: A research framework. Academy of Man-
agement Review, 21, 225–253.
Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organiza-
tional research: Problems and prospects. Journal of Management,
12, 531–544.
Robinson, S. L., & Greenberg, J. (1998). Employees behaving badly:
Dimensions, determinants, and dilemmas in the study of work-
place deviance. In C. L. Cooper & D. M. Rousseau (Eds.),
Trends in organizational behavior (Vol. 5, pp. 1–30). New York:
Wiley.
Robinson, S. L., & OÕLeary-Kelly, A. (1998). Monkey see, monkey do:
The influence of work groups on the antisocial behavior of
employees. Academy of Management Journal, 41, 658–672.
Romano, C. (1994). Workplace violence takes a deadly turn.
Management Review, 83,5.
Ross, C. E., & Van Willigen, M. (1996). Gender, parenthood, and
anger. Journal Marriage and the Family, 58, 572–584.
Salancik, G. J., & Pfeffer, J. (1978). A social information processing
approach to job attitudes and task design. Administrative Science
Quarterly, 23, 224–253.
Scheiman, S. (1999). Age and anger. Journal of Health and Social
Behavior, 40, 273–289.
Sennett, R., & Cobb, J. (1973). The hidden injuries of class. New York:
Random House.
Skarlicki, D., & Folger, R. (1997). Retaliation in the workplace: The
roles of distributive, procedural, and interactional justice. Journal
of Applied Psychology, 82, 434–443.
Skarlicki, D., Folger, R., & Tesluk, P. (1999). Personality as a
moderator in the relationship between fairness and retaliation. The
Academy of Management Journal, 42, 100–110.
Smith, K. C., Ulch, S. E., Cameron, J. E., Cumberland, J. A.,
Musgrave, M. A., & Tremblay, N. (1989). Gender-related effects in
the perception of anger expression. Sex Roles, 20, 487–499.
Speilberger, C. D. (1996). State-trait anger expression inventory,
research edition: Professional manual. Odessa, FL: Psychological
Assessment Resources Inc.
Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the
integrity of the self. In L. Berkowitz (Ed.), Advances in experimental
psychology (Vol. 21, pp. 261–302). New York: Academic Press.
Stuckless, N., & Goranson, R. (1992). The vengeance scale: Develop-
ment of a measure of attitudes toward revenge. Journal of Social
Behavior and Personality, 7, 25–42.
Tavris, C. (1982). Anger: The misunderstood emotion. New York:
Simon & Schuster.
Tavris, C. (1984). On the wisdom of counting to 10: Personal and
social dangers of anger expression. In P. Shaver (Ed.), Review of
personality and social psychology: Emotions, relationships, and
health. Beverly Hills, CA: Sage.
Tajfel, H., & Turner, J. C. (1985). The social identity theory of intergroup
behavior. In S. Worchel & W. G. Austin (Eds.), Psychology of
intergroup relations (2nd ed.., pp. 7–24). Chicago: Nelson-Hall.
Tedeschi, J., & Felson, R. (1994). Violence, aggression, and coercive
action. Washington, DC: American Psychological Association.
Tepper, B. J. (2000). Consequences of abusive supervision. Academy of
Management Journal, 43, 178–190.
Thomas, K. W. (1992). Conflict and negotiation processes in organi-
zations. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of
industrial and organizational psychology (pp. 651–717). Palo Alto,
CA: Consulting Psychologists Press.
Wicklund, R. A., & Gollwitzer, P. M. (1982). Symbolic self-completion.
Hillsdale, NJ: Erlbaum.
Received 7 March 2001
208 K. Aquino, S. Douglas / Organizational Behavior and Human Decision Processes 90 (2003) 195–208