Content uploaded by James R. Detert
Author content
All content in this area was uploaded by James R. Detert on Nov 25, 2014
Content may be subject to copyright.
Moral Disengagement in Ethical Decision Making:
A Study of Antecedents and Outcomes
James R. Detert
Cornell University
Linda Klebe Trevin˜o
The Pennsylvania State University
Vicki L. Sweitzer
Albion College
This article advances understanding of the antecedents and outcomes of moral disengagement by testing
hypotheses with 3 waves of survey data from 307 business and education undergraduate students. The
authors theorize that 6 individual differences will either increase or decrease moral disengagement,
defined as a set of cognitive mechanisms that deactivate moral self-regulatory processes and thereby help
to explain why individuals often make unethical decisions without apparent guilt or self-censure
(Bandura, 1986). Results support 4 individual difference hypotheses, specifically, that empathy and
moral identity are negatively related to moral disengagement, while trait cynicism and chance locus of
control orientation are positively related to moral disengagement. Two additional locus of control
orientations are not significantly related to moral disengagement. The authors also hypothesize and find
that moral disengagement is positively related to unethical decision making. Finally, the authors
hypothesize that moral disengagement plays a mediating role between the individual differences they
studied and unethical decisions. Their results offer partial support for these mediating hypotheses. The
authors discuss the implications of these findings for future research and for practice.
Keywords: moral disengagement, ethical decision making, empathy, trait cynicism, moral identity
Recent years have brought unrelenting news about unethical
behavior in virtually every sector of society (e.g., business, gov-
ernment, education, military, sports, religious institutions). The
tales of executives who enrich themselves at shareholders’ expense
(Horovitz, 2002; Samuelson, 2006), of corrupt government offi-
cials (Seglin, 2000), of cheating athletes (Lewis, 2006), and more
raise an obvious and critically important question, “Why do people
make unethical decisions?” Consistent with recent arguments that
the best explanations for unethical decision making may reside in
underlying psychological processes (Messick & Bazerman, 1996;
Tenbrunsel & Messick, 2004), we explore Bandura’s (1986) as-
sertion that people make unethical decisions when moral self-
regulatory processes that normally inhibit unethical behavior are
deactivated via use of several interrelated cognitive mechanisms
collectively labeled moral disengagement. Bandura (1986) argued
that moral disengagement explains why otherwise normal people
are able to engage in unethical behavior without apparent guilt or
self-censure.
Despite its potential importance for explaining unethical deci-
sion making, our understanding of moral disengagement remains
at an early stage. Research thus far has focused primarily on the
outcomes of moral disengagement, such as its positive relationship
to aggression in children (Bandura, Barbaranelli, Caprara, & Pas-
torelli, 1996; Bandura, Caprara, Barbaranelli, Pastorelli, & Rega-
lia, 2001; Bandura, Underwood, & Fromson, 1975) or its relation-
ship to decisions to support military action (Aquino, Reed, Thau,
& Freeman, 2007). Most importantly, we know little about the
antecedents of moral disengagement. While it is logical to assume
that individuals may differ in their propensity to morally disen-
gage, extant research has focused on simple demographics (e.g.,
age, nationality) as antecedents. For example, McAlister (2001)
found male subjects to be more morally disengaged than were
female subjects. Yet, if organizations knew more about whether
some individuals were more predisposed to moral disengagement
than others, perhaps they could target resources toward improving
these individuals’ decision making processes.
This study therefore seeks to contribute to knowledge about
moral disengagement by investigating the individual difference
antecedents of moral disengagement as well as the relationship
between moral disengagement and subsequent unethical decision
making. Figure 1 illustrates the key relationships we address. First,
we theorize that, because of relatively stable individual differences
associated with how individuals see others, events, and them-
selves, some people will be more predisposed to moral disengage-
James R. Detert, Johnson Graduate School of Management, Cornell
University; Linda Klebe Trevin˜o, Smeal College of Business, The Penn-
sylvania State University; Vicki L. Sweitzer, Economics and Management
Department, Albion College.
We gratefully acknowledge the advice received from Michael Brown,
Ethan Burris, Donald Hambrick, and David Harrison. We are also indebted
to Craig Crossland and Kristin Price for their research assistance and to the
Smeal College of Business for their financial support of this research
Correspondence concerning this article should be addressed to James R.
Detert, Cornell University, 318 Sage Hall, Ithaca, NY 14850. E-mail:
jdetert@cornell.edu
Journal of Applied Psychology Copyright 2008 by the American Psychological Association
2008, Vol. 93, No. 2, 374–391 0021-9010/08/$12.00 DOI: 10.1037/0021-9010.93.2.374
374
ment than others. We hypothesize that, because of their likely
facilitating or inhibitory influence on moral self-regulation pro-
cesses, the following individual differences will influence moral
disengagement: empathy, trait cynicism, locus of control orienta-
tions, and moral identity. We then investigate the relationship
between moral disengagement and unethical decision making,
defined as decisions to behave in ways that breach accepted moral
norms or standards of behavior. Finally, we explore whether moral
disengagement plays a mediating role between the individual
difference antecedents we examine and unethical decision making.
Theory and Hypotheses
Moral Disengagement Overview
Bandura (1999) developed the notion of moral disengagement
as an extension of social cognitive theory. Social cognitive theory
offers an agentic perspective on human behavior whereby individ-
uals exercise control over their own thoughts and behaviors
through self-regulatory processes (Bandura, 1986). According to
social cognitive theory, moral agency is governed by a self-
regulatory system that includes self-monitoring of one’s conduct
as well as self-reaction to that conduct in light of internal moral
standards. According to the theory, most people have developed
personal standards of moral behavior that serve a self-regulatory
role. These standards guide good behavior and deter bad behavior
because individuals use their personal standards to anticipate,
monitor, and judge their own actions. Behaving in ways that
counter these standards results in self-censure. Thus, individuals
usually behave in ways that are consistent with their internal moral
standards because they anticipate their own positive and negative
evaluations of possible conduct choices. However, this self-
regulatory function operates only if it is activated. Bandura (1999)
argued that moral self-regulation can be activated and deactivated
selectively, and he proposed moral disengagement as the key
deactivation process. Through moral disengagement, individuals
are freed from the self-sanctions and the accompanying guilt that
would ensue when behavior violates internal standards, and they
are therefore more likely to make unethical decisions.
Bandura (1986) suggested that moral self-regulation can be
deactivated or disengaged via eight interrelated moral disengage-
ment mechanisms: moral justification, euphemistic labeling, ad-
vantageous comparison, displacement of responsibility, diffusion
of responsibility, disregarding or distorting the consequences,
dehumanization, and attribution of blame. The first three mecha-
nisms—moral justification, euphemistic labeling, and advantageous
comparison—involve cognitive misconstrual of reprehensible behav-
ior in a way that increases its moral acceptability (Bandura, 1986). In
today’s society, for example, most people recognize that harming
others is wrong. However, with moral justification, individuals recon-
strue harm to others in ways that make it appear morally justifiable.
For example, hiring young children overseas may be justified by
stating that without such work the children would have to engage in
other more dangerous or degrading forms of employment to help their
desperate families. With euphemistic language, individuals use mor-
ally neutral language to make reprehensible conduct seem less harm-
ful or even benign. For example, lying to business competitors may be
called “strategic misrepresentation” (Safire, 1979), and killing civil-
ians in war may be referred to as “collateral damage” (Bandura,
1999). With advantageous comparison, unethical behaviors are com-
pared with even more harmful conduct, thus making the original
behavior appear acceptable. For instance, a student who asks another
student general questions about the content of an upcoming exam that
the latter has inappropriately obtained in advance may favorably
compare his or her actions with the behavior of other students who
review the specific exam questions by noting that just asking about
exam topics is less (or not) problematic compared with viewing the
entire exam.
The next three moral disengagement mechanisms, displacement
of responsibility, diffusion of responsibility, and distortion of
consequences, occur when an individual obscures or distorts the
effects of harmful actions (Bandura, 1986). When individuals view
their behaviors as a direct result of authoritative dictates (e.g., my
boss told me to do it), they may displace responsibility for their
actions to the authority figure, negating any personal accountabil-
ity for the unfavorable act. In addition, adverse group behavior
may trigger diffusion of responsibility because no one group
Antecedents Outcomes
Locus of Control:
• Internal
• Chance
• Power
Trait
Cynicism
Empathy
Moral
Disengagement
(Mediating hypotheses: H 6a-f)
H 5 Unethical
Decision
Making
H 4
H 3a-c
H2
H 1
Moral
Identity
Figure 1. Hypothesized antecedents and outcomes of moral disengagement. H ⫽hypothesis.
375
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
member feels personally liable for the collective’s destructive
behavior. These two moral disengagement tactics seem particu-
larly applicable in work organizations where individuals feel com-
pelled to follow the orders of authority figures (e.g., to change the
numbers in a report) and where responsibility for harmful out-
comes is often diffused to organizational teams or units. Individ-
uals may also disconnect harmful activities from self-sanctions by
distorting the consequences associated with a given act. For ex-
ample, customers may tell themselves that no one will be harmed
by not reporting an error in their favor because “this little bit of
money doesn’t affect anything in a huge company like X.”
Finally, dehumanization and attribution of blame can disengage
moral sanctions by reducing identification with the targets of
harmful acts. Research has consistently shown the tendency of
individuals to form groups and to quickly develop us-versus-them
thinking based on group membership (Brewer, 1979; Gaertner &
Insko, 2000). Such recasting of others as out-group members
makes it more likely that harm will be perpetuated because internal
standards and self-sanctions are less likely to be activated once
others have been cast as worthy of derogation or even lacking in
human qualities (Bar-Tal, 1989; Struch & Schwartz, 1989). Sim-
ilarly, attribution of blame can exonerate the self by placing fault
with the target of the harmful behavior. For example, torture may
be blamed on terrorists by noting they have brought such outcomes
on themselves.
The ideas captured in Bandura’s moral disengagement frame-
work are not completely new to psychological research, organiza-
tional behavior research, or behavioral ethics research. Others have
identified individual cognitive mechanisms that serve to discon-
nect an act from its moral valence. For example, the tendency of
individuals to displace responsibility for their actions onto pow-
erful others (e.g., Diener, 1977) or to diffuse responsibility by
pointing to collective decision making (e.g., Kelman, 1973) is well
documented. Likewise, much has been written about the processes
leading to moral exclusion and subsequent dehumanization (Opo-
tow, 1990; Staub, 1989) as well as the terrible consequences that
can flow from these processes (e.g., Kelman, 1973). More re-
cently, scholars have cited Bandura’s (1999) work when arguing
that some of these same cognitive processes can foster unethical
action in organizations. For example, Anand, Ashforth, and Joshi
(2005) linked selected rationalization tactics (denial of responsi-
bility, denial of injury, denial of the victim) and euphemistic
language to the facilitation of corruption. Similarly, Tenbrunsel
and Messick (2004) identified euphemistic language as a key
self-deceptive tactic that allows individuals to behave unethically
in organizations.
We view Bandura’s (1986) unique contribution to be the pro-
vision of a coherent theory that ties together all of these cognitive
mechanisms or tactics by explaining how they all serve to deacti-
vate moral self-regulatory processes. With colleagues, he has pre-
sented initial evidence that these cognitive processes operate as
part of a single overarching moral disengagement construct that
can be linked to outcomes such as childhood aggression or delin-
quency (Bandura et al., 1996, 2001).
Individual Differences and Moral Disengagement
If moral disengagement is an important precursor of unethical
decision making, it seems essential to understand whether some
people are more prone to moral disengagement than others. We
therefore identify a set of individual differences that have been
previously linked to moral cognition and action (e.g., Andersson &
Bateman, 1997; Aquino & Reed, 2002; Miller & Eisenberg, 1988;
Trevin˜o & Youngblood, 1990) and link them in Hypotheses 1– 4 to
the moral self-regulation process via their proposed influences on
moral disengagement. We will argue that these four individual
differences predispose individuals to see others, events, and them-
selves in ways that should make moral disengagement more or less
likely.
Empathy. Psychologists who study moral cognition and action
have highlighted the importance of imagining oneself in another’s
place or taking the perspective of others (see Eisenberg, 1986;
Kohlberg, 1969; Rest, 1986). Empathy is an individual difference
that describes the degree to which an individual notices and is
concerned about the needs or concerns of others (see Eisenberg &
Miller, 1987; Batson et al., 1989; Miller & Eisenberg, 1988). The
affective approach to understanding empathy emphasizes the ob-
server’s feeling of the target’s emotions, while the cognitive ap-
proach focuses on recognizing and understanding another’s
thoughts and feelings—that is, cognitively “taking the place” of
another or putting oneself in another’s shoes. Kohlberg (1969)
used the term role-taking to represent this cognitive process and
argued that role-taking was essential to moral development and
moral judgment. Rest (1986) similarly acknowledged the impor-
tance of role-taking to moral judgment, but also took a more
affective approach, suggesting that gut-level empathic feelings
often occur prior to moral judgment, increasing sensitivity to the
moral nature of the situation. Both approaches emphasize that
those high in empathy are more likely to take into account the
other person’s concerns. According to Bok (1998) “Empathy and
fellow feeling form the very basis of morality . . . Without some
rudimentary perception of the needs and feelings of others, there
can be no beginnings of felt responsibility toward them” (p. 70).
The dispositional view of empathy suggests that some individ-
uals are more predisposed than others toward vicarious empathic
experience and are more likely to engage in personalization and
“imaginative self-involvement” (Bandura, 1986, p. 314). Higher
levels of such arousal motivate the helping of others in need and
reduce motivation to harm others. Although social psychologists
have argued about whether such a predisposition toward empathy
exists (e.g., Batson, 1991), research has demonstrated support for
the idea that individuals differ in their concern for others and that
empathy emerges in childhood and is quite stable over time (Eisen-
berg et al., 1999).
We propose that being more acutely aware of the needs and
feelings of others should inhibit moral disengagement. Disposi-
tional empathy should be negatively related to moral disengage-
ment because individuals high on empathy are more likely to
vicariously experience the feelings of others and to be concerned
about those others’ needs. As a result, high empathy individuals
should be less likely to morally disengage through processes such
as the moral justification of acts that would harm others or the
dehumanization of the targets of those acts. They should also be
less likely to distort the consequences (i.e., potential harm) of their
actions or attribute blame to victims.
Hypothesis 1: Dispositional empathy is negatively related to
moral disengagement.
376 DETERT, KLEBE TREVIN
˜O, AND SWEITZER
Trait cynicism. Through early experience, individuals develop
philosophies of human nature that are difficult to change (Wrights-
man, 1992). One such philosophy of human nature is trait cyni-
cism, defined as a general attitude characterized by feelings of
frustration and disillusionment as well as distrust of other persons,
groups, ideologies, social conventions, and institutions (Abraham,
2000; Costa, Zonderman, McCrae, & Williams, 1985; Hochwarter,
James, Johnson, & Ferris, 2004). We propose that trait cynicism
will facilitate moral disengagement because individuals who are
high on trait cynicism have an underlying distrust of other people.
Thus, an individual who is high on trait cynicism will be more
likely to question the motives of others, including targets of harm,
and will be more likely to think that such targets are deserving of
their fate. Trait cynics should also be more likely to diffuse
responsibility because they think everyone is engaged in selfish
acts; in addition, they should be more likely to displace responsi-
bility to others, especially leaders, because they see such others as
lacking integrity or altruism (Kanter & Mirvis, 1989). Therefore,
we predict that those higher in trait cynicism will be predisposed
toward the types of morally disengaged reasoning that Bandura
argued are central to deactivation of moral self-regulation.
Hypothesis 2: Trait cynicism is positively related to moral
disengagement.
Locus of control. Locus of control orientations relate to how
individuals think about the events in their lives. These orientations
are relatively stable dispositions that differentiate between people
who believe they have personal control over the outcomes in their
lives and those who believe that such outcomes are controlled by
chance or powerful others (Lefcourt, 1966; Rotter, 1966). Individ-
uals with strong internal locus of control orientations see clear
connections between their own behavior and the outcomes of that
behavior (Levenson, 1981; Rotter, 1966). Trevin˜o (1986) theoret-
ically linked internal locus of control with ethical decision making,
arguing that those who see a clear connection between their own
behavior and its outcomes would be more likely to take personal
responsibility for that behavior. Trevin˜o and Youngblood (1990)
found empirical support for this link. This responsibility-taking, in
turn, activates moral norms (Schwartz, 1977). Thus, we propose
that individuals with higher internal locus of control orientations
should be less likely to morally disengage by displacing or diffus-
ing responsibility for unethical actions because they look within
rather than to others or to normative conditions for explanation or
justification of the relationship between their actions and conse-
quences (Maqsud, 1980). Internal locus of control should also be
negatively associated with moral disengagement because those
with stronger internal orientations are more likely to consider the
consequences of their actions and are less likely to attribute blame
for wrongdoing to others.
Hypothesis 3a: Internal locus of control is negatively related
to moral disengagement.
While Rotter (1966) treated locus of control as a single contin-
uum with internal locus of control at one end and external locus of
control at the other, Levenson (1974, 1981) asserted and found
evidence that (a) internal and external orientations represent dis-
tinct dimensions and (b) there are two conceptually and empiri-
cally distinct external orientations— one in which life’s outcomes
are largely attributed to chance, that is, seen as randomly deter-
mined by fate or luck, and another wherein outcomes are seen as
largely under the control of powerful others.
The chance external locus of control dimension describes how
much an individual believes that life experiences and outcomes are
a result of fate or luck rather than personal initiative (Levenson,
1981). We propose that individuals with higher chance locus of
control orientations are more prone to moral disengagement in part
because they see responsibility for outcomes as coming from
outside the self (Borrero-Hernandez, 1979). High chance locus of
control individuals may also be more likely to disregard or distort
consequences because they are more likely to think that an out-
come could not be helped.
Hypothesis 3b: Chance locus of control is positively related to
moral disengagement.
The powerful others locus of control dimension refers to the
belief that, although the world is relatively predictable, powerful
others are in control of events (Levenson, 1981). Having a higher
powerful others locus of control orientation should facilitate moral
disengagement because individuals who are high on this locus of
control dimension should be more likely to displace responsibility
for their own actions onto authority figures. For example, when
powerful others in an organization decide to change the rules
arbitrarily, subordinates with higher powerful others locus of con-
trol orientations may be particularly likely to go along with such
rules even if the changes have harmful effects on various stake-
holders because they reason that they have little personal control in
such situations. Individuals higher on the power locus of control
dimension may also be more likely to use other moral disengage-
ment tactics, such as moral justification, because they are more
likely to merely parrot what they hear from authority figures rather
than question those in charge.
Hypothesis 3c: Power locus of control is positively related to
moral disengagement.
Moral identity. Moral identity concerns how individuals think
about themselves. Moral identity has been defined in individual
difference terms as a relatively stable “self conception organized
around specific moral traits” (Aquino & Reed, 2002, p. 1,424) and
one of a number of hierarchically organized identities that com-
pose the sense of self. Because individuals have multiple identities
(Markus & Kunda, 1986), those identities that are the most salient
to the self are expected to most strongly influence thoughts and
feelings. For individuals with a highly self-important moral iden-
tity, moral concerns and commitments are central to their self-
definition and self-concept (Aquino & Reed, 2002). As a result,
Aquino and colleagues (Aquino et al., 2007) proposed that such
individuals are more likely to be concerned about the suffering of
others, including out-group members. We therefore hypothesize
that a highly self-important moral identity should inhibit moral
disengagement processes such as those that minimize or miscon-
strue harm to others (distortion of consequences) or dehumanize or
blame victims of harm (dehumanization, attribution of blame).
Thus, we predict that individuals with highly self-important moral
identities will be lower in moral disengagement.
377
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
Hypothesis 4: Moral identity is negatively related to moral
disengagement.
Moral Disengagement and Unethical Decision Making
Moral disengagement has been proposed to increase unethical
behavior because morally disengaged reasoning disconnects a con-
templated act from the guilt or self-censure that would otherwise
prevent it. This break between internal standards and contemplated
behavior reduces self-deterrents that would normally block indi-
viduals from carrying out unethical action (Bandura et al., 1996;
Duffy, Aquino, Tepper, Reed, & O’Leary-Kelly, 2005). Empirical
evidence supports this theoretical relationship. For example, Ban-
dura and colleagues found that moral disengagement decreased
prosocial behavior (helpfulness, cooperativeness) and increased
antisocial behavior (aggression, delinquency) in children (Bandura
et al., 1996, 2001; Bandura et al., 1975). With adults, the primary
application of moral disengagement has been in the area of atti-
tudes toward war and terrorism. McAlister (2001), for example,
found moral disengagement to be positively related to support for
military attacks against Iraq and Yugoslavia. Similarly, Aquino
and colleagues (Aquino et al., 2007) found moral disengagement
mechanisms to be positively related to the choice of death rather
than nonlethal options for the perpetrators of the 9/11 attacks. To
our knowledge, Duffy and colleagues (2005) have conducted the
only study on the relationship between moral disengagement and
adult unethical decision making or behavior. They found that
moral justification (a single moral disengagement mechanism) was
positively related to subsequent co-worker undermining (e.g.,
spreading rumors) among hospital workers. We propose here that
moral disengagement will have this same positive relationship
with undesirable outcomes in the realm of everyday decision
making wherein individuals make decisions that involve cheating,
lying, stealing, and other unethical behaviors.
Hypothesis 5: Moral disengagement is positively associated
with unethical decision making.
Moral Disengagement as a Mediator
Because some of the individual differences we hypothesized as
influences on moral disengagement (Hypotheses 1– 4) have also
been theorized and/or found to directly influence counterproduc-
tive or unethical behaviors (e.g., Andersson & Bateman, 1997;
Miller & Eisenberg, 1988; Trevin˜o & Youngblood, 1990), we
theorize that moral disengagement may play a mediating role
connecting stable individual differences to unethical decision mak-
ing.
Empathy, moral disengagement, and unethical decision making.
Psychologists have argued that empathy should inhibit aggressive
or antisocial conduct toward others because individuals who ex-
perience others’ negative feelings should be less inclined to con-
tinue harmful behavior or engage in it in the future (e.g., Eisen-
berg, 1986). Miller and Eisenberg’s (1988) meta-analysis revealed
that empathy was negatively related to antisocial actions such as
verbal and physical aggression in childhood, aggressive criminal
offenses, and administration of shock in a learning task. However,
the authors also noted that the processes underlying the link
between empathy and antisocial behavior are not well understood.
We propose that the relationship between empathy and unethical
decision making can be explained, in part, through moral disen-
gagement processes. Individuals higher in empathy are more likely
to be emotionally aroused by the needs of others and are more
likely to cognitively put themselves in others’ shoes. Thus, as
argued above, those higher in empathy should be less likely to
morally disengage via processes such as dehumanization or victim
blaming. This, in turn, should lead to a lower likelihood of making
decisions that would harm others.
Hypothesis 6a: Moral disengagement mediates the relation-
ship between empathy and unethical decision making.
Trait cynicism, moral disengagement, and unethical decision
making. Consistent with arguments that people who are distrust-
ful of others are more likely to engage in unethical behavior
(Rotter, 1980), research on cynicism in the workplace supports the
notion that cynicism increases unethical behavior. For example,
Andersson and Bateman (1997) found that trait cynics were sig-
nificantly more likely to approve an advertising plan that made
false claims. We propose that individuals high in trait cynicism
will be more likely to make unethical decisions that harm others in
part because of their moral disengagement. Such individuals are
less likely to trust others, and they are more likely to blame
potential victims and see them as deserving of bad outcomes. This,
in turn, should make unethical decisions more likely.
Hypothesis 6b: Moral disengagement mediates the relation-
ship between trait cynicism and unethical decision making.
Locus of control, moral disengagement, and unethical decision
making. Several studies have found locus of control orientations
to be related to ethical/unethical decision behavior. Trevin˜o and
Youngblood (1990), for example, showed that those with higher
internal locus of control orientations made more ethical decisions.
Likewise, Reiss and Mitra (1998) found that those higher on the
internal locus of control orientation were more likely to label a
variety of ethically ambiguous organizational actions as unaccept-
able. Research has also found that those with higher external locus
of control orientations behave unethically in experimental scenar-
ios (Hegarty & Sims, 1978), are willing to engage in insider
trading (Terpstra, Reyes, & Bokor, 1991), and are less likely to
report ethical beliefs that conflict with perceived organizational
interests (Mudrack & Mason, 1996).
Here, we hypothesize that the direct relationships between locus
of control orientations and unethical decision making will be
mediated by moral disengagement. For example, those with higher
internal locus of control orientations should be less likely to
morally disengage through processes such as diffusion or displace-
ment of responsibility. On the other hand, individuals with higher
chance locus of control orientations may make more unethical
decisions because they blame outcomes on fate or luck rather than
recognize their own role in producing harmful outcomes. In addi-
tion, individuals with higher power locus of control orientations
may make more unethical decisions because it is easy for them to
displace responsibility to powerful others. Thus, moral disengage-
ment processes should help to explain why these three locus of
control orientations make unethical decision making more or less
likely.
378 DETERT, KLEBE TREVIN
˜O, AND SWEITZER
Hypotheses 6c– 6e: Moral disengagement mediates the rela-
tionships between locus of control orientations (internal,
chance, and power are Hypotheses 6c, 6d, and 6e, respec-
tively) and unethical decision making.
Moral identity, moral disengagement, and unethical decision
making. When moral identity is highly self-important, moral
commitments are central to an individual’s self-concept. Such
individuals are expected to behave ethically in order to be true to
their moral conception of themselves (Aquino & Reed, 2002). In a
predictive validity study, Aquino and Reed found that high self-
importance of moral identity was positively associated with proso-
cial behaviors such as food donations and volunteering to help
others. Taking up Aquino and Reed’s call for future study of the
relationship between moral identity and antisocial behaviors, we
propose here that moral identity should reduce unethical decision
making. Further, we argue that this relationship can be explained,
in part, by the reduced use of moral disengagement mechanisms by
those whose moral identity is of higher self-importance. An indi-
vidual whose morality is more central to the self should be more
likely to activate moral cognitions (such as assessing the potential
for harm), less prone to morally disengaged reasoning, and, there-
fore, less likely to make unethical decisions.
Hypothesis 6f: Moral disengagement mediates the relation-
ship between moral identity and unethical decision making.
In sum, we hypothesized that a set of individual differences will
predict moral disengagement (Hypotheses 1– 4) and that moral
disengagement, in turn, will predict unethical decision making
(Hypothesis 5). We also hypothesized that moral disengagement
will play a mediating role between the hypothesized individual
difference antecedents and unethical decision making (Hypotheses
6a– 6f). These hypotheses are summarized visually in Figure 1.
Method
Research Design and Procedures
Testing the hypotheses outlined in Figure 1 required extensive
cooperation from individual respondents and a sponsoring institu-
tion. We therefore conducted a multi-wave survey study of busi-
ness and education undergraduates in the business and education
colleges of a large public research university in the Northeast. The
institution enrolls approximately 40,000 students with an entering
freshman class of approximately 6,500 students (of which approx-
imately 1,050 say they intend to major in business or education).
To survey students multiple times in a short time frame, we
obtained access to the students via 40 sections of a first year
seminar offered in both the business and education colleges. All
surveys were administered by one of the authors or a trained
graduate assistant. Students were told by the administrator prior to
Survey 1 that it was the first of three surveys they would be asked
to complete in class as part of a research study on decision making.
Students were informed that their participation was voluntary and
that all parts of the study had been reviewed and approved by the
University’s institutional review board. They were also informed
that their student ID was being requested on all surveys so that data
could be matched across the multiple instruments, and they were
promised that the data would be treated confidentially and would
bear no relation to their performance in the course.
The choice of a university sample was useful for testing this
study’s hypotheses. First, our goal was to explore individual dif-
ferences and psychological mechanisms underlying ethics-related
decision making. By college age, dispositions are presumably well
established. Conversely, while many individuals of college age
have some work experience, few have been fully socialized into
the world of work in ways that could be considered major influ-
ences on their attitudes or behavior. Further, divorcing the study
from an employment context minimized the likelihood that respon-
dents would distort their answers out of concern for potential
career consequences. A university context was also well suited to
aspects of our research design. For example, we were able to (a)
collect data on multiple constructs, (b) collect data from the same
individuals on several occasions, thereby minimizing concern for
response bias in self-report data (Ostroff, Kinicki, & Clark, 2002),
and (c) obtain behavioral data to help validate our measure of
unethical decisions.
Survey 1 was designed primarily to collect the individual dif-
ferences data to be used as control and independent variables in
our hypothesis testing. It was administered during the first 2 weeks
of the semester in small sections of the orientation seminar re-
quired for all freshmen. Survey 2 was designed primarily to
measure moral disengagement and was administered in class sev-
eral weeks later during the same seminar to the same sample of
business and education students. Survey 3, which contained the
items for the Unethical Decision Making scale, was administered
several weeks after Survey 2 during the first week of a new
semester. The ordering and temporal spacing of the three surveys
was designed to evaluate mediation and to minimize cognitive
carryover of prior questions and responses, a major source of
concern about common source variance (Harrison & McLaughlin,
1993). None of the respondents had completed the business
school’s required course in business ethics (a possible source of
contamination) at the time of Surveys 1–3. A final behavioral
measure used to provide validation evidence for the Unethical
Decision Making measure was obtained 10 weeks after Survey 3.
Research Subjects
Because our data collection occurred across three survey admin-
istrations spanning two semesters, the final sample for these anal-
yses came from the 307 respondents enrolled in colleges of busi-
ness and education who completed and provided their correct ID
number on all three surveys. This is approximately 29% of all
freshmen who plan to major in business or education at the
university. As noted in Table 1, respondents in the final sample
were primarily raised in the United States (96%), had an average
SAT score of 1214, and 54% were female students while 46%
were male students. Respondents were on average 18.4 years of
age; were 87% Caucasian American, 5% Asian American, 3%
African American, 2% Hispanic American, and 3% non-
American; and had an average of 19.3 months of paid full- or
part-time work experience in 2.88 different jobs (3% reported no
work experience). The population of business and education fresh-
men at this university was similar on these dimensions, having
only a slightly lower SAT average (M⫽1166). The 307 respon-
dents who completed all surveys were also similar to the more than
379
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
500 others who completed (with correct ID) one or two but not all
surveys on the study’s focal variables (e.g., trait cynicism: M⫽
4.80 for all surveys vs. 4.88 for some surveys; moral disengage-
ment: M⫽2.11 for all surveys vs. 2.24 for some surveys).
Independent Variables
Empathy. Empathy was measured in Survey 1 by using a
10-item scale from the International Personality Item Pool (IPIP;
Goldberg, 2002). Labeled as sympathy on the IPIP, the items
represent a general disposition of empathy. According to Bierhoff
(2002), with few exceptions (Eisenberg, 2000) the term sympathy
is an older term that is rarely used in current psychological dis-
cussions of empathic responses. The items on the IPIP scale aim to
tap an individual’s willingness to take others’ problems and emo-
tions into consideration. Sample items illustrating the affective and
cognitive aspects of empathy are “I suffer from others’ sorrows”
and “I am not interested in other people’s problems” (reverse
scored). The 10 items (␣⫽.81) were assessed by using a 7-point
Likert scale ranging from 1 (strongly disagree)to7(strongly
agree).
Trait cynicism. Trait cynicism was measured in Survey 1 by
using a five-item scale adapted by Johnson and O’Leary-Kelly
(2003) from Wrightsman’s (1992) subscale for cynicism as a
Philosophy of Human Nature (PHN). The items assessing cyni-
cism stem from Wrightsman’s (1992) assumption that cynicism
reflects a “virulent assumption about human nature,” namely, an
individual’s “attributions of selfishness and fakery to others” (p.
5). Therefore, the items assess the belief that people will refrain
from lying, cheating, or stealing only when it is easy to do so or
when they think they would get caught. The items were assessed
with a 7-point Likert scale ranging from 1 (strongly disagree)to7
(strongly agree). Cronbach’s alpha was .79.
Locus of control: internal, chance, and power. To measure the
three locus of control orientations, we used Levenson’s (1981)
scales for internal, chance, and power locus of control in Survey 1.
The eight-item Internal Locus of Control scale measures the extent
to which individuals believe that they have control over events in
their own lives (e.g., items assess the belief that those with more
power determine whether one experiences personal or professional
success). The eight-item Chance Locus of Control scale measures
one’s perceptions of how probable it is that events occur because
of fate or chance (e.g., items assess the belief that luck or fate
determine whether one experiences personal, professional, or
physical well-being). The eight-item Power Locus of Control scale
measures the extent to which individuals believe that their life
outcomes are largely determined or controlled by powerful others
(e.g., items assess the belief that one can make one’s plans happen
or protect one’s personal interests). All locus of control items were
assessed by using a 6-point Likert scale ranging from 1 (strongly
disagree)to6(strongly agree). Cronbach’s alpha for the Internal,
Chance, and Power Locus of Control scales were .65, .78, and .77,
respectively.
Moral identity. Moral identity was measured by using
Aquino and Reed’s (2002) Internalization subscale in Survey 2.
On this measure, subjects are first presented a set of nine
adjectives (e.g., caring, compassionate, fair, honest) along with
the statement that these represent “some characteristics that
might describe a person.” Subjects then rate, with a scale
ranging from 1 (strongly disagree)to5(strongly agree), five
items intended to assess the degree to which these characteris-
tics represent an important part of their own identity (sample
item: “Being someone who has these characteristics is an im-
portant part of who I am”; scale ␣⫽.77). The Internalization
subscale captures the degree to which a person’s moral identity
is rooted at the core of one’s being and has been found to be the
more robust predictor (compared with the Symbolization sub-
scale) of ethics-related attitudes and behavior (Aquino & Reed,
2002; Aquino et al., 2007).
Table 1
Means, Standard Deviations and Intercorrelations of Study Variables
Variable MSD123456789101112
1. Gender
a
.46 .50 —
2. Business vs. education
b
.70 .46 .41
***
—
3. SAT total 1,213.52 120.46 .23
***
.19
**
—
4. Native country
c
.96 .19 ⫺.02 ⫺.13
*
⫺.06 —
5. Empathy
d
3.86 .77 ⫺.34
***
⫺.26
***
⫺.09 ⫺.03 —
6. Trait cynicism
d
4.80 1.02 .15
*
.11 ⫺.12
*
⫺.03 ⫺.06 —
7. Internal locus of
control
e
4.28 .59 .15
*
.21
***
.11 ⫺.05 ⫺.21
***
.05 —
8. Chance locus of
control
e
2.71 .75 ⫺.02 ⫺.12 ⫺.02 ⫺.01 .07 .19
**
⫺.26
***
—
9. Power locus of control
e
2.93 .73 .11 .19
*
.04 ⫺.04 ⫺.11 .18
**
⫺.00 .47
***
—
10. Moral identity
f
4.58 .46 ⫺.16
**
.04 ⫺.01 .01 .22
***
.01 .00 ⫺.04 ⫺.08 —
11. Moral disengagement
f
2.11 .41 .30
***
.02 .07 ⫺.10 ⫺.27
***
.30
***
.04 .20
***
.14
*
⫺.24
***
—
12. Unethical decisions
d
3.30 1.00 .17
**
.11 ⫺.04 .04 ⫺.21
***
.31
***
.08 .11
*
.15
*
⫺.11 .34
***
—
a
Dichotomous variable for gender: male ⫽1; female ⫽0.
b
Dichotomous variable for intended major: business ⫽1; education ⫽0.
c
Dichotomous variable for country raised: United States ⫽1; outside United States ⫽0.
d
Measure was assessed using a 7-point scale.
e
Measure was assessed using a 6-point scale.
f
Measure was assessed using a 5-point scale.
*
p⬍.05.
**
p⬍.01.
***
p⬍.001.
380 DETERT, KLEBE TREVIN
˜O, AND SWEITZER
Dependent Variables
Moral disengagement. Moral disengagement was assessed on
the second survey with a measure similar to the one developed and
used in multiple studies by Bandura and others (Bandura et al.,
1996, 2001; Pelton et al., 2004). However, because Bandura’s
32-item scale was developed for use with children and young
adolescents, we adapted it to fit the population of our study (e.g.,
“It is unfair to blame a child who had only a small part in the harm
caused by the group” was changed to “You can’t blame a person
who plays only a small part in the harm caused by a group”; “If
kids fight and misbehave in school it is their teacher’s fault” was
eliminated and replaced with “People are not at fault for misbe-
having at work if their managers mistreat them”). Similar to
Bandura’s measure, the items were designed to equally tap the
eight subcomponents of the overarching moral disengagement
construct (see Appendix A for all items and the subcomponents
they are designed to tap). The items were assessed on a 5-point
Likert scale ranging from 1 (strongly disagree)to5(strongly
agree).
We randomly split the 828 responses to the moral disengage-
ment items in half and conducted exploratory factor analyses
(maximum-likelihood estimation with direct oblimin rotation) by
using 424 responses. Consistent with the theory, statistical criteria
(e.g., scree test examination, eigenvalues) suggested that an eight
factor solution was most appropriate. The pattern of item loadings
for the eight factor model, however, suggested that several items
should not be retained because they did not load on their expected
factor at ⬎.30 and/or had cross-loadings ⬎.30 (Fabrigar, Wegener,
MacCallum, & Strahan, 1999). To keep the Moral Disengagement
scale balanced across its eight components and as parsimonious as
possible, we selected for retention the 3 best-fitting items for each
moral disengagement subcomponent. The 24 items selected (see
Appendix A) all loaded on their expected factor at greater than .40
and had no cross-loadings greater than .25. We next conducted
confirmatory factor analysis with Lisrel 8.53 on the hold-out
sample (n⫽424), estimating a model with the 24 items loading
onto eight first order factors and those factors loading onto a single
second order factor. The fit statistics for this maximum-likelihood
estimated model indicate that the model fits the data well (e.g.,
root-mean-square error of approximation [RMSEA] ⫽.05; non-
normed fit index [NNFI] ⫽.95; comparative fit index [CFI] ⫽.96;
root-mean-square residual [RMR] ⫽.06), suggesting the validity
of Bandura and colleagues’ (Bandura et al., 1996) claim that moral
disengagement should be measured as a single higher order con-
cept. We therefore created an overall measure (␣⫽.87) of moral
disengagement by averaging the responses to the 24 items.
Unethical decision making. We created a measure of unethical
decision making by averaging responses to eight ethically charged
scenarios presented to subjects in Survey 3. Five scenarios were
created specifically for this study; three were adapted from Seg-
lin’s Ethics Survey (n.d.). A sample scenario is “You work as an
office assistant for a department at [University X]. You’re alone in
the office making copies and realize you’re out of copy paper at
home. You therefore slip a ream of paper into your backpack.” For
this and the other seven scenarios (see Appendix B, Scenarios
1– 8), respondents were asked to read the scenario and evaluate
“How likely is it that you would engage in the behavior de-
scribed?” using a 7-point scale ranging from 0 (not at all likely)to
6(highly likely). Just prior to the first scenario, students were
reminded that “there are no ‘incorrect’ answers” and that the
survey would have value to themselves and others only if they
gave “truthful responses, not those [they] think might seem more
desirable.” Respondents were also assured (on all three surveys)
that their responses would be confidential and not used to evaluate
them personally in any way. Cronbach’s alpha for this scale was
.74.
We conducted three independent exercises to provide evidence
regarding the construct validity of the Unethical Decision Making
measure. Before administering the measure, we contacted 20 busi-
ness ethics experts via e-mail and asked them to provide ratings on
a 5-point Likert scale on whether the behavior in each of the eight
scenarios composing the Unethical Decisions Making scale repre-
sents a violation of one or more ethical principles (e.g., virtue,
rights, justice). The experts were also asked to rate five randomly
interspersed scenarios written by the authors to not represent
unethical behavior. Sixteen experts responded. Findings indicated
that the panel of ethics experts found the eight scenarios used in
the scale to clearly represent violations of ethical principles: The
grand mean for the items used in our Unethical Decision Making
scale was 4.66 (SD ⫽.35). Further, this mean was significantly
different, t(15) ⫽15.3, p⬍.001, than the grand mean for the five
other scenarios (M⫽1.68, SD ⫽.62) rated by the experts but not
included in our scale. Appendix B provides additional details on
the expert rating exercise.
To demonstrate that our Unethical Decision Making scale is
correlated with actual unethical behaviors, we conducted two
additional validation exercises. Both are described in full in Ap-
pendix B. In one of these exercises, we obtained behavioral data
from a sample of 59 subjects that we then correlated with the
closest individual item on our Unethical Decision Making scale.
This exercise presented subjects with an actual opportunity to keep
or return $8 received in the mail that did not belong to them, a
choice somewhat similar to the scenario from the Unethical De-
cision Making scale that involved keeping or returning $10 in extra
change from a coffee shop. The polychoric correlation coefficient
of .34 between subjects’ actual behavior (keeping the cash) and
their response to the related scenario was significant at p⬍.01.
In the second exercise, we conducted two surveys of business
students at the same university. Students first provided responses
to our eight ethical decision making scenarios (␣⫽.80). Two
weeks later, these same students were asked to indicate how often
they had engaged in each of 13 cheating, lying, or stealing behav-
iors (e.g., “taking small amounts of money from my parent’s wallet
without their permission,” “copying from another student on a
test”; ␣⫽.80). We were able to match data for 58 respondents. At
.61 ( p⬍.001), the correlation between these scales indicates that
those who have engaged in an array of unethical behaviors are
significantly more likely to say that they intend to behave uneth-
ically as measured by our Unethical Decision Making scale.
Control Variables
To account for variance in the dependent variables (moral
disengagement for tests of Hypotheses 1– 4 and unethical decisions
for Hypotheses 5– 6f) that might be explained by factors other than
the hypothesized variables, we first entered four control variables
into each of our regression models. We coded male subjects as 1
381
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
and female subjects as 0 on a gender dummy variable. Although
often perceived to be different in their ethics (Schminke, Ambrose,
& Miles, 2003), small or nonsignificant differences have been
found between male subjects and female subjects on ethical rea-
soning and outcomes. When differences were found, female sub-
jects demonstrated slightly higher moral reasoning or more ethical
action (Ambrose & Schminke, 1999; Gephart, Harrison, &
Trevin˜o, 2007; O’Fallon & Butterfield, 2005; Rest, Thoma, Moon,
& Getz, 1986). Second, students with an intended major in busi-
ness were assigned a 1 and educationa0onanintended major
dummy variable because students with different intended majors
may systematically differ in their propensity to morally disengage
or behave unethically due to other unmeasured individual charac-
teristics (McCabe & Trevin˜o, 1995). We also entered each sub-
ject’s total SAT score (obtained with permission from university
records) into each regression equation because moral reasoning is
a cognitive task and has been associated with mental ability (Rest,
1986). Finally, we used a native country dummy variable indicat-
ing whether respondents had spent the majority of their first 18
years in the U.S. (1) or elsewhere (0) as a final control variable in
all analyses because the native country of individuals may influ-
ence their ethical considerations. For example, national K–12
educational approaches may have systematic effects on ethical
awareness in young adults; alternatively, national culture differ-
ences may alter individuals’ perceptions of right and wrong in
particular situations (McAlister, 2001).
Results
To assess construct independence among the eight independent
and dependent variables, we conducted confirmatory factor anal-
ysis on the items measuring empathy, cynicism, the three locus of
control orientations, and moral identity, moral disengagement, and
unethical decision making. The hypothesized eight factor structure
for the 76 items fit the data well (RMSEA ⫽.04, CFI ⫽.93,
NNFI ⫽.92, RMR ⫽.06). We also fit a number of other
models, including (a) a seven factor model that assessed how
the data fit a model with the moral disengagement and unethical
decision making items as a single factor and (b) a six factor
model that loaded moral disengagement and unethical decision
making onto a single factor as well as both externally oriented
locus of control dimensions (power and chance) onto a different
distinct factor. The intent of these alternative models was to
rule out more parsimonious models as providing equally good
or better fit to the data. For each model with fewer than eight
factors, all fit statistics suggested a significantly worse fit to the
data. Additionally, all chi-square differences per degrees of
freedom between the eight factor model and the models with
fewer factors were significant at p⬍.001. Thus, we found
strong support for our hypothesized eight factor structure for
this study’s independent and dependent variables.
Table 1 presents the means, standard deviations, and intercor-
relations of the study variables. The bivariate correlations between
the individual difference variables, moral disengagement, and un-
ethical decision making are nearly all in the predicted direction and
most are statistically significant. The mean (2.11) and standard
deviation (.41) for moral disengagement, assessed on a 5-point
scale, suggests most respondents are not highly prone to moral
disengagement, while the mean (M⫽3.30, which is slightly above
the midpoint of the 0 – 6 scale) and standard deviation (1.0) for
unethical decision making suggests reasonably high variability in
subjects’ willingness to say that they would engage in unethical
actions.
Ordinary least squares regression analyses were used to test the
hypotheses. Results are shown in Table 2. We first tested the
hypotheses involving individual difference influences on moral
disengagement (Hypotheses 1– 4) and then tested the hypothesis
that moral disengagement will predict subsequent unethical deci-
sion making (Hypothesis 5). Finally, we tested our hypotheses
about moral disengagement as a mediator of the individual
differences– unethical decision making relationships (Hypotheses
6a– 6f).
To account for the variance in moral disengagement explained
by the control variables, we first entered only respondents’ gender,
intended major, total SAT score, and native country into a regres-
sion model. As shown in Model 1 of Table 2, 11% of the variance
in respondents’ moral disengagement was accounted for by the
control variables. Male subjects were more likely to be morally
disengaged than were female subjects (B ⫽.29; 95% confidence
interval [CI] ⫽.19, .39; p⬍.001), while those intending to major
in business (vs. education) were less likely to be morally disen-
gaged (B ⫽–.12; 95% CI ⫽–.23, –.02; p⬍.05). The other control
variables, SAT score and native country, were nonsignificant.
In Model 2, we entered the control variables, followed by the six
independent variables to test the empathy, trait cynicism, locus of
control (internal, chance, and power), and moral identity influ-
ences on moral disengagement (Hypotheses 1– 4). The inclusion of
these individual difference variables explained an additional 16%
of the variance in moral disengagement, F(6, 300) ⫽10.3, p⬍
.001. As predicted in Hypothesis 1, empathy (B ⫽–.10; 95% CI ⫽
–.15, –.04; p⬍.01) was negatively related to moral disengage-
ment, indicating that those more likely to empathize with others
are less likely to morally disengage. Conversely, trait cynicism
(B ⫽.10; 95% CI ⫽.06, .14; p⬍.001) was positively related to
moral disengagement, lending support to the hypothesis (Hypoth-
esis 2) that the higher in trait cynicism an individual is, the more
likely he or she is to morally disengage. However, only one of the
three locus of control orientations (chance) was significant.
Chance locus of control (B ⫽.09; 95% CI ⫽.02, .15; p⬍.01) was
positively related to moral disengagement, supporting the hypoth-
esis (Hypothesis 3b) that those who see life’s events as due to fate
or luck are more likely to morally disengage. Neither the internal
(B ⫽.01; 95% CI ⫽–.06, .09; ns) nor power (B ⫽.09; 95% CI ⫽
–.07, .06; ns) locus of control orientations were related to moral
disengagement. Thus, Hypotheses 3a and 3c were not supported.
Finally, moral identity (B ⫽–.13; 95% CI ⫽–.22, –.04; p⬍.01)
was negatively related to moral disengagement, providing support
for the hypothesis that moral identity reduces the propensity to
morally disengage (Hypothesis 4).
As shown in Model 3 (Table 2), we next tested the hypothesis
that moral disengagement would significantly predict unethical
decision making (Hypothesis 5). To account for the variance in
unethical decision making explained by alternative potential
causes of unethical decision making, we first estimated a model
with this study’s 4 control variables and the 6 independent
variables. Collectively these 10 variables explain 16% of the
variance in unethical decision making, but only empathy (B ⫽
–.19; 95% CI ⫽–.34, –.04; p⬍.05) and trait cynicism (B ⫽
382 DETERT, KLEBE TREVIN
˜O, AND SWEITZER
.26; 95% CI ⫽.16, .37; p⬍.001) were significantly related to
this dependent variable. In Model 4, we added moral disen-
gagement to the model as a predictor of unethical decision
making. As predicted (Hypothesis 5), moral disengagement
(B ⫽.56; 95% CI ⫽.27, .86; p⬍.001) was significantly and
positively related to unethical decision making. The inclusion
of moral disengagement explained an additional 4% of the
variance in unethical decision making, F(1, 305) ⫽14.1, p⬍
.001. Thus, the results provide support for our hypothesis
(Hypothesis 5) that individuals who are more morally disen-
gaged are more likely to make unethical decisions.
Hypotheses 6a– 6f proposed that moral disengagement will me-
diate relationships between the focal individual differences and
unethical decisions. The traditional method for assessing media-
tion over the past 2 decades has been the multi-step process
outlined by Baron and Kenny (1986). However, recent research
has suggested that mediation can be established without significant
direct relationships between independent and dependent variables
(the first step in the Baron–Kenny method; MacKinnon, Lock-
wood, Hoffman, West, & Sheets, 2002; Shrout & Bolger, 2002).
We therefore tested our mediation hypotheses with the Sobel test
(Sobel, 1982), a method for assessing indirect effects that does not
require significant main effects for the independent variables and
that provides better balance between Type I and Type II errors.
(MacKinnon et al., 2002).
We conducted the Sobel tests for moral disengagement as a
mediator of relationships between empathy, trait cynicism, chance
locus of control, and moral identity, excluding tests for the internal
and power locus of control orientations because they did not
directly affect moral disengagement (the mediator). The Sobel test
statistics are significant at p⬍.05 for all four of the tested
independent variables: empathy (z⫽–2.50, p⫽.012), trait cyn-
icism (z⫽2.91, p⫽.004), chance locus of control (z⫽2.16, p⫽
.031), and moral identity (z⫽–2.33, p⫽.020). Thus, we find
support for the hypotheses that moral disengagement plays a
mediating role between empathy (Hypothesis 6a), cynicism (Hy-
pothesis 6b), chance locus of control (Hypothesis 6d), and moral
identity (Hypothesis 6f) and the dependent variable— unethical
decisions. Conversely, results do not support the mediation hy-
potheses involving the internal (Hypothesis 6c) and power (Hy-
pothesis 6e) locus of control orientations.
Discussion
Our goal in this research was to better understand the anteced-
ents and outcomes of moral disengagement, a set of cognitive
mechanisms that are thought to deactivate moral self-regulation
and allow individuals to make unethical decisions more easily
(Bandura, 1999). We found that individual differences in empathy,
trait cynicism, chance locus of control, and moral identity predict
moral disengagement and that moral disengagement predicts un-
ethical decision making. Further, we found support for the idea
that moral disengagement mediates the relationships between these
individual differences and unethical decisions.
Contributions and Implications for Future Research
Our study contributes to knowledge about moral disengagement
on several fronts. First, we identified four types of individual
Table 2
Results of Ordinary Least Squares Regression Analysis: Individual Differences, Moral Disengagement, and Unethical Decisions
Variable
Model 1 Model 2 Model 3 Model 4
DV ⫽moral
disengagement
DV ⫽moral
disengagement
DV ⫽unethical decision
making
DV ⫽unethical
decision making
B 95% CI B 95% CI B 95% CI B 95% CI
Control variable
Gender
a
.29
***
.19, .39 .18
***
.09, .28 .13 ⫺.11, .38 .03 ⫺.22, .28
Business vs. education
b
⫺.12
*
⫺.23,⫺.02 ⫺.13
*
⫺.23,⫺.02 .05 ⫺.23, .32 .12 ⫺.15, .39
SAT total .00 .00, .00 .00 .00, .00 .00 ⫺.00, .00 .00 ⫺.00, .00
Native country
c
⫺.22 ⫺.45, .00 ⫺.21 ⫺.42, .00 .27 ⫺.28, .81 .38 ⫺.16, .92
Independent variable
Empathy ⫺.10
**
⫺.15, .04 ⫺.19
*
⫺.34,⫺.04 ⫺.13 ⫺.28, .02
Trait cynicism .10
***
.06, .14 .26
***
.16, .37 .21
***
.10, .32
Internal locus of control .01 ⫺.06, .09 .08 ⫺.11, .28 .08 ⫺.11, .26
Chance locus of control .09
**
.02, .15 .09 ⫺.08, .27 .05 ⫺.13, .22
Power locus of control ⫺.01 ⫺.07, .06 .05 ⫺.12, .22 .05 ⫺.12, .22
Moral identity ⫺.13
**
⫺.22,⫺.04 ⫺.15 ⫺.38, .09 ⫺.07 ⫺.31, .16
Mediator
Moral disengagement .56
***
.27, .86
R
2
.11
***
.27
***
.16
***
.20
***
Adjusted R
2
.10 .24 .13 .17
⌬R
2
.16
***
.04
***
df (regression, residual) (4, 302) (10, 296) (10, 296) (11, 295)
a
Dichotomous variable for gender: male ⫽1; female ⫽0.
b
Dichotomous variable for intended major: business ⫽1; education ⫽0.
c
Dichotomous variable for native country: United States ⫽1; outside United States ⫽0.
*
p⬍.05.
**
p⬍.01.
***
p⬍.001.
383
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
differences that had been previously associated with moral cogni-
tion and action and that could be theoretically linked with the
facilitation or inhibition of moral disengagement. We proposed
that these individual differences would predispose individuals to
see others, events, and themselves in ways that make moral dis-
engagement more or less likely. We found that four individual
differences predict moral disengagement, explaining 16% of the
variance in moral disengagement beyond control variables.
Consistent with our theorizing that empathy would make it
harder, for example, to discount or dismiss harm to others, we
found that those who are more empathic are less likely to morally
disengage. This finding is consistent with previous research and
theory that has emphasized the importance for ethical decision
making of seeing a situation from another’s perspective both
cognitively and affectively (Eisenberg, 2000; Kohlberg, 1969). We
also found that trait cynicism facilitates moral disengagement.
Rather than heightening sensitivity to the plight of others as
empathy does, such thoughts appear to be suppressed in trait
cynics. We theorized that this is because of trait cynics’ distrust of
others, which should allow them to more easily distance them-
selves from and diffuse responsibility to others as well as blame or
dehumanize victims.
We developed and tested another set of hypotheses proposing
that locus of control orientations would inhibit or facilitate moral
disengagement. We found that chance locus of control orientation
was positively related to moral disengagement, suggesting that
those who believe life experiences and outcomes are due to forces
outside their control are more likely to morally disengage (perhaps
by displacing or diffusing responsibility). However, our hypothe-
ses regarding internal locus of control and power locus of control
were not supported. The nonsignificant finding for internal locus
of control orientation may be attributable to the relatively low
reliability obtained for the measure in our sample. The nonsignif-
icant finding for the powerful others locus of control orientation
may be because our decision making scenarios did not generally
highlight authority relationships. Future research should continue
to examine how locus of control orientations relate to moral
disengagement.
Finally, we argued from an identity perspective that an individ-
ual who thinks of the self in terms of moral concerns and com-
mitments (Aquino & Reed, 2002) will be less likely to morally
disengage in ways that minimize or misconstrue harm to others.
Our findings support this hypothesis, namely that higher self-
importance of moral identity was negatively related to moral
disengagement.
In sum, our research is among the first to propose and find that
moral disengagement is facilitated or inhibited by a number of
individual differences that reflect the ways individuals think of
others, events, and themselves. Although our hypotheses were
based on solid theorizing, our study did not measure the proposed
cognitive mechanisms associated with these individual differences,
such as perspective taking, trust, or responsibility taking. Future
work should explicitly measure these mechanisms. For example,
researchers might employ Davis’s (1980) Interpersonal Reactivity
Index to more directly assess perspective taking.
Beyond dispositional influences, social cognitive theory (Ban-
dura, 1986) and interactionist theories of ethical decision making
behavior (Trevin˜o, 1986) suggest that social context antecedents
are also likely to be important predictors of moral disengagement.
For example, parenting style has been found to influence moral
disengagement in children (Pelton et al., 2004). Also, in the
organizational literature, researchers have theorized that an ethical
environment can reduce employee use of corruption-inducing psy-
chological tactics (Anand et al., 2005; Tenbrunsel & Messick,
2004). Thus, future research should investigate the possibility that
contextual factors (e.g., leadership style, ethical climate or culture)
have independent and interactive (with individual difference fac-
tors) influences on moral disengagement. Adding contextual fac-
tors to future studies is likely to increase the amount of variance in
moral disengagement that researchers are able to explain and
provide additional practical insights for those seeking to reduce
unethical decision making.
With regard to the consequences of moral disengagement, we
found higher levels of moral disengagement to be positively asso-
ciated with increased unethical decision making, providing addi-
tional support beyond research previously conducted on aggres-
sion in children (Bandura et al., 1996, 2001) and beyond research
that found a relationship between a single moral disengagement
mechanism (moral justification) and coworker undermining
(Duffy et al., 2005). Moral disengagement explained 4% of the
variance in unethical decision making beyond that explained by an
array of individual differences. This finding suggests that moral
disengagement can help us understand why unethical decision
making occurs. Because Bandura’s theory applies to transgressive
behavior more generally (Bandura, 1999), we speculate that our
findings on moral disengagement’s influence on unethical deci-
sions should extend to a wider variety of behaviors that violate
societal or organizational norms and are therefore of great interest
to managers—for example, antisocial, counterproductive, and de-
viant behaviors (Bennett & Robinson, 2003; Vardi & Weitz,
2003).
Given previous research suggesting direct effects on unethical
decisions for the individual differences we studied as antecedents
of moral disengagement, we also asked whether moral disengage-
ment might help to explain why these relationships exist. We
found that moral disengagement plays a mediating role between
unethical decision making and individual differences in empathy,
trait cynicism, chance locus of control, and moral identity. Thus,
one of the reasons less empathic people and those with a less
self-important moral identity are more likely to make unethical
decisions is that such individuals are more likely to morally
disengage. Similarly, distrustful cynics and those high on chance
locus of control may be more likely to make unethical decisions in
part because they disconnect from self-censure mechanisms by
morally disengaging.
Although our research contributed to knowledge about the an-
tecedents and outcomes of moral disengagement, much remains to
be understood about the temporal aspects of moral disengagement.
While Bandura argued that moral disengagement precedes uneth-
ical behavior (e.g., “People do not ordinarily engage in reprehen-
sible conduct until they have justified to themselves the rightness
of their actions”; Bandura, 1996, p. 335), and while our Moral
Disengagement measure did precede the Unethical Decision Mak-
ing measure, our study did not attempt to access the internal
cognitive processing that immediately precedes or follows specific
unethical decisions. Others have noted that processes similar to
moral disengagement can also be invoked retrospectively (Anand
et al., 2005). For example, individuals queried later about their
384 DETERT, KLEBE TREVIN
˜O, AND SWEITZER
unethical behavior may find themselves resorting to attribution of
blame, diffusion of responsibility, and similar tactics to justify a
prior unethical act. Future research should explore these cognitions
and their timing more directly. Research might also investigate
whether and how post hoc rationalizations for an act at Time 1 feed
pre-act moral disengagement in future periods. In other words, is
there a self-reinforcing cycle of moral disengagement?
Future research can also help to determine where moral disen-
gagement fits within the ethical decision making process and about
the level of consciousness involved. For example, Rest’s (1986)
commonly cited four-stage model asserts that ethical decision
making involves the sequence of moral sensitivity, judgment,
intention, and action. It is currently unclear whether moral disen-
gagement is associated with moral sensitivity, moral judgment, or
both. Moral sensitivity involves the initial recognition that the
individual is facing a decision with moral overtones. It seems quite
possible that moral disengagement mutes moral sensitivity by
reducing moral issue recognition. This would suggest that moral
disengagement occurs below consciousness—that individuals us-
ing moral disengagement would not be aware that they are using
such cognition to completely ignore the ethical nature of a situa-
tion. Alternatively, moral disengagement might instead be more
related to the next step, moral judgment, the process of deciding
what is right or wrong in a situation. If moral disengagement is
associated primarily with moral judgment, it would suggest that
moral disengagement operates at a more conscious level and
involves rationalizing away recognized moral concerns. Much
more research will be required to situate moral disengagement
within these stages of ethical decision making and to determine the
level of conscious thought involved.
Finally, researchers may also wish to explore the theoretical and
empirical differences between moral disengagement and moral
engagement. We speculate that the two may be separate constructs
rather than opposite ends of a single continuum. That is, an
individual who does not routinely morally disengage need not be
engaged in active reflection about ethical issues. For example,
managers may passively accept extreme pay differentials between
executives and average workers without using moral disengage-
ment to justify these differentials but also without actively engag-
ing the potential moral aspects of the pay gap issue. Some addi-
tional cognitive effort may be required in order for an individual to
morally engage. Here, research on the moral intensity of the ethical
issue (Jones, 1991) and moral awareness (e.g., Butterfield,
Trevin˜o, & Weaver, 2000; Flannery & May, 2000; May & Pauli,
2002) seems relevant. For example, moral intensity of an issue
depends largely on the extent of the harm involved. A situation that
is likely to produce serious harm should be more likely to get
attention and instantiate moral standards (Jones, 1991).
Strengths and Limitations of the Research
We conducted a carefully designed study in a professionally
oriented sample of young adults, in which we collected data over
multiple time periods in order to demonstrate mediation via the
ideal condition of having the independent, mediator, and depen-
dent variables collected sequentially over time. Nonetheless, in the
absence of an experimental design, judgments about the causal
ordering among our key constructs must be rendered according to
the plausibility of the theory and logic we have provided rather
than conclusive empirical evidence.
Our college student sample raises concerns about generalizabil-
ity (Gordon, Slade, & Schmitt, 1986). However, we believe our
sample was appropriate and useful given our primary focus on a
psychological process and our desire to collect data from individ-
uals on multiple constructs at multiple time periods. Prominent
researchers (e.g., Greenberg, 1987) have also argued that student
samples are not necessarily more problematic than many nonrep-
resentative adult samples and that student samples may be partic-
ularly useful for understanding psychological processes. Others
have demonstrated quite convincingly (e.g., Locke, 1986) that
research findings on psychological processes such as goal setting
conducted with student and adult samples are quite similar. None-
theless, we recognize that future studies using diverse samples are
needed to help establish generalizability.
When designing the study, we also considered the issue of social
desirability bias, always a concern in ethics-related research (Ran-
dall & Fernandes, 1991). As noted in the Method section, prior to
completing all surveys, respondents were told that there were no
incorrect answers and that the research would have value only if
they gave truthful responses, not those they might think seem more
desirable. They were also assured on each survey that their re-
sponses would be kept completely confidential and would not be
used to evaluate them in any way. We did not use a measure of
social desirability bias in our research because, having discussed
ethical issues with similar students in our classes for years, we
have found them to be surprisingly open about their willingness to
engage in unethical actions such as lying, cheating, and stealing.
When discussing their reasons with them, we noticed a pattern of
responses that indicated moral disengagement. This experience
does not rule out the possibility that social desirability bias influ-
enced our results. However, to the extent that it did, the effect is
likely to attenuate rather than inflate the significance of our find-
ings. Nevertheless, future researchers should consider including a
measure of social desirability bias in their studies, particularly with
samples and in settings where respondents may be more likely to
fear being identified.
We also acknowledge that, despite its frequent use in behavioral
ethics research (Trevin˜o & Weaver, 2003; Weber, 1992) and
evidence that relationships between individual differences and
unethical intentions and behaviors are similar in direction and
significance (Gephart et al., 2007), a scenario-based measure can
only simulate unethical decision making behavior. We did, how-
ever, take substantial steps to validate the Unethical Decision
Making measure, designing two validation exercises to demon-
strate a significant statistical link between our measure and behav-
ior. In the “keep the cash” validation exercise, the significant
correlation was of modest absolute size (.34). However, the results
are within the range that might be expected considering that (a) the
two items that correlated (a single scenario from the scale and a
related behavioral indicator) were designed to be only roughly
equivalent to avoid priming effects (Feldman & Lynch, 1988), and
(b) modest effect sizes in the .30 range may be near the upper limit
of what is possible when explaining relationships between com-
plex variables that are likely to be determined by multiple factors
(Swann, Chang-Schneider, & McClarty, 2007). In the “cheat–lie–
steal” validation exercise, we correlated responses to our Unethical
Decision Making measure with respondents’ self-reported lying,
385
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
cheating, and stealing behavior (r⫽.608, p⬍.001). Thus, the
evidence from these validation exercises shows significant statis-
tical relationships between our intentions-based Unethical Deci-
sion Making measure and actual behaviors.
These validation exercises were not designed to provide evi-
dence of a relationship between moral disengagement and behav-
ior. However, we also regressed the “keep the cash” binary out-
come variable on moral disengagement and the other individual
difference variables in a logistic regression analysis (n⫽59).
1
The
results were nonsignificant but trended in the hypothesized direc-
tion (e.g., the odds ratio for moral disengagement was 1.79). These
nonsignificant results could reflect a mismatch between the very
broad nature of the independent variables and the very narrow
single behavior outcome (Ajzen & Fishbein, 1977; Fisher, 1980;
Harrison, Newman, & Roth, 2006), the likely meager reliability
associated with a single behavioral item (Wanous & Hudy, 2001),
and/or the low power associated with a logistic regression using a
small sample with a nonoptimal base rate of the phenomenon
(Agresti, 2002; Long, 1997; Tosteson, Buzas, Demidenko & Kara-
gas, 2003). In regard to the first possibility, we also examined the
correlation between moral disengagement and a broader range of
unethical behaviors. Specifically, as part of the third exercise to
validate our Unethical Decision Making measure (discussed in the
Method section and in Appendix B), we included this study’s
24-item measure of moral disengagement and then computed the
correlation between respondents’ moral disengagement and their
self-reported cheating, lying, and stealing behaviors (collected 2
weeks later). The statistically significant correlation (r⫽.389, p⬍
.01) provides support for a relationship between moral disengage-
ment and unethical decisions, but clearly more research is needed
to definitively link moral disengagement to a range of unethical
behaviors that cause harm to organizations and society.
Practical Implications
The subtext underlying this study rests on our normative belief
that moral disengagement should be reduced because it leads to
unethical decisions. Moral disengagement is likely to be particu-
larly important in organizations because bureaucratic structures
and the division of labor seem to lend themselves to moral disen-
gagement mechanisms such as the diffusion and displacement of
responsibility (Bandura, 1986). Euphemistic labeling is also com-
mon in organizations, such as when managers refer to layoffs as
“rightsizing.” Also, with victims out of sight, globalization makes
it easier to ignore or distort the harmful consequences of business
actions. Thus, moral disengagement seems highly relevant to un-
derstanding unethical behavior in 21st century organizations.
Our individual difference results suggest that organizations can
identify individuals who are prone to moral disengagement. For
example, job applicants determined as a result of selection assess-
ments to be high on trait cynicism and chance locus of control
orientation or low on moral identity and empathy should be con-
sidered at risk for moral disengagement. Given the relationship
between moral disengagement and unethical decisions, organiza-
tions may choose to avoid hiring individuals with higher propen-
sity to morally disengage, particularly for jobs that are unsuper-
vised or ethically sensitive. Organizations might also explicitly
search for individuals who are low on moral disengagement, or
high on predictors such as empathy, for placement into ethically
sensitive positions.
Organizations may also attempt to influence existing employees
who are found to be more prone to moral disengagement. Training
could be developed to help such individuals uncover the ethical
dimensions of their work. This might include helping employees
recognize the most common euphemisms, distortions, and external
attributions associated with their particular job, company, or in-
dustry. Training could also be designed, in part, to integrate our
findings on specific individual differences that predict moral dis-
engagement. For example, training could aim to enhance empathy
for certain types of stakeholders (e.g., customers, distant employ-
ees). Research will be required to determine if specific interven-
tions have their intended effects, but others have found that cog-
nitive biases and distortions can be significantly reduced via
training interventions (c.f. Bazerman, 1994).
Organizations might also be able to guard against moral disen-
gagement by instituting decision making systems that are designed
to explicitly surface and address ethical issues. For example,
decision makers could be required to have multiple stakeholders
review certain types of decisions, or a decision making group
could assign an “ethical decision making advocate” (someone who
is not prone to moral disengagement) to listen specifically for signs
of moral disengagement (e.g., “it’s just part of the game,” “it
doesn’t hurt anyone,” “it’s their own fault”). Standardized lists of
questions might also be developed to help individuals and groups
identify the potential moral blind spots in their decision processes.
Conclusions
The current management environment places a premium on
understanding the drivers of ethical decision making. We therefore
conducted a study focusing on moral disengagement as a set of
sociocognitive thought processes influenced by individual differ-
ences and associated with unethical decision making. Clearly, we
have a long way to go toward understanding these mechanisms and
their antecedents and consequences as well as how they interact
with other individual differences and contextual influences. How-
ever, this research provides a useful starting point, and the results
suggest the potential value of further efforts to better understand
moral disengagement processes so we can test interventions that
might counter their negative effects.
1
This regression was conducted at the request of an anonymous re-
viewer.
References
Abraham, R. (2000). Organizational cynicism bases and consequences.
Genetic, Social, and General Psychology Monographs, 126, 269 –292.
Agresti, A. (2002). Categorical data analysis. Hoboken, NJ: Wiley.
Ajzen, I., & Fishbein, M. (1977). Attitude– behavior relations: A theoret-
ical analysis and review of empirical research. Psychological Bulletin,
84, 888 –918.
Ambrose, M. L., & Schminke, M. (1999). Sex differences in business
ethics: The importance of perceptions. Journal of Managerial Issues, 11,
454 – 474.
Anand, V., Ashforth, B. E., & Joshi, M. (2005). Business as usual: The
386 DETERT, KLEBE TREVIN
˜O, AND SWEITZER
acceptance and perpetuation of corruption in organizations. Academy of
Management Executive, 19, 9 –23.
Andersson, L. M., & Bateman, T. S. (1997). Cynicism in the workplace:
Some causes and effects. Journal of Organizational Behavior, 18, 449 –
469.
Aquino, K., & Reed, A. (2002). The self-importance of moral identity.
Journal of Personality and Social Psychology, 83, 1423–1440.
Aquino, K., Reed, A., Thau, S., & Freeman, D. (2007). A grotesque and
dark beauty: How moral identity and mechanisms of moral disengage-
ment influence cognitive and emotional reactions to war. Journal of
Experimental Social Psychology, 43, 385–392.
Bandura, A. (1986). Social foundations of thought and action: A social
cognitive theory. Englewood Cliffs, NJ: Prentice Hall.
Bandura, A. (1999). Moral disengagement in the preparation of inhuman-
ities. Personal and Social Psychology Review, 3, 193–209.
Bandura, A., Barbaranelli, C., Caprara, G. V., & Pastorelli, C. (1996).
Mechanisms of moral disengagement in the exercise of moral agency.
Journal of Personality and Social Psychology, 71, 364 –374.
Bandura, A., Caprara, G. V., Barbaranelli, C., Pastorelli, C., & Regalia, C.
(2001). Sociocognitive self-regulatory mechanisms governing transgres-
sive behavior. Journal of Personality and Social Psychology, 80, 125–
135.
Bandura, A., Underwood, B., & Fromson, M. E. (1975). Disinhibition of
aggression through diffusion of responsibility and dehumanization of
victims. Journal of Research in Personality, 9, 253–269.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator distinction
in social psychology: Conceptual, strategic, and statistical consider-
ations. Journal of Personality and Social Psychology, 51, 1171–1182.
Bar-Tal, D. (1989). Stereotyping and prejudice: Changing conceptions.
New York: Springer-Verlag.
Batson, C. D. (1991). The altruism question: Toward a social–
psychological answer. Hillsdale, NJ: Erlbaum.
Batson, C. D., Batson, J. G., Griffitt, C. A., Barrientos, S., Brandt, J. R.,
Sprengelmeyer, P., & Bayly, M. J. (1989). Negative-state relief and the
empathy–altruism hypothesis. Journal of Personality and Social Psy-
chology, 56, 922–933.
Bazerman, M. (1994). Judgment in managerial decision making. New
York: Wiley.
Bennett, R. J., & Robinson, S. L. (2003). The past, present, and future of
workplace deviance research. In J. Greenberg (Ed.), Organizational
behavior: The state of the science (2nd ed., pp. 247–282). Mahwah, NJ:
Erlbaum.
Bierhoff, H. (2002). Prosocial behavior. New York: Taylor & Francis.
Bok, S. (1998). Mayhem: Violence as public entertainment. New York:
Perseus Publishing.
Borrero-Hernandez, A. (1979). Unidimensional and multidimensional
measures of locus of control and their relationships to selected person-
ality variables. Unpublished doctoral dissertation, TexasA&MUni-
versity.
Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A
cognitive–motivational analysis. Psychological Bulletin, 86, 307–324.
Butterfield, K. D., Trevin˜o, L. K., & Weaver, G. R. (2000). Moral aware-
ness in business organizations: Influences of issue-related and social
context factors. Human Relations, 53, 981–1018.
Costa, P., Zonderman, A., McCrae, R., & Williams, R. (1985). Cynicism
and paranoid alienation in the Cook and Medley Ho scale. Psychoso-
matic Medicine, 48, 283–285.
Daniel, L. G., Blount, K. D., & Ferrell, C. M. (1991). Academic miscon-
duct among teacher education students. Research in Higher Education,
32, 703–725.
Davis, M. A. (1980). A multidimensional approach to individual differ-
ences in empathy. JSAS Catalog of Selected Documents in Psychology,
10, 85.
Diener, E. (1977). Deindividuation: Causes and consequences. Social
Behavior and Personality, 5, 143–156.
Duffy, M. K., Aquino, K., Tepper, B. J., Reed, A., & O’Leary-Kelly, A. M.
(2005, August). Moral disengagement and social identification: When
does being similar result in harm doing? Paper presented at the annual
meeting of the Academy of Management, Honolulu, HI.
Eisenberg, N. (1986). Altruistic emotion, cognition, and behavior. Hills-
dale, NJ: Erlbaum.
Eisenberg, N. (2000). Emotion, regulation, and moral development. Annual
Review of Psychology, 51, 665– 697.
Eisenberg, N., Guthrie, I. K., Murphy, B. C., Shepard, S. A., Cumberland,
A., & Carlo, G. (1999). Consistency and development of prosocial
dispositions: A longitudinal study. Child Development, 70, 1360 –1372.
Eisenberg, N., & Miller, P. A. (1987). The relation of empathy to prosocial
and related behaviors. Psychological Bulletin, 101, 91–119.
Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J.
(1999). Evaluating the use of exploratory factor analysis in psycholog-
ical research. Psychological Methods, 4, 272–299.
Feldman, J. M., & Lynch, J. G. (1988). Self-generated validity and other
effects of measurement on belief, attitude, intention, and behavior.
Journal of Applied Psychology, 73, 421– 435.
Fisher, C. D. (1980). On the dubious wisdom of expecting job satisfaction
to correlate with performance. Academy of Management Review, 5,
607– 612.
Flannery, B. L., & May, D. R. (2000). Environmental ethical decision
making in the U.S. metal finishing industry. Academy of Management
Journal, 43, 642– 662.
Gaertner, L., & Insko, C. A. (2000). Intergroup discrimination in the
minimal group paradigm: Categorization, reciprocation, or fear? Journal
of Personality and Social Psychology, 79, 77–94.
Gephart, J., Harrison, D., & Trevin˜o, L. K. (2007, August). The who, when,
and where of unethical choices: Meta-analytic answers to fundamental
ethics questions. Paper presented at the annual meeting of the Academy
of Management, Philadelphia, PA.
Goldberg, L. R. (2001). International Personality Item Pool: A scientific
collaboratory for the development of advanced measures of personality
traits and other individual differences. Retrieved December 6, 2007, at
http://ipip.ori.org/
Gordon, M. E., Slade, L. A., & Schmitt, N. (1986). The “science of the
sophomore” revisited: From conjecture to empiricism. Academy of Man-
agement Review, 11, 191–207.
Greenberg, J. (1987). The college sophomore as guinea pig: Setting the
record straight. Academy of Management Review, 12, 157–159.
Harrison, D. A., & McLaughlin, M. E. (1993). Cognitive processes in
self-report responses: Tests of item context effects in work attitude
measures. Journal of Applied Psychology, 78, 129 –140.
Harrison, D. A., Newman, D. A., & Roth, P. L. (2006). How important are job
attitudes? Meta-analytic comparisons of integrative behavioral outcomes
and time sequences. Academy of Management Journal, 49, 305–325.
Hegarty, W. H., & Sims, H. P. (1978). Some determinants of unethical
decision behavior: An experiment. Journal of Applied Psychology, 63,
451– 457.
Hochwarter, W. A., James, M., Johnson, D., & Ferris, G. R. (2004). The
interactive effects of politics perceptions and trait cynicism on work out-
comes. Journal of Leadership and Organizational Studies, 10, 44 –57.
Horovitz, B. (2002). Scandals grow out of CEO’s warped mind-set. Re-
trieved on February 14, 2007, from http://www.usatoday.com/money/
companies/management/2002.html
Johnson, J. L., & O’Leary-Kelly, A. M. (2003). The effects of psycholog-
ical contract breach and organizational cynicism: Not all social ex-
changes are created equal. Journal of Organizational Behavior, 24,
627– 647.
Jones, T. M. (1991). Ethical decision making by individuals in organiza-
387
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
tions: An issue-contingent model. Academy of Management Review, 16,
366 –395.
Kanter, D. L., & Mirvis, P. H. (1989). The cynical Americans. San
Francisco: Jossey-Bass.
Kelman, H. C. (1973). Violence without moral restraint: Reflections on the
dehumanization of victims and victimizers. Journal of Social Issues, 29,
25– 61.
Kohlberg, L. (1969). Stage and sequence: The cognitive– developmental
approach to socialization. In D. Goslin (Ed.), Handbook of socialization
theory and research (pp. 347– 480). Chicago: Rand McNally.
Lefcourt, H. M. (1966). Internal versus external control of reinforcements:
A review. Psychological Bulletin, 65, 206 –220.
Levenson, H. (1974). Activism and powerful others: Distinctions within
the concept of internal– external control. Journal of Personality Assess-
ment, 38, 377–383.
Levenson, H. (1981). Differentiating among internality, powerful others,
and chance. In H. M. Lefcourt (Ed.), Research with the locus of control
construct: Vol. 1. Assessment methods (pp. 15– 63). New York: Aca-
demic Press.
Lewis, G. (2006). Why do athletes cheat? Retrieved on February 13, 2007,
from http://www.msnbc.msn.com/id/14234026.html
Locke, E. A. (1986). Generalizing from laboratory to field: Ecological
validity or abstraction of essential elements. In E. A. Locke (Ed.),
Generalizing from laboratory to field settings (pp. 3–9). Lexington, MA:
Lexington Books
Long, S. J. (1997). Regression models for categorical and limited depen-
dent variables. Thousand Oaks, CA: Sage.
MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., &
Sheets, V. (2002). A comparison of methods to test mediation and other
intervening variable effects. Psychological Methods, 7, 83–104.
Maqsud, M. (1980). Locus of control and stages of moral reasoning.
Psychological Reports, 46, 1243–1248.
Markus, H., & Kunda, Z. (1986). Stability and malleability of the self-
concept. Journal of Personality and Social Psychology, 51, 858 – 866.
May, D. R., & Pauli, K. P. (2002). The role of moral intensity in moral
decision-making. Business and Society, 41, 84 –117.
McAlister, A. L. (2001). Moral disengagement: Measurement and modi-
fication. Journal of Peace Research, 38, 87–99.
McCabe, D., & Trevin˜o, L. K. (1993). Academic dishonesty: Honor codes
and other contextual influences. Journal of Higher Education, 64, 520 –
538.
McCabe, D., & Trevin˜o, L. K. (1995). Cheating among business students:
A challenge for business leaders and educators. Journal of Management
Education, 19(2), 205–218.
Messick, D. M., & Bazerman, M. H. (1996). Ethical leadership and the
psychology of decision making. Sloan Management Review, 37, 9 –22.
Miller, P. A., & Eisenberg, N. (1988). The relation of empathy to aggres-
sion and externalizing/antisocial behavior. Psychological Bulletin, 103,
324 –344.
Mudrack, P. E., & Mason, E. S. (1996). Individual ethical beliefs and
perceived organizational interests. Journal of Business Ethics, 15, 851–
861.
O’Fallon, M. J., & Butterfield, K. D. (2005). A review of the empirical
ethical decision-making literature: 1996 –2003. Journal of Business Eth-
ics, 59, 375– 413.
Opotow, S. (1990). Moral exclusion and injustice: An introduction. Jour-
nal of Social Issues, 4, 1–20.
Ostroff, C., Kinicki, A. J., & Clark, M. A. (2002). Substantive and
operational issues of response bias across levels of analysis: An example
of climate-satisfaction relationships. Journal of Applied Psychology, 87,
355–368.
Pelton, J., Gound, M., Forehand, R., & Brody, G. (2004). The moral
disengagement scale: Extension with an American minority sample.
Journal of Psychopathology and Behavioral Assessment, 26, 31–39.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003).
Common method biases in behavioral research: A critical review of the
literature and recommended remedies. Journal of Applied Psychology,
88, 879 –903.
Randall, D. M., & Fernandes, M. F. (1991). The social desirability re-
sponse bias in ethics research. Journal of Business Ethics, 10, 805– 817.
Reiss, M. C., & Mitra, K. (1998). The effects of individual difference
factors on the acceptability of ethical and unethical workplace behaviors.
Journal of Business Ethics, 17, 1581–1593.
Rest, J. R. (1986). Moral development: Advances in research and theory.
New York: Praeger.
Rest, J., Thoma, S. J., Moon, Y. L., & Getz, I. (1986). Different cultures,
sexes, and religions. In J. Rest (Ed.), Moral development: Advances in
research and theory (pp. 89 –132). New York: Praeger.
Rotter, J. B. (1966). Generalized expectancies for internal versus external
control of reinforcement. Psychological Monographs: General and Ap-
plied, 80, 1–28.
Rotter, J. B. (1980). Interpersonal trust, trustworthiness, and gullibility.
American Psychologist, 35, 1–7.
Safire, W. (1979, May 13). The fine art of euphemism. San Francisco
Chronicle, p. 13.
Samuelson, R. J. (2006). Delinquency of the CEOs. Retrieved on February
13, 2007, from http://www.washingtonpost.com/wp-dyn/content/
article.2006.html
Schminke, M., Ambrose, M. L., & Miles, J. A. (2003). The impact of
gender and setting on perceptions of others’ ethics. Sex Roles, 48,
361–375.
Schwartz, S. H. (1977). Normative influences on altruism. In L. Berkowits
(Ed.), Advances in Experimental Social Psychology (Vol. 10, pp. 221–
279). New York: Academic Press.
Seglin, J. L. (2000, October 15). When bribery is lost in translation. The
New York Times. Retrieved from http://www.nytimes.com
Seglin, J. L. (n.d.) Ethics Survey. Retrieved December 19, 2007, from
http://www.mbajungle.com/monthlysurvey/ethicssurvey.cfm
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-
experimental studies: New procedures and recommendations. Psycho-
logical Methods, 7, 422– 445.
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in
structural equations models. In S. Leinhart (Ed.), Sociological method-
ology (pp. 290 –312). San Francisco: Jossey-Bass.
Staub, E. (1989). The roots of evil: The origins of genocide and other group
violence. New York: Cambridge University Press.
Struch, N., & Schwartz, S. H. (1989). Intergroup aggression: Its predictors
and distinctness from in-group bias. Journal of Personality and Social
Psychology, 56, 364 –373.
Swann, W. B., Chang-Schneider, C., & McClarty, K. L. (2007). Do
people’s self-views matter? American Psychologist, 62, 84 –94.
Tenbrunsel, A. E., & Messick, D. M. (2004). Ethical fading: The role of
self-deception in unethical behavior. Social Justice Research, 17, 223–236.
Terpstra, D. E., Reyes, M. G. C., & Bokor, D. W. (1991). Predictors of
ethical decisions regarding insider trading. Journal of Business Ethics,
10, 699 –710.
Tosteson, T., Buzas, J. S., Demidenko, E., & Karagas, M. (2003). Power
and sample size calculations for generalized regression models with
covariate measurement error. Statistics in Medicine, 22, 1069 –1082.
Trevin˜o, L. K. (1986). Ethical decision making in organizations: A person–
situation interactionist model. Academy of Management Review, 11,
601– 617.
Trevin˜o, L. K., & Weaver, G. R. (2003). Managing ethics in business
organizations: Social scientific perspectives. Palo Alto, CA: Stanford
University Press.
Trevin˜o, L. K., & Youngblood, S. A. (1990). Bad apples in bad barrels: A
causal analysis of ethical decision-making behavior. Journal of Applied
Psychology, 74, 378 –385.
388 DETERT, KLEBE TREVIN
˜O, AND SWEITZER
Vardi, Y., & Weitz, E. (2003). Misbehavior in organizations. Englewood
Cliffs, NJ: Erlbaum.
Wanous, J. P., & Hudy, M. J. (2001). Single-item reliability: A replication
and extension. Organizational Research Methods, 4, 361–375.
Weber, J. (1992). Scenarios in business ethics research: Review, critical
assessment, and recommendations. Business Ethics Quarterly, 2, 137–160.
Wrightsman, L. S. (1992). Assumptions about human nature: Implications
for researchers and practitioners (2nd ed.). London: Sage.
Appendix A
Items Used to Assess Moral Disengagement
1. It is alright to fight to protect your friends. [MJ]
2. It’s ok to steal to take care of your family’s needs.
[MJ]
3. It’s ok to attack someone who threatens your fami-
ly’s honor. [MJ]
4. It is alright to lie to keep your friends out of trouble.
[MJ]
5. Sharing test questions is just a way of helping your
friends. [EL]
6. Talking about people behind their backs is just part
of the game. [EL]
7. Looking at a friend’s homework without permission
is just “borrowing it.” [EL]
8. It is not bad to “get high” once in a while. [EL]
9. Damaging some property is no big deal when you
consider that others are beating up people. [AC]
10. Stealing some money is not too serious compared to
those who steal a lot of money. [AC]
11. Not working very hard in school is really no big deal when
you consider that other people are probably cheating. [AC]
12. Compared to other illegal things people do, taking
some things from a store without paying for them is
not very serious. [AC]
13. If people are living under bad conditions, they can-
not be blamed for behaving aggressively. [DISR]
14. If the professor doesn’t discipline cheaters, students
should not be blamed for cheating. [DISR]
15. If someone is pressured into doing something, they
shouldn’t be blamed for it. [DISR]
16. People cannot be blamed for misbehaving if their
friends pressured them to do it. [DISR]
17. A member of a group or team should not be blamed
for the trouble the team caused. [DIFR]
18. A student who only suggests breaking the rules should
not be blamed if other students go ahead and do it.
[DIFR]
19. If a group decides together to do something harmful,
it is unfair to blame any one member of the group for
it. [DIFR]
20. You can’t blame a person who plays only a small
part in the harm caused by a group. [DIFR]
21. It is ok to tell small lies because they don’t really do any
harm. [DC]
22. People don’t mind being teased because it shows
interest in them. [DC]
23. Teasing someone does not really hurt them. [DC]
24. Insults don’t really hurt anyone. [DC]
25. If students misbehave in class, it’s their teacher’s fault.
[AB]
26. If someone leaves something lying around, it’s their
own fault if it gets stolen. [AB]
27. People who are mistreated have usually done things
to deserve it. [AB]
28. People are not at fault for misbehaving at work if
their managers mistreat them. [AB]
29. Some people deserve to be treated like animals.
[DEH]
30. It is ok to treat badly someone who behaved like a
“worm.” [DEH]
31. Someone who is obnoxious does not deserve to be
treated like a human being. [DEH]
32. Some people have to be treated roughly because they
lack feelings that can be hurt. [DEH]
Notes. Items in bold compose the 24-item scale used in the study.
Items not in bold represent items dropped based on factor analysis.
MJ ⫽moral justification; EL ⫽euphemistic labeling; AC ⫽advan-
tageous comparison; DISR ⫽displacement of responsibility; DIFR ⫽
diffusion of responsibility; DC ⫽distortion of consequences; AB ⫽
attribution of blame; DEH ⫽dehumanization.
(Appendixes continue)
389
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
Appendix B
Development and Validation of the Unethical Decision Making Scale
The Unethical Decision Making Scale
Study respondents were asked “How likely is it that you
would engage in the behavior described?” Each of the following
scenarios was rated on a 7-point scale ranging from 0 (not at all
likely)to6(highly likely). Cronbach’s alpha for the eight-item
scale ⫽.74.
1. You work in a fast-food restaurant in downtown [City X]. It’s
against policy to eat food without paying for it. You came straight
from classes and are therefore hungry. Your supervisor isn’t
around, so you make something for yourself and eat it without
paying.
2. You work as an office assistant for a department at [Univer-
sity Y]. You’re alone in the office making copies and realize
you’re out of copy paper at home. You therefore slip a ream of
paper into your backpack.
3. You’re preparing for the final exam in a class where the
professor uses the same exam in both sections. Some of your
friends somehow get a copy of the exam after the first section.
They are now trying to memorize the right answers. You don’t
look at the exam, but just ask them what topics you should focus
your studying on.
4. You’ve waited in line for 10 minutes to buy a coffee and
muffin at Starbucks. When you’re a couple of blocks away, you
realize that the clerk gave you change for $20 rather than for the
$10 you gave him. You savor your coffee, muffin and free $10.
5. You get the final exam back from your professor and you
notice that he’s marked correct three answers that you got wrong.
Revealing his error would mean the difference between an A and
a B. You say nothing.
6. Your accounting course requires you to purchase a software
package that sells for $50. Your friend, who is also in the class, has
already bought the software and offers to lend it to you. You take
it and load it onto your computer.
7. Your boss at your summer job asks you to get confidential
information about a competitor’s product. You therefore pose as a
student doing a research project on the competitor’s company and
ask for the information.
8. You are assigned a team project in one of your courses. Your
team waits until the last minute to begin working. Several team
members suggest using an old project out of their fraternity/
sorority files. You go along with this plan.
Construct Validation Exercise 1
Twenty business ethics experts were contacted via e-mail and
asked to rate the scenarios based on whether the behavior in each
scenario represents a violation of one or more ethical principles
(e.g., utilitarianism, rights, justice). Sixteen experts responded. All
of the respondents hold doctoral degrees and have an average of
26.6 years of business ethics-related teaching experience and an
average of 55 ethics-related publications. The experts were given
a total of 13 scenarios to rate—the 8 composing the Unethical
Decision Making scale (see above) plus 5 randomly interspersed
additional scenarios (see below) created to represent behaviors that
were not thought to be unethical. Responses (on a 5-point Likert
scale) were used to determine whether the ethics experts viewed
the eight scenarios used in our scale as a violation of one or more
ethical principles and whether the grand mean for “violates one or
more ethical principles” for the 8 scenarios used in our Unethical
Decision Making scale differed significantly from the grand mean
on the other 5 scenarios. Findings indicated that the ethics experts
agreed that the 8 scenarios used in the scale represent violations of
ethical principles: The grand mean for the items used in our
Unethical Decision Making scale (Items 1– 8) was 4.66 (SD ⫽.35;
with a range of 4.44 –5.00 for the means of the individual scenar-
ios). Conversely, the grand mean for the 5 scenarios not included
in our scale (Items 9 –13 below) was 1.68 (SD ⫽.62) with a range
for the means of the individual scenarios from 1.25 to 2.06. The
difference between these means (M⫽4.66 vs. 1.68) was statisti-
cally significant, t(15) ⫽15.3, p⬍.001.
9. You and your roommates draw straws for choosing bedrooms
in the 3 bedroom house you’re renting. You draw the longest straw
and get first choice so you choose the biggest room.
10. Your professor in a large lecture class requires that your
assignments be submitted in hard copy before the end of class.
When you get to class on Thursday, you realize that you forgot to
bring your assignment. So, you decide to skip class, rush home,
print the assignment and rush back to turn it in before class is over.
11. You need 3 more general education credits. After asking
around, you narrow your choice to two. You’ve heard that one is
very valuable but hard. The other is less valuable but an easy A.
You choose the easy A.
12. Your professor asks you to complete a voluntary survey, but
you don’t because you’re just too busy.
13. You start college in September and receive a scholarship that
requires you to perform 40 hours of community service within 12
months. Since you already committed to volunteering at your
hometown church next summer, you decline other community
service opportunities, figuring that the service at your church will
be enough.
Construct Validation Exercise 2
Sixty subjects were randomly selected from the final sample of
307 subjects. Each subject was sent a letter from a strategy
doctoral candidate in the authors’ departments in which the doc-
toral candidate explained that, as part of his dissertation research
on strategic decision making, he had recently conducted a short lab
study (on make, buy, or ally decisions in international expansions)
for which $8 had been promised for participation. The doctoral
candidate’s letter further informed the subjects that his computer
had crashed and, as a result, his list of subjects had been lost. This
left him with only a paper copy of the original list of all potential
subjects (from which his actual subjects had been selected). The
doctoral candidate’s letter continued by stating that he was keeping
his end of the bargain by sending all of the potential subjects the
money but requesting that nonsubjects return the money because
390 DETERT, KLEBE TREVIN
˜O, AND SWEITZER
getting money back was critical for completing the study. Thus,
the doctoral candidate’s mailing to each of the students contained
his letter, $8 cash, and a return-addressed envelope to be sent using
free campus mail. The return envelope and bills (a $5 and three
$1s) were marked, using an ultraviolet pen, with a unique identifier
so that the behavior (keep or return the cash) could be linked to
responses on a similar unethical decision behavioral intention
scenario (see below).
In reality, the doctoral student had conducted no such lab study
and therefore none of the 60 recipients of the $8 was actually
entitled to it. The real purpose of the exercise was to see which
students would behave unethically by keeping the money. (All
subjects were subsequently thoroughly debriefed, and those who
had returned the cash were given the $8 to keep.) This scenario
was designed to be similar to, yet not entirely overlapping with (to
avoid cognitive priming or memory effects; Feldman & Lynch,
1988), the “keep the extra change at Starbucks” scenario from our
Unethical Decision Making measure (see Item 4 in this appendix)
so that a correlation between a stated behavioral intention and an
actual behavior could be computed. To be consistent with the
unethical decision scenarios (where unethical behavior was indi-
cated by responses at the high end of the scale), we coded “kept the
cash” in the bogus lab study as 1 and “returned the cash” as 0.
Thirty-nine of 59 recipients (one letter was returned as undeliver-
able) kept the cash despite the ease of returning it and the fact that
they had not earned it. Because neither of the variables is normally
distributed, we computed a polychoric correlation coefficient be-
tween the two variables (the behavioral intention regarding the
extra change at Starbucks and the behavioral choice regarding the
$8 in the bogus lab exercise). At .342, this correlation coefficient
is statistically significant at p⬍.01.
Construct Validation Exercise 3
On the first day of class, 85 students enrolled in a freshman
business class were invited to complete two surveys during class
time. Survey 1, administered by one of the authors during Class 1,
included the eight scenarios composing this study’s Unethical
Decision Making scale (see above) and the 24 items used in the
final Moral Disengagement scale (see Appendix A). Two weeks
later, one author again visited the class to administer a second
survey that asked students to report on their own and others’
cheating, lying, and stealing behaviors while in high school. Our
intent in developing the Cheat–Lie–Steal scale was to collect data
on multiple types of unethical behavior relevant to subjects this
age. Thus, the 13 items composing the Cheat–Lie–Steal scale
(shown below) represent a combination of items drawn from the
extant literature (4 items were used or adapted from McCabe and
Trevin˜o, 1993, and 3 from Daniel, Blount, and Ferrell, 1991) and
items we developed for this study.
1. Lying to my parents about my school performance
2. Exaggerating my accomplishments on my college ap-
plication
3. Lying about my age
4. Using a false excuse to delay taking an exam or turning
in an assignment
5. Claiming to have turned in an assignment when I have
not
6. Taking low-cost items from a retail store
7. Taking small amounts of money from my parents’ wal-
let without their permission
8. Copying from another student on a test
9. Collaborating or receiving substantial help on an assign-
ment when the instructor asked for individual work
10. Helping someone else to cheat on a test
11. Copying material and turning it in as your own work
12. Asking another student who has previously taken a quiz
or exam for the questions or the answers prior to taking
the test
13. Changing a response after a paper or exam is returned
and then reporting a grade error to the instructor
To decrease respondents’ propensity to underreport their own
behavior, they were first asked to rate how frequently they ob-
served others engaging in each of the 13 behaviors shown below
while in high school, using a 5-point Likert scale ranging from 1
(never)to5(many times). They were then presented the same 13
items and asked to indicate how frequently they engaged in each
behavior. Because 25 students were not at least age 18 at the time
of the first survey and because 2 did not provide a unique ID
required for matching the two surveys, the final sample for this
validation exercise was n⫽58. Estimated reliabilities for the
Moral Disengagement, Unethical Decision Making, and Cheat–
Lie–Steal (own behavior) scales were .81, .80, and .80, respec-
tively.
To assess whether reported intentions to engage in the eight
unethical behaviors indicated in this study’s Unethical Decision
Making scale were related to actual unethical behavior engaged in
by the same respondents, we computed the correlation between the
Unethical Decision Making and Cheat–Lie–Steal scales (which
had been collected with 2 weeks separation to reduce potential
common method bias—Podsakoff, MacKenzie, Lee, & Podsakoff,
2003). At r⫽.61 ( p⬍.001), results indicate that those who have
actually engaged in a variety of unethical cheating, lying, and
stealing behaviors were significantly more likely to report their
intention to behave unethically on the scenarios composing this
study’s dependent variable.
Received September 20, 2006
Revision received August 29, 2007
Accepted November 1, 2007 䡲
391
MORAL DISENGAGEMENT IN ETHICAL DECISION MAKING
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Applied Psychology
This content is subject to copyright. Terms and conditions apply.
