Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Journal of Applied Psychology Copyright 2008 by the American Psychological Association

2008, Vol. 93, No. 5, 1104 1117 0021-9010/08/$12.00 DOI: 10.1037/0021-9010.93.5.1104

When Employees Strike Back: Investigating Mediating Mechanisms


Between Psychological Contract Breach and Workplace Deviance

Prashant Bordia Simon Lloyd D. Restubog


University of South Australia University of New South Wales

Robert L. Tang
De La SalleCollege of Saint Benilde

In this article, psychological contract breach, revenge, and workplace deviance are brought together to
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

identify the cognitive, affective, and motivational underpinnings of workplace deviance. On the basis of
S. L. Robinson and R. J. Bennetts (1997) model of workplace deviance, the authors proposed that breach
(a cognitive appraisal) and violation (an affective response) initiate revenge seeking. Motivated by
revenge, employees then engage in workplace deviance. Three studies tested these ideas. All of the
studies supported the hypothesized relationships. In addition, self-control was found to be a moderator
of the relationship between revenge cognitions and deviant acts; the relationship was weaker for people
high in self-control.

Keywords: feelings of violation, psychological contract breach, revenge, self-control, workplace devi-
ance

Workplace deviance has been defined as voluntary behavior that ple, Aquino et al. (1999) found that interactional justice was
violates significant organizational norms and in so doing threatens the negatively related to organizational deviance and that distributive
well-being of an organization, its members, or both (Robinson & and interactional justice were negatively related to interpersonal
Bennett, 1995, p. 556). There are three key elements of this definition. deviance. From a social exchange perspective, an unfavorable or
First, deviant behaviors are volitional and motivated, not accidental. unsupportive work environment may be reciprocated with work-
Second, these behaviors are deviant, in that they violate the norms of place deviance (Colbert, Mount, Harter, Witt, & Barrick, 2004).
the dominant administrative coalition (Robinson & Bennett, 1997, Indeed, Colbert et al. found that perception of a supportive work
p. 6). Third, some of these behaviors (referred to as organizational environment was negatively related to organizational as well as
deviance) can be targeted at and therefore are harmful to the organi- interpersonal deviance. Finally, several dispositional characteris-
zation, and some (referred to as interpersonal deviance) can be tar- tics are related to workplace deviance (e.g., conscientiousness,
geted at other individuals at work. agreeableness, and emotional stability; Colbert et al., 2004).
Several theoretical perspectives have been used to examine Although prior research explicitly measured the situational and
workplace deviance. From the standpoint of equity and justice dispositional antecedents of workplace deviance, an empirical
theories, deviant behavior is considered an intentional act moti- examination of the mediating affective and motivational processes
vated by the need to restore equity or seek retributive justice (e.g., the need to restore equity or seek retribution) was lacking
(Aquino, Lewis, & Bradfield, 1999; Greenberg, 1990). For exam- (Griffin & Lopez, 2005; OLeary-Kelly, Griffin, & Glew, 1996;
Robinson & Bennett, 1997). In the research reported here, we
looked at the role of psychological contract breach as an anteced-
ent to deviant behaviors. In addition, we tested the mediating role
Prashant Bordia, School of Management, University of South Australia;
Simon Lloyd D. Restubog, Australian School of Business, University of of affect (feelings of violation) and motivational intent (revenge)
New South Wales; Robert L. Tang, School of Management and Informa- in the breach deviance relationship. Specifically, we predicted a
tion Technology, De La SalleCollege of Saint Benilde. chain of mediation as follows: breach, violation, revenge, devi-
An earlier version of this article was presented at the 66th Annual ance. Moreover, given the importance of dispositional variables in
Meeting of the Academy of Management, Atlanta, Georgia, August 2006. moderating the behavioral manifestations of antisocial and aggres-
We thank Elizabeth George, Maria Kraimer, and Carol Kulik for their sive behavior in the workplace (Douglas & Martinko, 2001), we
comments on the manuscript of this article and Cheri Ostroff for advice on tested the effects of self-control in moderating the relationship
the statistical analysis. We are grateful to Miguel Alfonso Abesamis, Eloisa between revenge motivations and workplace deviance.
Charmaine Bolinao, Kristina Cera, Geronimo Dimaano, Parker Ong, Mary
This article makes several contributions to the literature. In it,
Ann Parina, and Anne Margrette Siendo for assistance in data collection
and Peter Lemuel Cayayan for logistical support and data entry.
we bring together the literatures on psychological contract breach
Correspondence concerning this article should be addressed to Pra- and violation, organizational revenge, workplace deviance, and
shant Bordia, School of Management, University of South Australia, individual differences to develop a framework that addresses two
GPO Box 2471, Adelaide, South Australia 5001, Australia. E-mail: critical questions. First, what is the process by which psycholog-
prashant.bordia@unisa.edu.au ical contract breach leads to deviant behavior? Although empirical

1104
CONTRACT BREACH, REVENGE, AND WORKPLACE DEVIANCE 1105

research has made considerable gains in identifying the situational limit the translation of revenge motivations into workplace devi-
and dispositional antecedents of deviance, the role of motivational ance. However, there is a lack of research on dispositional char-
intent underlying deviance has been ignored (Griffin & Lopez, acteristics that constrain the likelihood of revenge seeking (Brad-
2005), even though intentionality is a defining characteristic of field & Aquino, 1999). Self-control, as a dispositional variable,
workplace deviance (Bennett & Robinson, 2000). Indeed, this is serves as a self-regulatory mechanism that holds people back from
the case with much of the research on negative behaviors in reacting to provocations in aggressive ways (Brown, Westbrook,
organizations. In a review of this literature, Griffin and Lopez & Challagalla, 2005). This article examines the direct effect of
(2005) concluded that in virtually all cases, researchers studying self-control, as well as the interactive effect of self-control and
bad behaviors have focused almost exclusively on the behaviors revenge, in the prediction of workplace deviance.
themselves . . . [and] there have been few detailed discussions and We tested the proposed relationships in a programmatic series of
virtually no in-depth analyses of why individuals elect to pursue three studies conducted with working adults as participants. Study
[these] behaviors (p. 995). The test of the mediating processes 1 involved a longitudinal test of the relationship between employee
between breach and deviance therefore fills a major gap in the reports of psychological contract breach at Time 1 and workplace
literature; a test of mediators contributes to our understanding of deviance obtained from employee records at Time 2 (20 months
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

how and why breach leads to deviance. From a theoretical stand- later). Study 2 is a survey study in which we used a facet measure
point, focusing solely on the overt behavior can be misleading, as of contract breach (i.e., transactional and relational dimensions of
an ostensibly harmful act might have pro-social intent (as demon- the psychological contract) and tested the mediating roles of feel-
strated by pro-social rule breaking; Morrison, 2006). Moreover, an ings of violation and revenge seeking in the prediction of organi-
understanding of the intent underlying deviant behavior is impor- zational workplace deviance (i.e., deviant acts directed toward the
tant for gauging the threat to organizational well-being. By exam- organization). Study 3 is a survey study in which we replicated the
ining the role of revenge seeking, we explicitly test for motiva- mediating roles of violation and revenge cognitions in the predic-
tional underpinnings of deviant behaviors. tion of organizational and interpersonal workplace deviance (i.e.,
Similarly, an empirical analysis of motivational intent is missing deviant acts directed toward coworkers). In addition, we tested the
from the psychological contract breach literature (Coyle-Shapiro moderating role of self-control in the relationship between revenge
& Conway, 2004). The harmful effects of breach are explained by cognitions and the two types of workplace deviance. Research on
social exchange theory (Coyle-Shapiro & Conway, 2004; Rous- workplace deviance has largely relied on self-report methodology
seau, 1995); the employer employee relationship is governed by and cross-sectional designs (Fox & Spector, 2005). Fox and Spec-
the rules of social exchange, and when the organization is per- tor have strongly encouraged researchers to search for creative
ceived to break a promise, employees reciprocate by hurting or- means to study these types of behaviors. We obtained deviance
ganizational interests (e.g., withholding effort, engaging in anti- data from organizational records in Study 1 and from supervisor
citizenship behaviors, and exiting the organization). However, the reports in Studies 2 and 3. Finally, we took steps to minimize
mediating mechanisms that demonstrate exchange-based motiva- common method variance and strengthen internal validity. For
tional intent (e.g., retaliation for perceived harm) are usually not example, self-reports of breach, violations, and revenge cognitions
measured. Our proposed model extends research on consequences might be influenced by social desirability motives. To test for these
of psychological contract breach by demonstrating that revenge effects, we measured social desirability and incorporated it in the
can be an outcome of breach and can provide the motivational data analysis.
intent in the relationship between breach and workplace deviance.
Further, the theoretical work on psychological contracts distin- Theoretical Framework and Hypotheses
guishes between perceptions of contract breach (a cognitive as-
sessment) and feelings of violation (an affective reaction that Our approach is based on the model of workplace deviance
follows from a perception of contract breach; Robinson & Morri- proposed by Robinson and Bennett (1997). According to this
son, 2000). However, the mediating role of feelings of violation model, deviance is instigated by provocations in the workplace,
between perceptions of breach and outcomes has received limited such as unfair or poor working conditions and injustice. These
attention. In particular, the relationship of violation with facet workplace instigators lead to cognitions of disparity (i.e., an as-
measures of contract breach (i.e., breach of transactional and sessment that the situation is lacking in what was expected) and an
relational inducements) has never been tested. Although Robinson emotional reaction of outrage. These cognitions and emotional
and Morrison provided an empirical test of their theoretical model reactions instigate motivations to restore the disparity (instrumen-
(i.e., relationship between psychological contract breach and feel- tal motivation) or to vent the feelings of outrage (expressive
ings of violation), they did not discuss the consequences of feel- motivation). The two motivations are closely intertwined, and the
ings of violations, aside from the emotional response of violation same deviant behavior can fulfill both motivations. Finally, not
itself. every impulse to engage in deviance is acted upon; constraints
We address a second question: What factors constrain the en- such as fear of punishment or internalized norms against deviance
actment of deviance? From a practical standpoint, the identifica- act to suppress the impulse.
tion of the series of cognitions and emotions elicited by the In our proposed model, the breach of psychological contract acts
experience of breachand ending in deviant behaviorswill en- as a provocation for deviance. Perceptions of breach represent
hance our understanding of the stepwise process by which the cognitions of disparity (i.e., a perceived mismatch between what
experience of broken promises results in deviant behaviors; inter- was promised by the organization vs. what was received). These
ventions can then be developed at each of these steps to curtail the cognitions lead to feelings of violation (i.e., feelings of betrayal,
enactment of deviance. In addition, personality characteristics can frustration, and anger in response to the breach) and represent the
1106 BORDIA, RESTUBOG, AND TANG

outrage element of the workplace deviance model (Robinson & seau, 1994). To improve the definitional clarity, Morrison and
Bennett, 1997). These cognitions and emotions develop into a need Robinson (1997) distinguished between psychological contract
for revenge. Revenge fulfills both instrumental and expressive breach and feelings of violation accompanied by strong emotions.
motivations; the aim of revenge is to vent the frustration, redress Psychological contract breach, a cognitive appraisal, emanates
the imbalance in the exchange relationship, and punish the orga- from the employees perception that there is a discrepancy be-
nization for the breach. The revenge motivation leads to deviant tween what was promised versus what was actually delivered.
behaviors. Finally, the personality variable of self-control acts as a Violation refers to the affective reaction that follows from this
constraint and moderates the relationship between revenge and cognitive appraisal of breach. It consists of various emotions
deviant behaviors (i.e., the relationship between revenge and de- resulting from the fact that expected outcomes did not eventuate
viance will be stronger for people low in self-control). In the (e.g., disappointment and frustration) andat a deeper level
following sections, we review the literature on each of the key feelings of betrayal, anger, and bitterness due to broken promises
variables and formally propose the hypotheses. (Morrison & Robinson, 1997; Robinson & Morrison, 2000). In
addition, Morrison and Robinson argued that violation would
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

mediate the relationship between perceptions of breach and attitu-


Psychological Contract Breach and Feelings of Violation
This document is copyrighted by the American Psychological Association or one of its allied publishers.

dinal and behavioral outcomes. The mediating role of violation can


A psychological contract consists of a set of beliefs involving be explained using affective events theory (Weiss & Cropanzano,
terms and exchange agreement between the employee and his or 1996; Zhao et al., 2007), which argues that experiences at work
her employing organization (Rousseau, 1995). An important at- evoke affective reactions that, in turn, influence attitudes and
tribute of the psychological contract is that it is inherently percep- behaviors. Indeed, there is evidence to suggest that feelings of
tual and exists in the eye of the beholder (Rousseau, 1995, p. 6). violation serve as an intermediary variable that links contract
In other words, all employees have their own perceived under- breach and work attitudes (Raja, Johns, & Ntalianis, 2004; Zhao et
standing of the reciprocal obligations that exist between them- al., 2007). Contract breach elicits these negative work attitudes
selves and the organization. Psychological contracts can be clas- (i.e., reduced affective commitment and job satisfaction and in-
sified into two dimensions: transactional and relational (Rousseau, creased turnover intentions) as a result of the strong sense of
1995). The transactional dimension, characterized as inflexible and violation experienced. In the context of the current research, we
short term, includes organizational benefits, competitive wages, predicted that violation would mediate the relationship between
and merit pay based on current performance. Conversely, the breach and revenge. In the next section, we introduce the revenge
relational dimension is contingent on less specific exchanges and literature and formally propose the hypotheses linking breach,
a long-term relationship with the organization and includes career violation, and revenge.
growth opportunities, training and development, and support from
supervisors (Robinson & Rousseau, 1994). The distinction be-
tween the transactional and relational dimensions of the psycho- Revenge in an Organizational Context
logical contract has been empirically demonstrated (Restubog &
Revenge can be defined as an effort by the victim of harm to
Bordia, 2006; Robinson & Rousseau, 1994; Turnley, Bolino,
inflict damage, injury, discomfort, or punishment on the party
Lester, & Bloodgood, 2003).
judged responsible for causing the harm (Aquino, Tripp, & Bies,
Psychological contract breach takes place when one party in a
2001, p. 53). What makes aggressive actions revengeful is the
relationship perceives that the other party has neglected to fulfill
intention to retaliate or reciprocate perceived harm by harming in
what has been committed or promised (Rousseau, 1995). As pre-
return. There has been growing interest in investigating revenge
dicted by social exchange theory, employees respond to a breach
dynamics in an organizational context. Research has focused on
of psychological contract by the organization in a variety of
individual differences (e.g., negative reciprocity norm endorse-
negative ways (Rousseau, 1995; Zhao, Wayne, Glibkowski, &
ment, Eisenberger, Lynch, Aselage, & Rohdieck, 2004; agreeable-
Bravo, 2007). Employee perceptions of breach of the organiza-
ness, Skarlicki, Folger, & Tesluk, 1999) and organizational ante-
tional psychological contract are negatively related to job satisfac-
cedents (e.g., injustice, violation of promises, loss of status or
tion (Robinson & Rousseau, 1994), organizational citizenship be-
power, Bies & Tripp, 2001; interpersonal offense, Aquino et al.,
haviors (Restubog, Bordia, & Tang, 2006, 2007; Robinson &
2001) that contribute to revenge seeking.
Morrison, 1995), trust (Robinson, 1996), and employee perfor-
Revenge may be used to restore equity or remedy perceived
mance (Restubog et al., 2006; Turnley et al., 2003). Along similar
harm (Bies & Tripp, 2001). Bies, Tripp, and Kramer (1996)
lines, breach is positively related to anticitizenship behaviors
proposed that revenge may unfold through several stages. First,
(Kickul, Neuman, Parker, & Finkl, 2001), absenteeism (Deery,
sparking events (Bies et al., 1996, p. 20), such as perceived
Iverson, & Walsh, 2006), and intentions to leave the organization
injustice or victimization, may trigger the need for revenge. These
(Robinson & Rousseau, 1994). On the basis of these studies, we
events are characterized by broken promises or violated norms and
made the following hypothesis:
expectations and may involve hurt esteem or social identity. These
Hypothesis 1: Psychological contract breach will be posi- events lead to feelings of anger and thoughts of revenge and
tively related to workplace deviance. constitute a heating up phase in which strong emotions accom-
panied by thoughts of revenge are experienced (Bies et al., 1996,
In research literature of the mid-1990s, the term violation was p. 21). Finally, people may react in various ways (Tripp & Bies,
used interchangeably with breach (both implying perception of 1997). Some may choose to forgive or do nothing. Fear of retal-
broken promises; Robinson & Morrison, 1995; Robinson & Rous- iation may hold some back from seeking revenge. Motivated by
CONTRACT BREACH, REVENGE, AND WORKPLACE DEVIANCE 1107

the need to seek revenge, some may act in ways to harm the work; Marcus & Schuler, 2004). Moreover, people low in self-
wrongdoer. control are more likely to react to situational triggers with criminal
Revenge acts are generally considered to be a thought-out, acts (i.e., self-control would moderate the relationship between
calculated, and considered course of action, often preceded by triggers and hostile reactions). On the basis of this characteriza-
ruminations that may involve fantasies of hurting or causing tion, we expected self-control would have direct and moderating
harm to the wrongdoer (Bies et al., 1996; Bradfield & Aquino, effects in the context of workplace deviance. Indeed, there is
1999). These revenge cognitions represent the motivational intent research evidence for both types of effects (Douglas & Martinko,
and precede the enactment of harmful behaviors directed at the 2001; Hepworth & Towler, 2004). For example, Sarchione et al.
target of revenge (Bradfield & Aquino, 1999). In our model, (1998) found that police officers who engaged in dysfunctional
revenge is treated as a motivational antecedent (operationalized in behavior (including insubordination, use of excessive force and
the form of revenge cognitions); workplace deviance results from racial slurs toward the public, and sexual misconduct) were sig-
revenge cognitions. We argued that psychological contract breach nificantly lower in self-control in comparison with a control group
and the associated feelings of anger and betrayal (i.e., violation) of police officers who did not engage in dysfunctional behaviors.
would lead to thoughts of revenge. These revenge cognitions occur On the basis of this research, we expected that self-control would
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

prior to the enactment of deviant behaviors (Bradfield & Aquino, be negatively related to workplace deviance (i.e., there would be a
1999). direct effect).

Hypothesis 2: The relationship between psychological con- Hypothesis 4: Self-control will be negatively related to work-
tract breach and revenge cognitions is mediated by feelings of place deviant behaviors.
violation.
Self-control has been found to attenuate (moderate) the adverse
Hypothesis 3: The relationship between feelings of violation effects of negative emotions on work performance (Brown et al.,
and workplace deviant behaviors is mediated by revenge 2005; Marcus & Schuler, 2004). Therefore, we expected that
cognitions. self-control would keep people from acting on revenge cognitions.
The relationship between revenge cognitions and workplace devi-
ance would be stronger for people low in self-control, compared
Dispositional Constraints on Workplace Deviance: Direct with people high in self-control (i.e., there would be a moderating
and Moderating Effects of Self-Control effect).
Workplace deviance is influenced by situational as well as
Hypothesis 5: Self-control will moderate the relationship be-
dispositional factors (Colbert et al., 2004; Douglas & Martinko,
tween revenge cognitions and workplace deviant behaviors.
2001; Marcus & Schuler, 2004). Dispositional factors may have a
There will be a stronger positive relationship between re-
direct effect on workplace deviance. For example, Douglas and
venge cognitions and workplace deviant behaviors for those
Martinko found that a combination of individual difference mea-
employees with low levels of self-control than for employees
sures (e.g., trait anger, attribution style, and self-control) accounted
with high levels of self-control.
for 62% of variance in self-reported workplace aggression. Dis-
positional variables may also moderate the relationship between Study 1
situational factors and workplace deviance (Colbert et al., 2004).
For example, Skarlicki et al. (1999) found that negative affectivity In this study, we examined the direct relationship between
interacted with perceptions of distributive and interactional justice psychological contract breach and workplace deviance. Much of
in the prediction of retaliatory behaviors. When situational factors the research linking deviance and its antecedents has been cross-
instigate deviant behaviors, dispositional factors may amplify this sectional in nature (Fox & Spector, 2005). This fact brings into
effect (as in the case of trait anger; Hepworth & Towler, 2004) or question the causal direction of relationships. For example, a
attenuate this effect (as in the case of self-control; Brown et al., conclusion that perceptions of injustice lead to workplace deviance
2005). There is a growing literature on the relationship between is open to an alternative explanation that the perceived (i.e.,
self-control and aggressive or harmful behaviors in the workplace. self-reported) injustice may be a justification or rationalization by
However, the role of self-control in inhibiting the enactment of employees who are engaged in deviant acts.1 In light of this
revenge intentions has not been investigated. In this study, we limitation of existing research, we conducted a longitudinal study.
considered both direct and moderating effects of self-control on We measured perceptions of psychological contract breach at
workplace deviance. Time 1 and collected data on workplace deviance, in the form of
Self-control refers to the level of an individuals restraint from workplace offenses committed by the employees, at Time 2 (20
indulging in negative action tendencies that might further compli- months after the Time 1 data collection).
cate or damage the situation (Brown et al., 2005, p. 794). Self-
Method
control has been referred to as freedom from impulsivity (Sar-
chione, Cuttler, Muchinsky, & Nelson-Gray, 1998, p. 905) and has Participants and Procedure
been found to be negatively related to a range of dysfunctional The study was conducted in a public-sector organization in the
behaviors in the workplace (Marcus & Schuler, 2004). According Philippines. At Time 1, surveys were distributed to 300 employees
to the general theory of crime (Gottfredson & Hirschi, 1990),
criminal behavior is largely a result of low self-control in the
1
perpetrators of crime (including counterproductive behaviors at We are grateful to an anonymous reviewer for this comment.
1108 BORDIA, RESTUBOG, AND TANG

during a training session. The content of the training was not in any of psychological contract breach). These offenses are an appropri-
way related to the current study. The questionnaires were prepared ate operationalization of workplace deviance, because they repre-
in English because this language is spoken by a vast majority of sent acts considered counternormative and harmful to the organi-
the Filipino population, especially in its educated social strata zations interest by the management (referred to as the dominant
(Bernardo, 2004). Simon Lloyd D. Restubog read the accompany- administrative coalition by Robinson & Bennett, 1997, p. 6). The
ing instructions and ensured anonymity and confidentiality. All organization classified these offenses into two types: major offense
participants were invited to participate, and 215 participants chose and minor offense. An independent person matched survey re-
to participate in the Time 1 survey (response rate 71.67%). At sponses with the frequency of major and minor offenses (we did
Time 2, 20 months after Time 1 data collection, we obtained actual not have access to the exact form of offense). A major offense was
workplace deviant behaviors for those participants who gave con- one that warranted a suspension or dismissal. Examples included
sent for us to access to their personnel records. The result was a negligence resulting in damage of organizational property/
matched sample of 153. To assess response bias, we conducted a equipment/facilities; unauthorized use of office materials, equip-
series of t tests. There were no significant differences across all ment, facilities, or resources for personal or commercial purposes;
demographic characteristics (i.e., age, gender, and tenure) between and tampering with ones own or someone elses log-in record/
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

those employees who participated in the data collection for Time punch card or allowing someone else to tamper with ones log-in
1 and Time 2. Approximately 68% of the sample was female record/punch card. In contrast, a minor offense was one that
(7.8% failed to report their gender). This gender distribution is warranted a sanction of reprimand, verbal admonition, or warning.
congruent with findings of the Asian Development Bank (2004) Examples of minor offenses included willful disrespect of cowork-
that a large proportion of Philippine government employees are ers, superiors, or visitors/clients (e.g., making vicious remarks
women. The largest group of participants (57.2%) was 4150 years intended to offend or annoy the person addressed) during office
old, and average organizational tenure was 21.39 years. To main- hours and/or within the organizational environment; leaving the
tain confidentiality, an independent person matched the informa- workplace during office hours without permission from the super-
tion from the employees personnel records with the survey re- visor; reading a newspaper/magazine/book; watching television;
sponse using control numbers. Neither the organizational and browsing over the Internet during official work hours.
representatives nor the research team had access to both pieces of Control variables. Demographic variables that are likely to
information. influence the study results were measured and controlled for in the
subsequent analysis. Gender was controlled because there is evi-
Measures dence to suggest that men tend to exhibit overt angry reactions
more frequently than do women (Spielberger, 1996). Age and
Psychological contract breach. Perceptions of psychological tenure were controlled because older employees are less likely to
contract breach were assessed with the five-item scale (7-point engage in negative and disruptive behaviors than are younger
Likert scale ranging from 1 Strongly Disagree to 7 Strongly employees. This proposition is supported by the social emotional
Agree) developed by Robinson and Morrison (2000), as it provides selectivity theory, which suggests that aging is often accompanied
an overall assessment of the extent to which the employees by an increased emotional maturity that may lessen the tendency to
psychological contract has been fulfilled. Example items include react negatively to distressing situations (Carstensen, 1992).
Almost all the promises made by my employer during recruitment
have been kept so far and So far my employer has done an
Results and Discussion
excellent job of fulfilling its promises to me. This scale yielded a
reliability coefficient of .74. The means, standard deviations, and zero-order correlations for
Workplace deviance. We assessed workplace deviance by the study variables are displayed in Table 1. Zero-order correla-
drawing on archival personnel records pertaining to the number of tions were all in the expected direction. A check for multivariate
workplace offenses committed by the employees in a period span- outliers using Mahalanobis distance scores (Tabachnick & Fidell,
ning 3 months at Time 2 (20 months after the Time 1 measurement 1996) identified two cases as outliers. Removing these cases did

Table 1
Means, Standard Deviations, Coefficient Alpha, and Intercorrelations of Variables (Study 1)

Variable M SD 1 2 3 4 5 6

1. Gender 0.74 0.44


2. Age 6.49 1.37 .16
3. Tenure 21.38 7.12 .15 .70
4. Psychological contract breach
(Time 1) 3.57 1.03 .13 .18 .11 (.74)
5. Minor offenses (Time 2) 0.76 0.80 .01 .01 .03 .38 na
6. Major offenses (Time 2) 0.40 0.60 .09 .10 .06 .41 .39 na

Note. n 153. Gender is coded as 0 male and 1 female. Age is coded as 1 less than 20 years, 2 2125 years, 3 26 30 years, 4 3135
years, 5 36 40 years, 6 41 45 years, 7 46 50 years, 8 above 50 years. Tenure was measured by an open-ended question: How long have you
been working for this organization (in years)? Alpha reliabilities are along the diagonal. na not available.

p .05. p .001.
CONTRACT BREACH, REVENGE, AND WORKPLACE DEVIANCE 1109

not change the results, so we have retained the cases in the tained an assurance of confidentiality. Furthermore, participants
following analyses. Hierarchical multiple regression analyses were were assured that no one in the organization would see their
performed separately for the two types of offenses (i.e., major and individual survey responses. The survey kit contained two forms:
minor). Results (see Table 2) showed that psychological contract a self-report form and a supervisor rating form. The self-report
breach was positively related to both minor offenses ( .44, p questionnaire assessed the facet measures of the psychological
.001) and major offenses ( .49, p .001), over and above the contract, feelings of violation, and revenge cognitions, and the
effects of the demographic variables. Thus, Hypothesis 1 was supervisor rating form contained the measure of organizational
supported. These findings are consistent with those of previous workplace deviance. We requested that participants ask their su-
studies that reported negative organizational outcomes of psycho- pervisors to complete the enclosed supervisor rating form and
logical contract breach (Deery et al., 2006; Kickul et al., 2001). In return it to the research team in a sealed envelope. To ensure the
addition, the use of archival records and the association of breach integrity of the supervisor ratings, we requested that supervisors
and deviance over time strengthened our confidence in this rela- sign across the flap of the sealed envelope that contained the rating
tionship. As the relationship between the antecedent (breach) and form. To allow matching of the data collected, we requested that
dependent variable (deviance) was demonstrated by Study 1, the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

each participant fill in an anonymous code for the self-survey and


This document is copyrighted by the American Psychological Association or one of its allied publishers.

subsequent studies tested the role of the mediators (violation and supervisory form.
revenge seeking).

Study 2 Measures

Method Unless otherwise specified, all study variables were measured


with a 7-point Likert type scale (1 strongly disagree to 7
Participants and Procedure strongly agree). The variables were measured with previously
established scales. Items were coded such that the higher the score,
Sales personnel and their direct supervisors from a large phar-
the greater the response to the focal constructs.
maceutical organization in the Philippines participated in this
Psychological contract breach. This measure assessed the ex-
study. Surveys were distributed to a total of 350 staff, and 207 staff
tent of breach on transactional and relational dimensions. Partici-
completed the surveys (response rate 59.14%). Sixty-one su-
pants were presented with a list of obligations or commitments
pervisors completed behavioral ratings that resulted in 168
generally made by employers/organizations to their employees
matched surveys. Only the matched surveys were included in the
(Restubog & Bordia, 2006; Robinson & Morrison, 1995) and were
analysis. Supervisors evaluated an average of 2.75 subordinates.
instructed to rate the extent to which each commitment was ful-
Of the 168 participants, 51.8% were males. Almost half (48.2%) of
filled by their employer on a 7-point Likert scale (1 Not fulfilled
the participants were in the 26- to 30-year-old age group. Approx-
imately 82% of the participants had worked for their organization at all to 7 Extremely fulfilled). All items were reverse scored so
for 15 years. they would operate as an index of psychological contract breach.
Survey kits were distributed to staff via the human resource Seven items measured breach in transactional obligations (e.g.,
department and were prefaced by a letter that outlined the objec- high salary, health care benefits, regularity of pay raises), and nine
tives of the research, discussed voluntary participation, and con- items assessed breach in relational obligations (e.g., professional
guidance in carrying out ones job, opportunities for training and
development, support and understanding regarding personal con-
Table 2 cerns). Cronbachs alphas were .86 and .91 for transactional and
Summary of Hierarchical Regression Analysis for Variables relational breach, respectively.
Predicting Archival Records of Workplace Deviant Behaviors Feelings of violation. We used a four-item measure developed
(Study 1) by Robinson and Morrison (2000) to assess the emotional experi-
ence of anger, frustration, and betrayal associated with broken
Minor offenses Major offenses
promises. Example items are I feel that the organization has
(Time 2) (Time 2)
violated the contract between us and I feel betrayed by the
Adjusted Adjusted organization. Alpha reliability for the scale was .85.
Variable R2 R2 R2 R2 Revenge cognitions. Four items taken from Bradfield and
Aquino (1999) assessed thoughts of revenge. A sample item is
Step 1
Gender .00 .14 Im going to get even. The scale yielded a coefficient alpha of
Age .03 .07 .90.
Tenure .03 .00 .02 .00 Workplace deviance. We used the eight-item measure devel-
Step 2 oped by Aquino et al. (1999) to assess workplace deviance directed
Gender .07 .07
Age .09 .06 at the organization. Example items are intentionally arrived late
Tenure .05 .04 for work and took undeserved breaks to avoid work. Supervi-
Psychological sors were requested to provide an assessment of the frequency with
contract breach which their employees demonstrated deviant behaviors (1 Never
(Time 1) .44 .15 .18 .49 .22 .22
to 5 More than 20 times). The reliability coefficient of the
Note. n 153. eight-item scale was .65. Supervisor and peer ratings have previ-

p .01. p .001. ously been used to measure anticitizenship behaviors (Kickul et
1110 BORDIA, RESTUBOG, AND TANG

al., 2001) and organizational retaliatory behaviors (Skarlicki & 0 and 1, and values closer to 1 represent a good fitting model. A
Folger, 1997), respectively. value of .08 or less for RMSEA is indicative of a good fit.
Control variables. We measured demographic characteristics
of age, gender, and tenure. Given that self-report measures of
Measurement Model
feelings of violation, revenge cognitions, and workplace deviant
behaviors are vulnerable to self-report bias, we controlled for the The hypothesized measurement model had a weak fit with the
effects of social desirability and assessed social desirability using observed data, 2(454, N 168) 931, p .001, 2/df 2.05,
the short version of the CrowneMarlowe Social Desirability CFI .84, TLI .82, RMSEA .08. Several items in the breach
Scale. The scale consists of 10 items that describe desirable and measure pertained to common dimensions (three items each on
undesirable behaviors. It serves as a measure of impression man- pay, benefits, and supervision relationships). We created parcels of
agement and can be used to determine the extent to which test items that derived from a common dimension. Parceling leads to a
scores have been influenced by the desire of the individual to better variables-to-sample-size ratio and is justified when it is
present a positive image of his or her character (Crowne & Mar- conducted on theoretically justifiable categories (Little, Cunning-
lowe, 1960). Previous research has revealed strong correlations
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ham, Shahar, & Widaman, 2002). In the measure of transactional


This document is copyrighted by the American Psychological Association or one of its allied publishers.

between the short version and the 33-item version of the scale breach, we combined items related to pay (high salary, pay based
(Strahan & Gerbasi, 1972). In this study, the coefficient alpha was on level of performance, regularity of pay raises) and benefits
.61. Previous studies have found reliability estimates that are (health care benefits, retirement benefits, overall benefits). In the
comparable with our findings; for example, reliability estimates measure of relational breach, we combined items referring to
obtained from male university students, female university students, supervision relationships (feedback about performance, support
and a combination of student and nonstudent samples were .70, and understanding regarding personal concerns, professional guid-
.66, and .59, respectively (Strahan & Gerbasi, 1972). ance in carrying out ones job). We also dropped one item from
workplace deviance that did not load significantly on the latent
construct (called in sick when he/she was not really ill). These
Results
modifications led to a substantial improvement of the measure-
The means, standard deviations, and zero-order correlations for ment model, as indicated by the various fit indices, 2(265, N
the study variables are displayed in Table 3. Zero-order correla- 168) 422.39, p .001, 2/df 1.59, CFI .92, TLI .91,
tions were all in the expected direction. None of the demographic RMSEA .06. The standardized path estimates of the manifest
characteristics were related to the study variables. We undertook a indicators (range .24 .97) were all statistically significant. We
two-step procedure (Anderson & Gerbing, 1988) using AMOS also compared the final measurement model with alternative mod-
(analysis of moment structure) to examine the hypothesized rela- els comprising one, two, three, and four factors. The five-factor
tionships. In the first step, we estimated a full measurement model model had the best fit.
with all the manifest indicators to establish the factor structure of
the variables in the model (i.e., psychological contract breach, Structural Model
feelings of violation, revenge cognitions, and supervisor-rated
deviant behaviors directed toward the organization). The second Consistent with Anderson and Gerbings (1988) recommenda-
step involved testing the hypothesized structural model. We used tions for examining nested models, we compared two competing
several fit indices to assess the adequacy of the model (Byrne, models, a fully mediated structural model and a partially mediated
2001), namely, chi-square, the comparative fit index (CFI), the structural model. The hypothesized mediated model (Model A)
TuckerLewis index (TLI), and the root-mean-square error of had a good fit, 2(271, N 168) 426.96, p .001, 2/df
approximation (RMSEA). Values for the TLI and CFI are between 1.58, CFI .92, TLI .92, RMSEA .06. All the predicted

Table 3
Means, Standard Deviations, Coefficient Alphas, and Intercorrelations of Variables (Study 2)

Variable M SD 1 2 3 4 5 6 7 8 9

1. Gender 0.48 0.50


2. Age 2.77 0.72 .18
3. Tenure 2.10 0.40 .09 .28
4. Social desirability 7.77 1.92 .03 .02 .13 (.61)
5. Transactional breach 3.96 0.78 .15 .05 .13 .24 (.86)
6. Relational breach 3.73 0.70 .11 .12 .09 .24 .69 (.91)
7. Feelings of violation 2.20 1.08 .03 .07 .10 .23 .21 .27 (.85)
8. Revenge cognitions 2.05 1.24 .02 .03 .01 .13 .12 .16 .36 (.90)
9. Organizational deviance 1.17 0.22 .06 .02 .02 .17 .09 .02 .25 .38 (.65)

Note. n 168. Gender is coded as 0 male and 1 female. Age is coded as 1 less than 20 years, 2 2125 years, 3 26 30 years, 4 3135
years, 5 36 40 years, 6 41 45 years, 7 46 50 years, 8 above 50 years. Tenure is coded as 1 less than 1 year, 2 15 years, 3 6 10 years,
4 1115 years, 5 16 20 years, 6 2125 years, 7 26 30 years. Alpha reliabilities are along the diagonal.

p .05. p .01. p .001.
CONTRACT BREACH, REVENGE, AND WORKPLACE DEVIANCE 1111

paths were significant at p .001, with the exception of the path the relational component of the psychological contract was more
linking transactional breach and feelings of violation. To examine frequently mentioned than was the transactional component and
the partially mediated structural model (Model B), we added four therefore seemed more important. Similarly, a recent meta-
additional paths representing direct effects: (a) a path linking analysis (Zhao et al., 2007) revealed that the relationship between
relational breach and revenge cognitions; (b) a path linking trans- relational breach and job satisfaction, turnover intentions, and
actional breach and organizational deviance; (c) a path linking organizational citizenship behaviors was stronger than was the
relational breach and organizational deviance; and (d) a path relationship between transactional breach and these variables.
linking feelings of violation and organizational deviance. Results Revenge cognitions mediated the relationship between feelings
showed that these additional paths were nonsignificant. The fit of violation and organizational deviance. These findings imply that
indices of the partially mediated model (Model B) were 2(269, when promises to employees are broken, the employees are likely
N 168) 423.53, p .001, 2/df 1.58, CFI .92, TLI .91, to feel violated and, in turn, to contemplate revenge and carry it out
RMSEA .06. Results of the chi-square difference test suggest in the form of organizational deviance. Thus, our findings dem-
that this model did not substantively improve the fit compared with onstrate the motivational underpinnings of organizational devi-
Model A, diff
2
(2) 3.43, ns. Thus, the partially mediated struc- ance. However, we did not test whether the revenge motivation
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

tural model was not supported. Finally, we undertook the Sobel leads to workplace deviance directed at coworkers (interpersonal
test (Sobel, 1982) to formally assess the significance of the pro- deviance). Study 3 extended the model to include interpersonal
posed mediators. Results revealed that all indirect paths were deviance and also tested the direct and moderating role of self-
significant (i.e., z 1.99, p .05, for the path linking relational control.
breach and revenge cognitions through violation; z 2.75, p Although our hypothesized model was supported, note the ab-
.01, for the path linking feelings of violation and organizational sence of a correlation between breach and organizational deviance.
deviance through revenge cognitions). After the removal of the This absence could result from distal mediation (i.e., the predictor
nonsignificant paths, the model had adequate overall fit, 2(271, and the criterion variables are conceptually or temporally far apart
N 168) 426.96, p .001, 2/df 1.58, CFI .92, TLI .92, and are likely to be only weakly correlated; Shrout & Bolger,
RMSEA .06. 2002), as there is a long chain of mediators between breach and
We tested another alternative model (Model C), in which social deviance. Indeed, similar absence of direct relationship has been
desirability and demographic variables were included in the struc- noted in other models that involve mediators between breach and
tural model. The overall model had an average fit, 2(319, N outcomes, such as family undermining (Hoobler & Brass, 2006)
168) 509.65, p .001, 2/df 1.60, CFI .91, TLI .90, and absenteeism (Deery et al., 2006). A second possibility is that
RMSEA .06. Results of the chi-square difference test suggested facet-based measures of breach, such as the ones used in this study,
that the hypothesized structural model (Model A) had better fit have weaker relationships with outcomes than do global measures
than did this alternative model (Model C), diff
2
(48) 82.69, p of breach (Zhao et al., 2007). The facet measures may not com-
.001. Figure 1 presents the final model (Model A) with the stan- prehensively capture the experience of breach for every respon-
dardized estimates. dent. For some respondents, the facets may not be highly relevant,
and, for others, significant facets may be underrepresented (Cross-
Discussion ley, Bennett, Jex, & Burnfield, 2007). Therefore, global measures
are likely to be better at capturing the respondents assessments of
Overall, the results provide support for the hypotheses. Feelings breach. In Study 3, reported below, we used a global measure of
of violation mediated the relationship between relational contract breach.
breach and revenge cognitions. It is interesting that transactional
breach was not related to feelings of violation. These findings are
Study 3
consistent with previous research that has found stronger effects
for breach of relational contract compared with transactional con- In this study, we extended the previous study by including both
tract for some outcome variables (Restubog & Bordia, 2006; organizational and interpersonal deviance. Given that the breach of
Robinson & Morrison, 1995; Turnley et al., 2003). For example, in psychological contract is an organization-level transgression, we
a qualitative study, Herriot, Manning, and Kidd (1997) found that might expect that revenge would be directed at the organization in

Transactional
breach

.39*** .51***

.75***
Feelings of Revenge Organizational
violation cognitions deviance
.30***

Relational
breach


Figure 1. Final structural model in Study 2 (n 168). p .001.
1112 BORDIA, RESTUBOG, AND TANG

the form of organizational deviance. However, we argued that both Feelings of violation. As in Study 2, feelings of violation were
the organization and coworkers could be targeted. There are two assessed with the four-item scale ( .95) developed by Robinson
lines of argument to support this prediction. First, deviance could and Morrison (2000).
be directed at other coworkers who are thought to be aligned to the Revenge cognitions. As in Study 2, revenge cognitions (
management and therefore represent the transgressor; harming .89) were assessed with the measure taken from Bradfield and
coworkers could be perceived as a way of getting back at the Aquino (1999).
organization. Second, there is evidence to suggest that aggression Self-control. Self-control was assessed with five items devel-
can be displaced from the original foci to other targets (Marcus- oped by Scott (1965).2 In this study, the scale for self-control
Newhall, Pederson, Carlson, & Miller, 2000), particularly when yielded a reliability coefficient of .87.
the original transgressor is intangible, not available, or more pow- Workplace deviance. Workplace deviance was assessed along
erful and likely to be feared. In the context of psychological two dimensions on the basis of previous conceptualizations (orga-
contract breach, it might not be clear who is responsible for the nizational workplace deviance and interpersonal workplace devi-
original promises or for the subsequent breach (Hoobler & Brass, ance; Robinson & Bennett, 1995). Nine items (i.e., five items
2006). Moreover, an employee would be fearful of blatantly tar- measuring organizational deviance and four items assessing inter-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

personal deviance)3 with the highest factor loadings on their re-


This document is copyrighted by the American Psychological Association or one of its allied publishers.

geting the organization with acts of aggression. In such circum-


stances, coworkers and subordinates might be more available and spective factors were taken from Aquino et al. (1999). As in Study
easier targets. Therefore, we expected that our model would apply 2, immediate supervisors were instructed to rate the frequency with
to both organizational and interpersonal deviance. which their employees engaged in workplace deviant behaviors.
Study 3 built on Study 2 in two additional ways. First, we Example items for interpersonal deviance include This employee
examined the role of self-control as a moderator of the relationship made an obscene comment or gesture at a coworker and This
between revenge cognitions and workplace deviant behaviors. employee teased a coworker in front of other employees. Cron-
Second, in Study 2 we utilized a facet-based measure of psycho- bachs alphas for scales for organizational and interpersonal devi-
logical contract breach. In Study 3, we used a global assessment of ance were .87 and .90, respectively.
breach, in which participants were asked to provide an overall Control variables. As in Studies 1 and 2, we measured demo-
evaluation of the extent to which their organization had fulfilled its graphic characteristics of age, gender, and tenure. In addition,
promises to them. social desirability bias was assessed with the short version of the
CrowneMarlowe Social Desirability Scale ( .77; Strahan &
Gerbasi, 1972).
Method

Participants and Procedure Results

A survey was administered to 250 employees of a large com- The means, standard deviations, intercorrelations, and reliability
mercial bank in the Philippines, and 204 employees completed the coefficients for all measures are shown in Table 4. Zero-order
survey (response rate 81.6%). Fifty-nine supervisors completed correlations were all moderate in size. The pattern of correlations
behavioral ratings that resulted in 187 matched surveys. On aver- was in the predicted direction. Age and tenure were negatively
age, supervisors evaluated 2.79 subordinates. Of the participants, correlated with revenge cognitions. We include the demographic
58% were female, 80% were between 21 and 30 years of age, and variables and social desirability in the analysis reported below.
87% had been working for their organization for 15 years. Before testing the hypothesized relationships, we conducted a
Items in the survey were prepared in English. Each participant confirmatory factor analysis on the study variables (psychological
completed the self-reported questionnaire and passed on the be- contract breach, violation, revenge cognitions, interpersonal devi-
havioral rating scale to his or her immediate supervisor. Partici- ance, organizational deviance, and self-control). Initial analysis of
pation in the survey was voluntary, and all participants were our proposed model in this sample suggested that our hypothesized
assured of the confidentiality of their responses. Completed survey model had a moderate fit with the observed data, 2(284, N
forms were returned directly to Robert L. Tang in sealed and
preaddressed envelopes. The self- and supervisor ratings were 2
matched with a respondent-generated code (placed on both the We examined whether the shortened 5-item measure of self-control
used in the present study was equivalent with the full measure. To achieve
self-report and the supervisor questionnaire).
this objective, we administered the 10-item scale (full scale) to an inde-
pendent sample of 160 government employees in the Philippines. Bivariate
correlations indicated that the shortened and complete versions were highly
Measures
correlated (r .89, p .001).
3
Participants responded using a 7-point Likert scale (1 To examine whether the shortened measure of organizational and
interpersonal deviance dimensions used in the present study were equiva-
Strongly Disagree to 7 Strongly Agree). The study variables
lent with the full scale, we administered the (complete) 16-item scale used
were measured with previously established scales. However, due
by Aquino et al. (1999) to an independent sample of 267 full-time em-
to survey length constraints posed by the participating organiza- ployees enrolled in a part-time MBA program in the Philippines. Bivariate
tion, we had to shorten some of the measures. correlations suggest that the shortened and complete versions were highly
Psychological contract breach. We used the five-item global correlated (short and long versions of organizational deviance scales, r
psychological contract breach measure developed by Robinson .93, p .001; short and long versions of the interpersonal deviance scales,
and Morrison (2000) in this study ( .81). r .90, p .001).
CONTRACT BREACH, REVENGE, AND WORKPLACE DEVIANCE 1113

Table 4
Means, Standard Deviations, Coefficient Alphas, and Intercorrelations of Variables (Study 3)

Variable M SD 1 2 3 4 5 6 7 8 9 10

1. Gender 0.58 0.50


2. Age 2.78 1.21 .20
3. Tenure 1.79 1.06 .26 .58
4. Social desirability 5.39 2.69 .01 .14 .08 (.77)
5. Psychological contract breach 3.61 1.32 .03 .00 .02 .37 (.81)
6. Feelings of violation 2.76 1.58 .06 .02 .05 .31 .53 (.95)
7. Revenge cognitions 2.82 1.65 .12 .23 .15 .30 .20 .42 (.89)
8. Self-control 4.48 1.28 .03 .07 .02 .44 .29 .15 .21 (.87)
9. Organizational deviance 1.54 0.71 .03 .00 .00 .10 .19 .19 .21 .22 (.87)
10. Interpersonal deviance 2.22 1.46 .06 .10 .04 .51 .16 .31 .43 .35 .19 (.90)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Note. n 187. Gender is coded as 0 male and 1 female. Age is coded as 1 less than 20 years, 2 2125 years, 3 26 30 years, 4 3135
This document is copyrighted by the American Psychological Association or one of its allied publishers.

years, 5 36 40 years, 6 41 45 years, 7 46 50 years, 8 above 50 years. Tenure is coded as 1 less than 1 year, 2 15 years, 3 6 10 years,
4 1115 years, 5 16 20 years, 6 2125 years, 7 26 30 years. Alpha reliabilities are along the diagonal.

p .05. p .01. p .001.

187) 693.17, p .001, 2/df 2.44, CFI .89, TLI .87, Figure 3 presents the final path model with the standardized path
RMSEA .088. One item from the psychological contract breach estimates. The final path model had a very good fit, as indicated by
measure (My employer has broken many of its promises to me the various fit indices, 2(23, N 187) 32.46, ns, 2/df 1.41,
even though Ive upheld my side of the deal) had highly corre- CFI .97, TLI .95, RMSEA .047. All the paths were
lated error terms with the latent factor feelings of violation and statistically significant.
with the item I have not received everything promised to me in To examine a partially mediated path model, we added three
exchange for my contributions. Removal of this item led to a direct paths: (a) a path linking psychological contract breach and
substantial improvement in fit, 2(260, N 187) 515.15, p revenge cognitions, (b) a path linking feelings of violation and
.001, 2/df 1.98, CFI .93, TLI .92, RMSEA .073. All the organizational deviance, and (c) a path linking feelings of violation
manifest indicators were statistically significant. We compared this and interpersonal deviance. However, all these direct paths were
six-factor measurement model to eight different measurement nonsignificant. A Sobel test showed that the indirect effect of
models, each with a different configuration of factors. The six- psychological contract breach to revenge cognitions via feelings of
factor measurement model had the best fit. violation (z 4.39, p .001) was significant. Similarly, the
We tested the predicted model using path analysis that included indirect effects of feelings of violation to organizational deviance
a multiplicative term to test for the interaction between self-control (z 2.15, p .05) and interpersonal deviance (z 3.7, p .01)
and revenge cognitions (see Figure 2). Because age and social via revenge cognitions were significant. Overall, we found support
desirability were related to some of the study variables, we in- for the mediated model.
cluded these variables in the path model. The path between age and Further examination of the path model showed that self-control
revenge cognitions was significant and was therefore retained. and the product term (Revenge Cognitions Self-Control) signif-
Similarly, social desirability was related to breach, revenge cog- icantly predicted both types of deviant behaviors. To aid interpre-
nitions, and interpersonal deviance; these paths were retained. tation, we plotted the slopes for high (one standard deviation above

Social .45***
Age Self - control
desirability
- .15*
- .38***
-.17*
Interpersonal
- .38*** - .22*
.27*** deviance

Psychological
Feelings of Revenge
contract
violation cognitions
breach
.15*
.53*** .37*** - .21**

- .14*
Organizational
Self - control
deviance
x
Revenge cognitions - .19**
Interaction Term


Figure 2. Final path model in Study 3 (n 187). p .05. p .01. p .001.
1114 BORDIA, RESTUBOG, AND TANG

2.5 Theoretical Implications


Low self-control High self-control
Supervisor-rated organizational

This article extends the research on the harmful consequences of


psychological contract breach in several important ways. First, we
2.0
provide strong empirical evidence for the relationship between
deviance

breach and deviant behaviors, including a longitudinal relationship


and multiple operationalizations of deviance (employee records
1.5 and supervisor ratings). Second, by relating breach and violation to
revenge cognitions, we have demonstrated that breach results in
motivated deviant acts aimed at harming the organization. There is
1.0 a great deal of evidence for the harmful consequences of breach
Low High (Kickul et al., 2001; Morrison & Robinson, 1997; Restubog et al.,
Revenge cognitions 2007; Turnley et al., 2003), but the motivational processes that link
contract breach and negative outcomes are rarely explored. To our
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Figure 3. Interactive relationship between revenge cognitions and self- knowledge, this is the first study to measure and test the role of
This document is copyrighted by the American Psychological Association or one of its allied publishers.

control in predicting supervisor-rated organizational deviance. revenge seeking as a consequence of contract breach.
Although there is a great deal of research on the consequences
of contract breach, there is limited work on violation as an affec-
the mean) and low (one standard deviation below the mean) levels tive response to breach and as a mediator between breach and
of self-control for each type of deviant behavior. With organiza- outcomes. Our results demonstrate that a cognitive assessment of
tional deviance as the outcome variable, the slope for high self- breach results in feelings of violation. Violation in turn leads to
control was not significant, t(177) .20, ns. However, the slope thoughts of revenge and workplace deviance. These findings sup-
for low self-control was statistically significant, t(177) 3.30, p port the theoretical propositions of Morrison and Robinson (1997)
.001. Figure 3 shows that there is a stronger positive association and can be interpreted in light of recent work on emotions that has
between revenge cognitions and organizational deviance for em- argued that emotions mediate the relationship between workplace
ployees with low levels of self-control than for employees with events and resulting attitudes and behaviors (Judge, Scott, & Ilies,
high levels of self-control. As with organizational deviance, the 2006).
simple slope of revenge cognitions on interpersonal deviance was The identification of revenge as an antecedent of workplace
significant for low self-control, t(177) 4.52, p .001, but not deviance makes an important contribution to the wider literature
for high self-control, t(177) 1.29, ns. Figure 4 suggests that there on negative, retaliatory, and counterproductive behaviors in the
is a stronger positive relationship between revenge cognitions and workplace. Although research has identified workplace anteced-
interpersonal deviance for employees with low levels of self- ents to negative behaviors (e.g., injustice), many of the motiva-
control than for employees with high levels of self-control. Over- tional underpinnings, such as reciprocity (the need to retaliate,
all, Hypotheses 4 and 5 were supported. seek revenge, or restore equity), are left unmeasured. There is a
range of complex motivations underlying ostensibly deviant be-
Discussion haviors (Griffin & Lopez, 2005), and a behavior that appears
counternormative and detrimental to the organization may actually
Once again, these results supported predicted relationships. Us- be driven by good intentions (Morrison, 2006). This article has
ing a global measure of contract breach, we found that violation drawn attention to one motivationrevenge underlying deviant
mediated the relationship between breach and revenge cognitions behaviors. We recommend that future empirical research that links
(Hypothesis 2). Also, revenge cognitions mediated the relationship workplace antecedents (e.g., injustice, breach of psychological
between violation and interpersonal and organizational deviance contract) and consequences (e.g., job performance, citizenship,
(Hypothesis 3). As predicted, self-control had a role that was direct
(Hypothesis 4) and moderating (Hypothesis 5). People high in
self-control were less likely to engage in organizational and inter-
5.0
personal deviant acts. Also, high self-control seems to have kept
Low self-control High self-control
Supervisor-rated interpersonal

people from acting on the revenge cognitions. This finding adds to 4.5
the growing literature on the role of individual differences, and
self-control in particular, in minimizing harmful behavior in an 4.0
deviance

organizational context (Douglas & Martinko, 2001).


3.5

General Discussion 3.0

Our aim in this article was to study the cognitive, affective, and 2.5
motivational antecedents and dispositional buffers of workplace
2.0
deviance. In doing so, we strove to connect various streams of
Low High
research in organizational behavior: psychological contracts, feel- Revenge cognitions
ings of violation, workplace revenge, self-control, and workplace
deviance. Our proposed model and the associated hypotheses were Figure 4. Interactive relationship between revenge cognitions and self-
largely supported. control in predicting supervisor-rated interpersonal deviance.
CONTRACT BREACH, REVENGE, AND WORKPLACE DEVIANCE 1115

deviant or withdrawal behaviors) incorporate the motivational ventions can be targeted. At the startto prevent this process from
intent driving the behavioral consequences. initiating organizations should strive to fulfill the psychological
An important line of future research would be on constraints that contract. However, given the competitive pressures confronting
limit the likelihood of revenge cognitions being acted out as most organizations and the resulting incidence of organizational
deviant behaviors. Results from Study 3 showed that self-control change, it may not be possible to fulfill every promise made to a
acted as a buffer between revenge cognitions and both organiza- job incumbent (Robinson & Rousseau, 1994; Rousseau, 1995). In
tional and interpersonal deviance. Other personality variables, such circumstances, organizations should strive to reduce the feel-
such as negative reciprocity (Eisenberger et al., 2004) and trait ings of violation by providing adequate explanation for the causes
anger (Spielberger, 1996), may strengthen the association between of breach and redressing the loss to the employee by other means
revenge and deviance. In addition to personality variables, contex- (Morrison & Robinson, 1997). Further, organizations should strive
tual variables, such as formal structures, informal norms, or com- to reduce the conversion of feelings of violation into motivation to
mitment to other referents in the organization (e.g., supervisor or seek revenge by providing other means of expressing the frustra-
work team), may constrain the enactment of workplace deviance tion and anger. For example, the expression and disclosure of these
(Robinson & Bennett, 1997). feelings to a supportive supervisor may provide a more construc-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

By linking revenge with workplace deviance, we have taken the tive coping mechanism (Grandey, 2000). Revenge can be a coun-
stance that revenge has negative consequences for organizational terproductive way of coping with perceived transgressions, as it
interests. However, we acknowledge that revenge is not always can lead to further retaliation and an escalation of conflict (Brad-
bad and indeed may have constructive outcomes for individuals field & Aquino, 1999). The moderating role of self-control sug-
and workplaces (Tripp & Bies, 1997). For example, revenge may gests that better self-regulation of emotional and cognitive reac-
stop abusive or bullying behaviors, restore justice, and boost tions can help in curbing the devastating consequences of revenge
self-esteem of the revenge seeker (Tripp & Bies, 1997). Moreover, seeking. Thus, self-control could be a desirable trait in the context
revenge is not the only response to transgressions. People do of employee selection (OLeary-Kelly et al., 1996). In addition,
choose to forgive. There is evidence to suggest that these responses training in regulation of emotions and negative ruminations may
are dynamic (McCullough, Fincham, & Tsang, 2003). For exam- help employees and organizations better manage the consequences
ple, a person might contemplate revenge soon after perceived of breach.
victimization but forgive the transgressor over time. Longitudinal
research will allow exploration of the temporal progression of References
revenge and forgiveness in the workplace.
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in
practice: A review and recommended two-step approach. Psychological
Limitations Bulletin, 103, 411 423.
A major limitation of Studies 2 and 3 is the cross-sectional Aquino, K., Lewis, M. U., & Bradfield, M. (1999). Justice constructs,
design. Although we have taken steps to minimize common negative affectivity, and employee deviance: A proposed model and
empirical test. Journal of Organizational Behavior, 20, 10731091.
method variance (by the use of supervisor reports and by control-
Aquino, K., Tripp, T. M., & Bies, R. J. (2001). How employees respond to
ling for social desirability), we cannot draw causal inferences from personal offense: The effects of blame attribution, victim status, and
data collected via a cross-sectional design. Measures of social offender status on revenge and reconciliation in the workplace. Journal
desirability and organizational deviance had low internal consis- of Applied Psychology, 86, 5259.
tency in Study 2. Also, the mean levels of revenge cognitions Asian Development Bank. (2004). ADB country gender assessment: Phil-
(Study 2, M 2.05, Study 3, M 2.82, on 7-point scales) and ippines. Manila, Philippines: Author.
workplace deviance (Study 2, M organizational deviance 1.17; Bennett, R. J., & Robinson, S. L. (2000). Development of a measure of
Study 3, M organizational deviance 1.54; M interpersonal de- workplace deviance. Journal of Applied Psychology, 85, 349 360.
viance 2.22, on 7-point scales) are low. However, previous Bernardo, A. B. I. (2004). McKinleys questionable bequest: Over 100
literature has reported similar means for revenge cognition (on years of English in Philippine education. World Englishes, 23, 1731.
Bies, R. J., & Tripp, T. M. (2001). A passion for justice: The rationality and
5-point scales, 1.85 in Bradfield & Aquino, 1999; 1.21 in Aquino
morality of revenge. In R. Cropanzano (Ed.), Justice in the workplace:
et al., 2001) and for workplace deviance (M organizational devi- From theory to practice (Vol. 2, pp. 197208). Mahwah, NJ: Erlbaum.
ance 1.30 in Aquino et al., 1999, and 1.81 in Dunlop & Lee, Bies, R. J., Tripp, T. M., & Kramer, R. M. (1996). At the breaking point:
2004; M interpersonal deviance 1.42 in Aquino et al., 1999, and Cognitive and social dynamics of revenge in organizations. In R. A.
1.87 in Dunlop & Lee, 2004). These cognitions and behaviors may Giacalone & J. Greenberg (Eds.), Antisocial behavior in organizations
have low incidence or low visibility, and respondents are likely to (pp. 18 36). Thousand Oaks, CA: Sage.
be reluctant to report them. However, their impact is considerable, Bradfield, M., & Aquino, K. (1999). The effects of blame attributions and
and they therefore require continued research attention. offender likeableness on revenge and forgiveness in the workplace.
Journal of Management, 25, 607 631.
Brown, S. P., Westbrook, R. A., & Challagalla, G. (2005). Good cope, bad
Practical Implications cope: Adaptive and maladaptive coping strategies following a critical
negative work event. Journal of Applied Psychology, 90, 792798.
The findings of this research have practical implications for the Byrne, B. M. (2001). Structural equation modeling with AMOS: Basic
management of workplace deviance as well as for employee psy- concepts, applications, and programming. Mahwah, NJ: Erlbaum.
chological contracts. We have identified a chain of cognitions and Carstensen, L. L. (1992). Social and emotional patterns in adulthood:
emotions that combine to provide the motivational intent for Support for socio-emotional selectivity theory. Psychology and Aging, 7,
deviance. This sequence suggests various points at which inter- 331338.
1116 BORDIA, RESTUBOG, AND TANG

Colbert, A. E., Mount, M. K., Harter, J. K., Witt, L. A., & Barrick, M. R. Displaced aggression is alive and well: A meta-analytic review. Journal
(2004). Interactive effects of personality and perceptions of the work of Personality and Social Psychology, 78, 670 689.
situation on workplace deviance. Journal of Applied Psychology, 89, McCullough, M. E., Fincham, F. D., & Tsang, J. (2003). Forgiveness,
599 609. forbearance, and time: The temporal unfolding of transgression-related
Coyle-Shapiro, J., & Conway, N. (2004). The employment relationship interpersonal motivations. Journal of Personality and Social Psychol-
through the lens of social exchange. In J. A. M. Coyle-Shapiro, L. M. ogy, 84, 540 557.
Shore, M. S. Taylor, & L. E. Tetrick (Eds.), The employment relation- Morrison, E. W. (2006). Doing the job well: An investigation on pro-social
ship: Examining psychological and contextual perspectives (pp. 528). rule-breaking. Journal of Management, 32, 528.
New York: Oxford University Press. Morrison, E. W., & Robinson, S. L. (1997). When employees feel be-
Crossley, C. D., Bennett, R. J., Jex, S. M., & Burnfield, J. L. (2007). trayed: A model of how psychological contract violation develops.
Development of a global measure of job embeddedness and integration Academy of Management Review, 22, 226 256.
into a traditional model of voluntary turnover. Journal of Applied OLeary-Kelly, A. M., Griffin, R. W., & Glew, D. J. (1996). Organization-
Psychology, 92, 10311042. motivated aggression: A research framework. Academy of Management
Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability Review, 21, 225253.
independent of psychopathology. Journal of Consulting Psychology, 24,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Raja, U., Johns, G., & Ntalianis, F. (2004). The impact of personality on
349 354.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

psychological contracts. Academy of Management Journal, 47, 350


Deery, S. J., Iverson, R. D., & Walsh, J. T. (2006). Toward a better 367.
understanding of psychological contract breach: A study of customer Restubog, S. L. D., & Bordia, P. (2006). Workplace familism and psycho-
service employees. Journal of Applied Psychology, 91, 166 175. logical contract breach in the Philippines. Applied Psychology: An
Douglas, S. C., & Martinko, M. J. (2001). Exploring the role of individual International Review, 55, 563585.
differences in the prediction of workplace aggression. Journal of Applied Restubog, S. L. D., Bordia, P., & Tang, R. L. (2006). Effects of psycho-
Psychology, 86, 547559. logical contract breach on performance of IT employees: The mediating
Dunlop, P. D., & Lee, K. (2004). Workplace deviance, organizational role of affective commitment. Journal of Occupational and Organiza-
citizenship behavior, and business unit performance: The bad apples do tional Psychology, 79, 299 306.
spoil the whole barrel. Journal of Organizational Behavior, 25, 67 80. Restubog, S. L. D., Bordia, P., & Tang, R. L. (2007). Behavioral outcomes
Eisenberger, R., Lynch, P., Aselage, J., & Rohdieck, S. (2004). Who takes
of psychological contract breach in a non-western culture: The moder-
the most revenge? Individual differences in negative reciprocity norm
ating role of equity sensitivity. British Journal of Management, 18,
endorsement. Personality and Social Psychology Bulletin, 30, 789 799.
376 386.
Fox, S., & Spector, P. E. (2005). Counterproductive work behavior:
Robinson, S. L. (1996). Trust and breach of psychological contract. Ad-
Investigations of actors and targets. Washington, DC: American Psy-
ministrative Science Quarterly, 41, 574 599.
chological Association.
Robinson, S. L., & Bennett, R. J. (1995). A typology of deviant workplace
Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime.
behaviors: A multidimensional scaling study. Academy of Management
Stanford, CA: Stanford University Press.
Journal, 38, 555572.
Grandey, A. (2000). Emotional regulation in the workplace: A new way to
Robinson, S. L., & Bennett, R. J. (1997). Workplace deviance: Its defini-
conceptualize emotional labor. Journal of Occupational Health Psychol-
tion, its manifestations, and its causes. In R. J. Lewicki & R. J. Bies
ogy, 5, 95110.
(Eds.), Research on negotiation in organizations (pp. 327). Greenwich,
Greenberg, J. (1990). Employee theft as a reaction to underpayment ineq-
CT: JAI.
uity: The hidden cost of pay cuts. Journal of Applied Psychology, 75,
Robinson, S. L., & Morrison, E. W. (1995). Psychological contracts and
561568.
Griffin, R. W., & Lopez, Y. P. (2005). Bad behavior in organizations: A OCB: The effect of unfulfilled obligations on civic virtue behavior.
review and typology for future research. Journal of Management, 31, Journal of Organizational Behavior, 16, 289 298.
988 1105. Robinson, S. L., & Morrison, E. W. (2000). The development of psycho-
Hepworth, W., & Towler, A. (2004). The effects of individual differences logical contract breach and violation: A longitudinal study. Journal of
and charistmatic leadership on workplace aggression. Journal of Occu- Organizational Behavior, 21, 525546.
pational Health Psychology, 9, 176 185. Robinson, S. L., & Rousseau, D. M. (1994). Violating the psychological
Herriot, P., Manning, W. E., & Kidd, J. M. (1997). The content of the contract: Not the exception but the norm. Journal of Organizational
psychological contract. British Journal of Management, 8, 151162. Behavior, 15, 245259.
Hoobler, J. M., & Brass, D. J. (2006). Abusive supervision and family Rousseau, D. M. (1995). Psychological contracts in organizations: Un-
undermining as displaced aggression. Journal of Applied Psychology, derstanding written and unwritten agreements. Thousand Oaks, CA:
91, 11251133. Sage.
Judge, T. A., Scott, B. A., & Ilies, R. (2006). Hostility, job attitudes, and Sarchione, C. D., Cuttler, M. J., Muchinsky, P. M., & Nelson-Gray, R. O.
workplace deviance: Test of a multilevel model. Journal of Applied (1998). Prediction of dysfunctional job behaviors among law enforce-
Psychology, 91, 126 138. ment officers. Journal of Applied Psychology, 83, 904 912.
Kickul, J., Neuman, G., Parker, C., & Finkl, J. (2001). Settling the score: Scott, W. A. (1965). Values and organizations: A study of fraternities and
The role of organizational justice in the relationship between psycho- sororities. Chicago: Rand-McNally.
logical contract breach and anti-citizenship behavior. Employee Respon- Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-
sibilities and Rights Journal, 13, 7793. experimental studies: New procedures and recommendations. Psycho-
Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). logical Methods, 7, 422 445.
To parcel or not to parcel: Exploring the question, weighing the merits. Skarlicki, D. P., & Folger, R. (1997). Retaliation in the workplace: The
Structural Equation Modeling, 9, 151173. roles of distributive, procedural, and interactional justice. Journal of
Marcus, B., & Schuler, H. (2004). Antecedents of counterproductive be- Applied Psychology, 82, 434 443.
havior at work: A general perspective. Journal of Applied Psychology, Skarlicki, D. P., Folger, R., & Tesluk, P. (1999). Personality as a moderator
89, 647 660. in the relationship between fairness and retaliation. Academy of Man-
Marcus-Newhall, A., Pederson, W., Carlson, M., & Miller, N. (2000). agement Journal, 42, 100 108.
CONTRACT BREACH, REVENGE, AND WORKPLACE DEVIANCE 1117

Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in The impact of psychological contract fulfillment on the performance of
structural equation models. Sociological Methodology, 13, 290 312. in-role and organizational citizenship behaviors. Journal of Manage-
Spielberger, C. D. (1996). StateTrait Anger Expression Inventory, re- ment, 29, 187206.
search edition: Professional manual. Odessa, FL: Psychological Assess- Weiss, H. M., & Cropanzano, R. (1996). Affective events theory: A
ment Resources. theoretical discussion of the structure, causes, and consequences of
Strahan, R., & Gerbasi, K. C. (1972). Short, homogeneous versions of the affective experiences at work. Research in Organizational Behavior, 18,
MarloweCrowne Social Desirability Scale. Journal of Clinical Psy- 174.
chology, 28, 191193. Zhao, H., Wayne, S. J., Glibkowski, B. C., & Bravo, J. (2007). The impact
Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics (3rd of psychological contract breach on work-related outcomes: A meta-
ed.). New York: Harper Collins College. analysis. Personnel Psychology, 60, 647 680.
Tripp, T. M., & Bies, R. J. (1997). Whats good about revenge: The
avengers perspective. In R. J. Lewicki, R. J. Bies, & B. H. Sheppard
(Eds.), Research on negotiation in organizations (Vol. 6, pp. 145160). Received June 20, 2006
Greenwich, CT: JAI. Revision received March 12, 2008
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Turnley, W. H., Bolino, M. C., Lester, S. W., & Bloodgood, J. M. (2003). Accepted March 17, 2008
This document is copyrighted by the American Psychological Association or one of its allied publishers.

E-Mail Notification of Your Latest Issue Online!


Would you like to know when the next issue of your favorite APA journal will be available
online? This service is now available to you. Sign up at http://notify.apa.org/ and you will be
notified by e-mail when issues of interest to you become available!

You might also like