Are Counterproductive Work Behavior and Withdrawal Empirically Distinct? A Meta-Analytic Investigation

Abstract

The purpose of this meta-analysis was to clarify the current understanding of the relationship between counterproductive work behavior (CWB) and withdrawal. First, we articulate theoretical and conceptual reasons for the confusion on important issues, such as their conceptualization, labeling, and measurement. Second, we conduct a meta-analysis between current CWB and withdrawal measures. We found that, as measured, CWB and withdrawal are strongly related and have patterns of nomological relationships with common correlates that are nearly identical. The relationship between organizational-target CWB and withdrawal is particularly strong. The results suggest that withdrawal may be best represented as a facet in the hierarchical model of CWB, perhaps even as a facet of organizational-target CWB. We also discuss important avenues and needs for future research.

Keywords

counterproductive work behavior withdrawal behavior meta-analysis construct validity

Introduction

Understanding employee negative work behavior has been a long-standing focus of organizational researchers and practitioners alike as a result of the costs to organizations and employees. Most of the current scholarship pertaining to negative work behavior occurs under the counterproductive work behavior (CWB) or workplace deviance labels, which both refer to employee behavior that violates organizational interests and norms and harms the organization itself and/or coworkers or supervisors (Robinson & Bennett, 1995). Indeed, the current popularity of CWB research has at times been attributed to Robinson and Bennett’s (R. J. Bennett & Robinson, 2000) taxonomy and operationalization of CWB, which contributed clarity to both the definition and measurement of employee deviance (e.g., Berry, Ones, & Sackett, 2007).

Importantly, negative work behavior also has a rich tradition of being studied using the withdrawal behavior label, which is used to represent employees’ avoidance of or disengagement from the work environment, tasks, or the organization (e.g., Hanisch & Hulin, 1991; Hulin, 1991). However, withdrawal research has decreased over the past decade, and Harrison and Newman recently speculated that this is because “withdrawal behavior has been subsumed to a large extent under the concept of CWB” (2013: 283). This suggests that CWB and withdrawal may be assumed to refer to redundant classes of employee behaviors, whereby only one label (i.e., CWB) is required. Although reasonable, this assumption is premature as it has yet to be formally tested and, furthermore, critical questions regarding the CWB-withdrawal relationship remain. Notably, it is unknown whether the empirical relationship between CWB and withdrawal supports them as distinct constructs.

The purpose of the present study is to contribute clarity to the understanding of CWB and withdrawal, particularly regarding whether the magnitude of their empirical relationship suggests that the constructs are distinct. First, we conduct a comprehensive meta-analysis of the bivariate CWB-withdrawal relationship to determine their degree of empirical overlap and distinctness. Second, we meta-analyze and then compare the relationships CWB and withdrawal have with variables in their theorized nomological networks—composed of dispositions, organizational characteristics and perceptions, job attitudes, and work performance behaviors—in order to evaluate whether CWB and withdrawal have similar or different relationships with common correlates. These two pieces of meta-analytic evidence are important for determining the degree of overlap between CWB and withdrawal. Indeed, if the evidence suggests that CWB and withdrawal can be empirically distinguished, then this would mean that future scholarship on employee deviance should acknowledge such uniqueness. On the other hand, if the evidence suggests CWB and withdrawal are not empirically distinct, future researchers should recognize this and better integrate CWB and withdrawal. As such, this study represents a necessary clarification of the CWB-withdrawal relationship and provides important implications for the future measurement and theoretical understanding of CWB and withdrawal.

The Conceptual Relationship Between CWB and Withdrawal

Some of the confusion regarding CWB and withdrawal stems from the fact that the development of these constructs occurred not only rather independently but also prior to an understanding of their respective theoretical underpinnings (Harrison, 2002). CWB developed as an explication of the varied negative behaviors (beyond theft) employees may enact at work (e.g., Hollinger & Clark, 1983), while withdrawal’s development is rooted in the job satisfaction literature, with withdrawal conceptualized as an important outcome of employees’ dissatisfaction (Hanisch & Hulin, 1990, 1991). Thus, until recently these literatures existed in relative isolation from one another and, as a result, advances in the measurement and conceptualization of these constructs have not always been integrated or acknowledged.

Despite this disconnect, CWB and withdrawal are similarly theorized as labels for behavioral families that are undesired by organizations (see Berry et al., 2007; Hanisch & Hulin, 1991; Hanisch, Hulin, & Roznowski, 1998). The major divergence regarding the CWB-withdrawal relationship pertains to whether the behaviors composing CWB and withdrawal (a) represent two distinct constructs, such that they are separate tendencies or sets of responses to negative events (Lehman & Simpson, 1992; Murphy, 1989, 1990); or (b) reflect a single construct, meaning they serve as similarly viable responses to negative work events and/or reflect a single, core propensity of employees to enact negative work behaviors (e.g., Hanisch & Hulin; Hogan & Hogan, 1989; Marcus & Schuler, 2004; Rotundo & Spector, 2010; Spector & Fox, 2005). To be clear, the latter perspective does not necessarily entail that CWB and withdrawal are completely redundant (although this is one possibility). CWB and withdrawal could reflect the same construct by being related facets of a broader negative work behavior construct, much like interpersonal-target CWB (CWBI) and organizational-target CWB (CWBO) are related facets of a broader CWB construct (Berry et al.).

Thus, a key issue is also the conceptual relationship between CWB and withdrawal. That is, what are the important areas of overlap or the similarities and differences in the conceptualizations of these constructs? O’Leary-Kelly, Duffy, and Griffin (2000; see also Buss, 1961) outlined a taxonomic framework of dimensions on which antisocial work behaviors can differ from other similar constructs. In Table 1, we have adapted the framework of O’Leary-Kelly et al. to compare similarities and differences between CWB and withdrawal. We note important similarities and differences in the (a) constitutive elements, (b) targets, and (c) motives/intentions of CWB and withdrawal that are critical to understanding their conceptual relationship.

Table 1

Framework of Conceptual Similarities and Differences Between Counterproductive Work Behavior (CWB) and Withdrawal

Dimension/Factors^a	CWB	Withdrawal
Dimensions on which CWB and withdrawal are similar
Active and Passive	CWB reflects both active and passive behaviors, as CWB includes the perpetrator’s action (theft) or withholding of action (withholding information).	Withdrawal reflects both active and passive behaviors, as employees may arrive late or withhold effort or attention from their job.
Attempted and Completed	CWB reflects both attempted and completed behaviors; for example, CWB includes an employee attempting to harm someone and whether the attempt is successful.	Withdrawal includes both attempts and completed behaviors. Employees can attempt to avoid their work duties or they can succeed in being absent.
Perpetrator	CWB is enacted by organizational members.	Withdrawal is enacted by organizational members.
Dimensions on which CWB and withdrawal are different
Verbal and Physical	CWB includes both physical and verbal behaviors. Interpersonal CWB reflects gossip and verbal harassment, as examples.	Withdrawal includes only physical behaviors, such as daydreaming, staying home, or being late.
Intentionality	CWB includes behaviors intended to harm, as the behaviors are defined as purposeful and voluntary behaviors.	Withdrawal is not regarded as intended to harm. For example, lateness, absence, and daydreaming may reflect an intent to avoid the organization.
Target	CWB can be directed towards the organization or towards other individuals (e.g., employees, clients).	Withdrawal can be directed towards the organization but is not defined as having an individual target.

Adapted from the framework of O’Leary-Kelly, Duffy, and Griffin (2000; see also Buss, 1961).

First, CWB and withdrawal may reflect the same construct if they are theorized as having the same constitutive elements. As shown in Table 1, CWB and withdrawal have core behavioral commonalities such that they both (a) can reflect active and passive behavior, (b) can refer to attempted and completed actions, and (c) are perpetrated by organizational members. These shared elements in the behaviors composing CWB and withdrawal support the likely existence of a broad, single propensity underlying CWB and withdrawal. Indeed, this is reflected in theories suggesting that negative job attitudes, such as job dissatisfaction (Hulin, Roznowski, & Hachiya, 1985), or dispositional traits (e.g., Hogan & Hogan, 1989; Johns, 1997) likely influence employees to display behavioral patterns such that they tend to attempt and complete active and passive forms of both CWB and withdrawal. On the other hand, differences in constitutive elements would suggest that employees may have two distinct tendencies to enact either CWB or withdrawal. For example, CWB is clearly conceptualized as reflecting both verbal and physical actions (Neuman & Baron, 2005), as CWB includes an employee physically harming a coworker or destroying organizational property and arguing with or gossiping about someone (e.g., R. J. Bennett & Robinson, 2000). However, withdrawal appears exclusively defined in terms of physical behaviors such as daydreaming, staying home from work, or being late.

CWB and withdrawal also differ in their respective targets or victims. Negative work behaviors that target the organization are expressly reflected in the conceptualizations of both CWB (e.g., CWBO; theft or destruction of organizational property) and withdrawal (e.g., attendance problems). However, CWB is additionally regarded as reflecting actions that are directly targeted towards individuals, such as coworkers, supervisors, and/or customers (CWBI; Neuman & Baron, 1997, 2005; Robinson & Bennett, 1995). Indeed, we note that CWB is often conceptualized as a hierarchical construct, with an overall dimension at the top level and CWBI and CWBO dimensions at the second level (Berry et al., 2007). Although an employee’s withdrawal actions, such as lack of attendance, attention, or effort, may certainly have an eventual effect on other organizational members, withdrawal behaviors are not expressly regarded as being directly targeted towards others.

Finally, CWB and withdrawal may differ in the motive theorized to underlie the respective behaviors. Many CWBs may be regarded as behaviors intended to harm the organization or organizational members (Neuman & Baron, 1997, 2005; Robinson & Bennett, 1995; Robinson & Greenberg, 1999; Spector & Fox, 2002), while withdrawal could be driven by a distinct intent to avoid (Hanisch & Hulin, 1991; Hulin, 1991). For example, CWB and withdrawal have been linked as distinct “fight or flight” responses to work events that provoke negative affect (Dalal, Lam, Weiss, Welch, & Hulin, 2009). This distinction between motives is not always clear, though. For example, one could reason that an employee’s absence, defined as his or her lack of physical presence at a defined setting and time (Harrison, 2002), could be due to the intent to harm or avoid a supervisor. Additionally, employees do not need to intend harm to carry out many CWBs (e.g., Gruys & Sackett, 2003), for example, as an employee who engages in unsafe behaviors does not necessarily do so to harm others or the organization.

Therefore, as summarized in Table 1, CWB and withdrawal are similar on the first three dimensions of the O’Leary-Kelly et al. (2000) taxonomy but differ somewhat on the verbal/physical, target, and intentionality dimensions. Importantly, most of the differences between CWB and withdrawal (i.e., on the verbal/physical and target dimensions) are actually differences between CWBI and withdrawal, not CWBO and withdrawal. That is, most verbal CWBs are CWBIs (e.g., verbal abuse, harassment, gossip) and both withdrawal behaviors and CWBOs target the organization. As a result, it may be that CWBI and withdrawal are separate constructs while CWBO and withdrawal are not. In general, the conceptual similarities and differences beg the question of whether these conceptual differences are enough to outweigh the many conceptual similarities between these constructs. Thus, a key line of meta-analytic evidence provided by the present study is the empirical relationship between withdrawal and CWB and the CWBI and CWBO dimensions.

Predictors of CWB and Withdrawal

For two constructs such as CWB and withdrawal to be considered distinct, they should have different relationships with their predictors and other variables in their nomological networks (Berry et al., 2007). Alternatively, if CWB and withdrawal reflect the same construct, they should be driven by similar variables. We highlight two different categories of theoretical perspectives that have been used to understand what drives (and may distinguish) CWB and withdrawal: person- and situation-oriented perspectives.

Person-oriented perspectives focus on employee traits and dispositional characteristics as antecedents of CWB and withdrawal. From this theoretical lens, CWB and withdrawal could reflect a common construct because the same stable dispositions, such as personality traits, lead employees to enact a consistent behavioral repertoire or pattern composed of both CWB and withdrawal actions. These theoretical positions are found in the general theory of crime (Gottfredson, 1990), the delinquency theoretical perspective (Hogan & Hogan, 1989), and the deviance hypothesis (Johns, 1987, 1997) and commonly assert that a general dispositional trait or constellation of traits (Harrison & Martocchio, 1998) influences an employee’s general tendency to engage in a broad range of negative behaviors (e.g., CWBs and withdrawal behaviors). In this case, CWB and withdrawal would reflect a common construct because both types of behaviors are manifestations of the same overall deviance tendency or propensity. Of course, person-oriented perspectives could also support CWB and withdrawal being distinct constructs if CWB and withdrawal are associated with separate sets of dispositions. Because CWBs have sometimes been defined as reflecting an intent to harm (e.g., Dalal et al., 2009; James et al., 2005), this might suggest they are driven by things like trait aggression. As described above, withdrawal behaviors are not typically linked to aggression or an intent to harm, suggesting that employee characteristics other than aggression might be at play.

Person-oriented theories are complemented by situation-oriented theories, which focus on the characteristics of workplace situations as driving the enactment of CWB and withdrawal. Affective events theory (Weiss & Cropanzano, 1996) and the stressor-strain model (Spector & Fox, 2005) are situation-oriented theories that provide additional insight into the theoretical relationship between CWB and withdrawal. According to these theories, the enactment of CWB or withdrawal “depends” on the precipitating context or situation. Thus, one employee could plausibly enact CWB and another employee could engage in withdrawal, depending on the organizational situation, context, or their respective reactions to a precipitating event. For example, a conflict with a coworker could lead an employee to respond by sabotaging his or her work to punish the coworker (i.e., CWB), while an unjust action from the organization could lead the same employee to leave his or her workstation (i.e., withdrawal). Thus, CWB and withdrawal would be systematically precipitated by different events, and maintaining the distinction between CWB and withdrawal would be critical for both theoretical and empirical reasons. However, if employees are, for example, just as likely to intentionally sabotage their productivity as they are to be absent or late in response to an unfair work decision, then CWB and withdrawal could reflect the same construct.

Person- and situation-oriented theories can lend conceptual support to the common construct or distinct constructs perspectives for CWB and withdrawal. As such, an empirical comparison of the relationships that CWB and withdrawal have with their person- and situation-based antecedents, along with other variables in their nomological networks, is important to shed light on whether CWB and withdrawal reflect the same or distinct constructs.

Current Study: Hypotheses and Research Questions

The purpose of the current study is to advance understanding of the CWB-withdrawal relationship as the constructs are currently measured in the literature. That is, this meta-analysis focuses on and isolates the current operationalizations of CWB and withdrawal. As we show in Table 2, this measurement focus is fundamentally important to the aims of this study because many of the prevalent measures used to assess CWB and withdrawal unfortunately contain instances of overlapping content, meaning that similarly worded items are used to measure both constructs. This general issue of content-overlapping items has been recognized in several domains and by several authors (e.g., Bozeman & Perrewé, 2001; Carpenter, Newman, & Arthur, 2011; Dalal, 2005; Spector, Bauer, & Fox, 2010; Stone-Romero, Alvarez, & Thompson, 2009), and it was important to acknowledge that this issue also applies to the understanding of the CWB-withdrawal relationship. Indeed, one possibility is that these shared items reflect actual construct commonalities in CWB and withdrawal. Thus, the present study serves as a quantitative summary of this commonality.

Table 2

Examples of Content-Overlapping Counterproductive Work Behavior (CWB) and Withdrawal Items

Example CWB Items	Example Withdrawal Items
Called in sick when not ill^d Called in sick when you were not^a	I might call in sick occasionally because I don’t feel like working on my job^c Sometimes when I just don’t feel like working I will call in sick^c
Come in late to work without permission^a	Came to work late without permission^e
Spent too much time fantasizing or daydreaming instead of working^a	Daydreaming^b
Intentionally worked slower^d Intentionally worked slower than you could have worked^a Purposely worked slowly when things needed to get done^e	I have lost motivation for my assigned job; I might work more slowly or make more errors^c Sometimes when I don’t feel like working I will work slowly or make errors^c
Left work early without permission^a	Left work early without permission^b Left work earlier than you were allowed to^e
Put little effort into your work^a	Now and then there are workdays where I just don’t put much effort into my work^c Put less effort into job than should have^b
Spent time on personal matters while at work^d Worked on a personal matter instead of work for your employer ^a	Spent work time on personal matters^b
Taken an additional or longer break than is acceptable at your workplace^a Took an extended coffee or lunch break^d	Taken a longer break than you were allowed to take^e Taken longer lunch or rest break than allowed^b
Taken property from work without permission^a,d	Taken supplies or equipment without permission^b

Note: The table presents items from scales that are commonly used to measure both CWB and withdrawal.

R. J. Bennett and Robinson (2000).

Lehman and Simpson (1992).

Rusbult, Farrell, Rogers, and Mainous (1988).

Skarlicki and Folger (1997).

Spector, Fox, Penney, Bruursema, Goh, and Kessler (2004).

Empirical relationship between CWB and withdrawal

If CWB and withdrawal theoretically reflect a common construct, then we would expect measures of both sets of behaviors to be strongly interrelated. The specific boundary conditions of this expectation should align with existing guidelines and empirical evidence regarding indicators of a common construct (Edwards & Berry, 2010). If two variables are completely redundant and simply measure the same construct, their intercorrelation should approach unity, especially when corrected for error of measurement. However, as mentioned previously, two variables do not need to be completely redundant to reflect a common construct; they could, for instance, be related facets or indicators of a common construct. There is no consensus pertaining to how strongly variables must be intercorrelated before one can conclude they reflect a common construct. However, it is reasonable that if two variables reflect a common construct, their level of intercorrelation should generally meet or exceed a “large” effect size (i.e., r is at least .50; Cohen, 1992). Further, LePine, Erez, and Johnson (2002) regarded an average organizational citizenship behavior (OCB) facet intercorrelation of .67 as evidence of a common OCB construct, while Berry et al. (2007) interpreted the correlation of .62 between CWBI and CWBO as evidence that both reflected an overall CWB factor. Thus, on the basis of this prior evidence, relationship magnitudes in excess of .50, and especially nearing .70, would suggest CWB and withdrawal are indicators of a common construct. Conversely, if CWB and withdrawal reflect distinct constructs, the empirical relationships should show moderately sized correlations, such as small or medium effect sizes (of r = .30 or less; Cohen). For example, Berry, Lelchook, and Clark (2012) interpreted withdrawal facet intercorrelations ranging from .01 to .26 as evidence of distinctiveness, while Dalal’s (2005) meta-analytic estimate of –.32 for the relationship between CWB and OCB was also regarded as evidence of separate constructs. CWB and withdrawal both represent negative workplace behaviors and, as noted in Table 1, also share important features. Thus, we generally expected that CWB and withdrawal would have some degree of covariation and, therefore, be positively related. Similarly, as we noted above, CWBO and withdrawal share more correspondence than CWBI and withdrawal, which suggested that withdrawal should be more strongly related to CWBO than to CWBI.

Hypothesis 1: CWB and withdrawal measures will be positively related.

Hypothesis 2: Withdrawal will be more strongly related to CWBO than CWBI.

Research Question 1: Which theoretical perspective does the CWB-withdrawal empirical relationship support: The common construct or distinct constructs perspective?

Nomological networks

A second line of evidence is the extent to which CWB and withdrawal measures have similar or different relationships with their correlates. In line with person- and situation-oriented theories, CWB and withdrawal both are regarded as being related to dispositions, such as personality traits (Berry et al., 2007; Ones, Viswesvaran, & Schmidt, 2003) and negative affect (Hershcovis et al., 2007; Holtom, Burton, & Crossley, 2012); job attitudes, such as job satisfaction (Berry, Carpenter, & Barratt, 2012; Hulin, 1991) and organizational commitment (Dalal, 2005; Mathieu & Zajac, 1990); organizational perception and context variables, such as organizational justice (e.g., Cohen-Charash & Spector, 2001; Colquitt, Conlon, Wesson, Porter, & Ng, 2001) and stress (e.g., Hershcovis et al.; Podsakoff, LePine, & LePine, 2007); work performance behaviors, such as OCB (Harrison, Newman, & Roth, 2006); and demographic variables, such as age (Berry et al., 2007; Hom & Griffeth, 1995). Thus, we conducted new meta-analyses of the relationships CWB and withdrawal each have with these important common correlates.

If CWB and withdrawal behavior are aspects of the same negative work behavior construct, then they should have similar patterns and magnitudes of relationships with these correlates. One threshold for this empirical relationship might be that CWB and withdrawal’s correlations with their correlates not differ in magnitude by more than about .10, which denotes a small difference between correlations (Cohen, 1992). For example, Berry, Carpenter, and Barratt (2012) concluded that self- and observer-rated CWB had similar relationships with their common correlates and, thus, were redundant with each other, in part on the basis of average differences in correlations with common correlates between .07 and .13. Thus, differences between nomological network relationships larger than .10 would be evidence that CWB and withdrawal may be considered separate constructs.

Research Question 2: Do CWB and withdrawal measures have similar or different patterns and magnitudes of relationships with variables in their nomological networks?

Moderators

Several relevant measurement-related variables exist that could plausibly affect the CWB-withdrawal relationship. For example, common operationalizations not only align with different theoretical frameworks of CWB and withdrawal but also contain content-overlapping items (see Table 2), which suggests that the relationship could depend on the measure that is used. Thus, we examined whether the CWB-withdrawal relationship differed on the basis of the measures used to assess the behaviors. We also expected the relationship to be stronger when the same individual (e.g., an employee or supervisor rates both behaviors) provides ratings of CWB and withdrawal than when different sources rate each behavior, particularly because of common method variance and halo error. Finally, we examined whether the CWB-withdrawal relationship depends on whether withdrawal is measured as an actual behavior (e.g., psychological withdrawal, lateness, absenteeism) or an intention (i.e., turnover intentions). For example, because intentions are considered precursors of actual behavior (Ajzen, 1985, 1987), we expected that CWB’s relationship with withdrawal intentions would be significantly smaller than the relationship with withdrawal behavior. However, we discovered that in all studies in our meta-analytic database, withdrawal intentions were measured as turnover intentions. As such, our moderator analysis focuses only on the CWB-withdrawal relationship when withdrawal is measured as a behavior versus as turnover intentions. There certainly exist additional measurement-related factors that could influence the relationship between CWB and withdrawal, for example, whether (a) withdrawal is measured objectively (i.e., absence records) or subjectively (e.g., self-reported absence), (b) employee absence or lateness was described as voluntary/involuntary (e.g., avoidable or unavoidable) or excused/unexcused, or (c) actual turnover versus turnover intentions were measured. Unfortunately, there were not enough samples located to empirically examine these relevant moderators.

Moderator Hypothesis 1: CWB and withdrawal will be more strongly related when both are rated by the same rater (e.g., supervisor) than by different raters (e.g., self-rated CWB and supervisor-rated withdrawal).

Moderator Research Question 1: Does the type of CWB or withdrawal measure influence the CWB-withdrawal relationship?

Moderator Hypothesis 2: CWB and withdrawal will be more strongly related when withdrawal is measured as a behavior than as turnover intentions.

Method

Literature Search

To locate primary studies, we first conducted a keyword search for published articles and unpublished papers through 2013 using the ABI/Inform, PsycINFO, and Proquest Dissertations databases. Next, conference papers from 2007 to 2013 were located through keyword searches of conference proceedings for the Academy of Management and Society for Industrial and Organizational Psychology. A comprehensive list of search terms was used, with examples including devian-*, counterproductiv-*, abuse, aggression, antisocial, bullying, dysfunctional work behavior, drug/alcohol use, harmdoing, incivility, property destruction, retaliation, sabotage, social undermining, workplace victimization, attendance, absence, lateness/tardiness, withdrawal, and turnover.¹ Both constructs’ terms were included in literature searches, and searches were conducted separately for each construct as well. Finally, prominent researchers of both CWB and withdrawal were contacted with requests for unpublished data.

Inclusion Criteria

A primary study was included if it provided enough information to obtain a correlation between (a) overall CWB or a specific CWB dimension (e.g., theft, CWBI) and (b) overall withdrawal or a specific withdrawal dimension (e.g., absence, lateness). Within these studies, relationships among CWB, withdrawal, and common correlates (e.g., job satisfaction, commitment) were also coded. One study that met these criteria, Hunt (1996), was excluded because (a) it used many of the same items to measure CWB and withdrawal and (b) included very large sample sizes (Ns between 2,000 and 6,500) relative to other studies and was therefore an influential outlier (e.g., a number of relationships were large, greater than –.60, with Hunt included, but were essentially 0 when Hunt was removed).² Using these criteria resulted in the inclusion of 45 studies with 46 independent samples. Of these samples, 32 were published and 14 were unpublished.

Procedure

For each sample, a correlation was coded for (a) the relationship between CWB and withdrawal (overall and/or specific dimensions) and, if applicable, (b) the relationships between CWB or withdrawal and a common correlate. To estimate the relationship between overall CWB and overall withdrawal when studies provided only correlations between multiple narrow dimensions of CWB or withdrawal (e.g., correlations between withdrawal and CWBI and CWBO), we used composite formulas (Ghiselli, Campbell, & Zedeck, 1981) to estimate the overall correlation. For coding of moderators, we first recorded the precise label and scales used to represent and measure CWB and withdrawal.³ In all, there were 16 different labels and 18 different scales used to represent CWB, while 12 labels and 18 scales were used to represent withdrawal. In addition, we coded the rating source (i.e., self, supervisor, or coworker) of CWB and withdrawal as well as whether withdrawal was measured as a behavior or as turnover intentions. All coding was initially completed by the first author of this study. To ensure the accuracy and reliability of the coding, we also had two independent raters (doctoral students in industrial and organizational psychology with master’s degrees) code approximately one quarter (k = 11) of the total samples. With the exception of minor transcription errors, the overall agreement between coders on the subset of coding was 100%.

Meta-analyses

Artifact distribution meta-analytic procedures specified by Hunter and Schmidt (2004) were used to calculate mean correlations and variability of relationships between CWB, withdrawal, and their common correlates. We corrected correlations for unreliability in both variables (see Table 3 for artifact distributions). Next, when there were at least three samples in each moderator category, moderator analyses were performed on these subsamples using the same meta-analytic procedures. To determine whether there was a significant difference between correlations across moderator categories, we used formulas from Raju and Brand (2003) to test whether the differences between corrected correlations were significant.

Table 3

Reliability (Alpha) Artifact Distributions

Type	r_xx	SD	N	k	Source
CWB
CWB	.77	—	16,721	49	Dalal (2005)
CWBI	.84	.07	6,878	26	Berry, Ones, & Sackett (2007)
CWBO	.82	.07	6,080	22	Berry et al. (2007)
Withdrawal
Withdrawal	.79	.09	18,523	37	Present Study
Absenteeism	.68	.12	5,780	8	Present Study
Turnover Intentions	.82	.13	7,654	15	Present Study
Lateness	.70	.11	329	3	Present Study
Common Correlates
Overall OCB	.79	—	16,455	47	Dalal (2005)
Task Performance	.85	—	2,697	17	Viswesvaran, Ones, & Schmidt (1996)
Job Satisfaction	.84	.06	1,415	7	Connolly & Viswesvaran (2000)
Commitment	.85	—	42,032	166	Mathieu & Zajac (1990)
Conscientiousness	.78	.10	—	307	Viswesvaran & Ones (2000)
Extraversion	.78	.09	—	307	Viswesvaran & Ones (2000)
Emotional Stability	.78	.11	—	370	Viswesvaran & Ones (2000)
Agreeableness	.75	.11	—	123	Viswesvaran & Ones (2000)
Openness to Experience	.73	.12	—	251	Viswesvaran & Ones (2000)
Negative Affect	.79	.11	3,326	15	Connolly & Viswesvaran (2000)
Positive Affect	.81	.07	3,326	15	Connolly & Viswesvaran (2000)
Self-Control	.78	.10	—	307	Viswesvaran & Ones (2000)
Social Desirability	.76	—	—		Higgins, Judge, & Ferris (2003)
Stress	.87	—	903	4	Lee & Ashforth (1996)
Conflict	.74	—	3,363	13	Spector & Jex (1998)
Constraints	.85	—	1,746	8	Spector & Jex (1998)
Procedural Justice	.91	—	—	66	Hauenstein, McGonigle, & Flinder (2001)
Distributive Justice	.91	—	—	66	Hauenstein et al. (2001)
Interpersonal Justice	.89	.04	645	3	Berry et al. (2007)
Organizational Support	.90	—	—	73	Rhoades & Eisenberger (2002)

Note: Type = type of reliability coefficient used in the artifact distribution; r_xx = reliability artifact distribution mean; SD = reliability artifact distribution standard deviation; N = reliability artifact distribution sample size; k = number of samples contributing to artifact distributions; Source = study from which artifact distributions were drawn; CWB = counterproductive work behavior; CWBI = interpersonal CWB; CWBO = organizational CWB; OCB = organizational citizenship behavior.

For the overall relationships between withdrawal and CWB, CWBI, and CWBO, we also reported correlations corrected for indirect range restriction in CWB because the employee/incumbent sample variance in CWB (and withdrawal) was restricted relative to that in the population as a result of a number of factors, such as selecting individuals on the basis of characteristics that make them less likely to commit CWBs, firing individuals caught engaging in CWB, and putting into place organizational policies and/or constraints that curtail CWB enactment. Restriction-corrected and restriction-uncorrected correlations address different, meaningful research questions. The restriction-uncorrected correlations index the relationship between CWB and withdrawal for the actual employees on the job at a given point in time. The restriction-corrected correlations more directly address whether CWB and withdrawal measures are measuring the same thing. For example, if the restricted CWB-withdrawal correlation is well below the threshold for two measures reflecting the same construct, this smaller correlation might not mean CWB and withdrawal are actually measuring different things; the smaller relationship could rather be due to low variance on CWB and withdrawal as a result of these restriction processes. The restriction-corrected correlations estimate the relationship free of these restriction processes.

We used Hunter, Schmidt, and Le’s (2006) indirect range restriction correction method. This required four pieces of information: the uncorrected/restricted CWB-withdrawal correlation (calculated in the present meta-analysis), the restricted reliabilities of CWB and withdrawal (these two pieces of information were drawn from artifact distributions listed in Table 3), and the u ratio for CWB (i.e., the restricted standard deviation of CWB divided by the unrestricted standard deviation of CWB). Berry et al. (2007) is one of the largest existing CWB meta-analyses, so we obtained all of the studies in Berry et al. and examined them for information needed to calculate u ratios. Thirteen samples (combined N = 3,612) provided the necessary information (a list of these samples is available on request from the first author). These samples each listed restricted means and standard deviations of CWB. These samples used scales ranging from 1 to 4 to 1 to 7 to measure CWB; we converted all CWB scales to 7-point (1–7) scales. The unrestricted standard deviations of CWB were calculated using a method introduced by Hunter (1983). Specifically, pooling the data across these 13 samples estimates the standard deviation of CWB in the work population, and a basic formula from analysis of variance shows that total variance (i.e., the variance pooled across studies) is equal to the variance of the CWB means plus the mean of the CWB variances. These methods resulted in estimates of the mean restricted standard deviations of CWBI and CWBO (.84 and .69, respectively), the unrestricted standard deviations of CWBI and CWBO (1.23 and 1.22, respectively), and u ratios for CWBI and CWBO (.68 and .56, respectively). Averaging these u ratios provides an estimate of the u ratio for overall CWB (.62). These u ratios were used in correcting for indirect range restriction.

Results

Relationship Between CWB and Withdrawal

The meta-analytic results for the relationships between CWB and withdrawal measures are presented in Table 4. Hypothesis 1 proposed that the relationship between measures of CWB and withdrawal would be positive. The sample-size-weighted correlation between CWB and withdrawal was .45 and the reliability-corrected correlation was .58. The 95% confidence interval suggested a large overall CWB-withdrawal relationship that differed from 0 and was less than unity, supporting Hypothesis 1. When corrected for indirect range restriction, the CWB-withdrawal correlation was .79, also supporting Hypothesis 1.⁴

Table 4

Meta-Analytic Results: Relationships Between CWB and Withdrawal Measures

Variable	N	k	r _m	SD_r	ρ	SD _ρ	% Var	CV₁₀	CV₉₀	CI_L	CI_U	z
CWB—Withdrawal	20,827	46	.45	.16	.58	.19	5.80	.34	.83	.52	.64
CWB Dimensions
CWBI—Withdrawal	5,865	18	.40	.20	.50	.24	5.13	.19	.81	.38	.61	5.00a
CWBO—Withdrawal	13,080	28	.46	.16	.58	.19	5.05	.33	.83	.50	.65
Theft—Withdrawal	9,082	15	.44	.13	.55	.16	6.35	.35	.75	.47	.63	3.06^a,b
Sabotage—Withdrawal	4,071	12	.39	.19	.49	.23	5.65	.20	.79	.36	.63	−2.32^a,c
Production Deviance—Withdrawal	5,038	14	.47	.16	.59	.19	6.31	.34	.84	.48	.69	−4.66^a,d
Withdrawal Dimensions
CWB—Absenteeism	7,676	11	.37	.18	.52	.25	3.12	.20	.84	.37	.67
CWBI—Absenteeism	1,164	4	.15	.18	.19	.22	10.24	−.09	.48	−.04	.43	−7.18a
CWBO—Absenteeism	7,066	7	.36	.23	.48	.30	1.47	.10	.86	.26	.70
CWB—Turnover Intentions	8,049	17	.23	.09	.29	.10	24.30	.17	.41	.24	.34
CWBI—Turnover Intentions	1,526	6	.21	.15	.25	.17	16.45	.04	.46	.11	.39	−0.61
CWBO—Turnover Intentions	6,878	10	.22	.06	.27	.05	40.55	.20	.34	.23	.31
CWB—Lateness	329	3	.26	.16	.33	.16	32.52	.12	.53	.10	.55

Note: r_m = mean sample-size-weighted correlation; SD_r = sample-size-weighted observed standard deviation of correlations; ρ = mean sample-size-weighted correlation corrected for unreliability; SD_ρ = corrected standard deviation of corrected correlations; % Var = percentage of variance attributable to statistical artifacts; CV₁₀ and CV₉₀ = 10% and 90% credibility values, respectively; CI_L and CI_U = lower and upper 95% confidence interval values, respectively; CWB = counterproductive work behavior; CWBI = interpersonal CWB; CWBO = organizational CWB.

Indicates significant differences between correlations (α = .05).

Refers to the comparison of Theft—Withdrawal and Sabotage—Withdrawal.

Refers to the comparison of Theft—Withdrawal and Production Deviance—Withdrawal.

Refers to the comparison of Sabotage—Withdrawal and Production Deviance—Withdrawal.

Next, we examined the relationship between measures of CWB and withdrawal at the level of the narrow dimensions. In support of Hypothesis 2, we found that CWBO (reliability-corrected r = .58) was significantly more strongly related to withdrawal than was CWBI (reliability-corrected r = .50, z = 5.00, p < .05), although both were correlated at least .50 with withdrawal. However, the u ratios for CWBI (.68) and CWBO (.56) suggested CWBO was more restricted in range than CWBI. When corrected for indirect range restriction, CWBO is even more strongly related to withdrawal than CWBI (rs of .80 vs. .66, respectively; we are not aware of a significance test for correlations corrected for indirect range restriction). We also found that three specific forms of CWBO—theft (ρ = .55), sabotage (ρ = .49), and production deviance (ρ = .59)—were strongly related to withdrawal (not enough studies reported such relationships for specific forms of CWBI). In general, the relationships between CWB dimensions and overall withdrawal also supported CWB and withdrawal as strongly related, with CWBO and its more specific forms (theft, sabotage, and production deviance) being, in many cases, more strongly related than CWBI to withdrawal.

We also examined relationships between withdrawal’s dimensions and CWB (broad and dimensions). Overall CWB was strongly related to absenteeism (ρ = .52), but this relationship was largely driven by CWBO. Specifically, CWBO (ρ = .48) was significantly more strongly related to absenteeism than was CWBI (ρ = .19, z = −7.18, p < .05), which provided further support for Hypothesis 2. We also found that CWB was moderately related to turnover intentions and lateness (i.e., corrected correlations ranged from .25 to .33; we note that the lateness meta-analysis included only three samples).

Moderator analyses

The moderator results are presented in Table 5. We first examined if the type of CWB and withdrawal scale moderated the relationship between broad CWB and withdrawal measures. The use of the Counterproductive Work Behavior Checklist (CWB-C; Spector et al., 2004; ρ = .68) led to significantly larger relationships than both the R. J. Bennett and Robinson (2000; ρ = .45, z = 9.26, p < .05) and Lehman and Simpson (1992; ρ = .57, z = 5.46, p < .05) measures. The Lehman and Simpson measure also led to significantly larger relationships than the Bennett and Robinson measure (z = 4.38, p < .05). When corrected for indirect range restriction, these differences in the CWB-withdrawal correlation between measures remain; however, the absolute magnitudes of the CWB-withdrawal relationship become substantial for all measures (.85, .77, and .66 for the CWB-C, Lehman and Simpson, and Bennett and Robinson measures, respectively). Similarly, for CWBI and CWBO relationships with withdrawal, the CWB-C measure led to significantly stronger relationships than when the Bennett and Robinson measure was used. Importantly, the relationship between CWB (particularly CWBO) and withdrawal remained quite large regardless of the type of measure used. For example, when corrected for indirect range restriction, the CWBI-withdrawal relationships were .84 for the CWB-C measure but only .23 for the Bennett and Robinson measure; however, the CWBO-withdrawal relationships were large for both measures (.87 and .68 for the CWB-C and Bennett and Robinson measures, respectively). This further supports Hypothesis 2.

Table 5

Meta-Analysis of CWB-Withdrawal Measures: Measurement Moderator Results

Variable	N	k	r _m	SD_r	ρ	SD _ρ	% Var	CV₁₀	CV₉₀	CI_L	CI_U	z
CWB—Withdrawal
Type of Measure
CWB-C	6,246	18	.53	.18	.68	.22	4.87	.40	.96	.58	.78	5.46^a,b
Lehman & Simpson (1992)	3,985	5	.44	.13	.57	.16	5.08	.37	.77	.43	.71	9.26^a,c
R. J. Bennett & Robinson (2000)	2,654	9	.35	.16	.45	.20	9.66	.19	.70	.31	.59	4.38^a,d
Rating Source
Self-Rating (same)	19,506	39	.46	.15	.59	.19	5.43	.36	.83	.53	.66	1.27^e
Supervisor Rating (same)	746	4	.42	.11	.54	.12	30.27	.40	.69	.41	.68	9.12^a,f
Self-Other (different)	1,083	6	.19	.08	.24	.03	89.69	.20	.28	.16	.32	5.55^a,g
Behavior vs. Intention
Withdrawal Behavior	18,832	36	.47	.15	.61	.18	5.40	.38	.84	.55	.67	20.36^a
Turnover Intention^h	8,296	17	.23	.11	.30	.13	15.69	.13	.46	.23	.36
CWBI—Withdrawal
Type of Measure
CWB-C	3,490	10	.50	.18	.62	.21	5.05	.35	.89	.48	.75	15.92^a
R. J. Bennett & Robinson (2000)	2,000	7	.10	.15	.12	.17	15.53	−.09	.33	−.01	.26
Rating Source
Self-Rating (same)	5,718	17	.41	.21	.50	.24	4.90	.19	.82	.38	.62	5.16^a
Self-Other (different)	552	3	.19	.02	.23	.00	100.00	.23	.23	.21	.26
Behavior vs. Intention
Withdrawal Behavior	5,636	16	.41	.21	.51	.24	4.63	.20	.82	.38	.63	7.87^a
Turnover Intention^h	1,526	6	.21	.15	.25	.17	16.45	.04	.46	.11	.39
CWBO—Withdrawal
Type of Measure
CWB-C	5,073	14	.55	.14	.68	.17	7.01	.47	.89	.59	.77	7.91^a
R. J. Bennett & Robinson (2000)	1,548	7	.35	.21	.44	.25	7.92	.12	.75	.24	.63
Rating Source
Self-Rating (same)	12,683	26	.47	.16	.58	.19	4.89	.34	.83	.51	.66	7.93^a
Self-Other (different)	604	3	.14	.09	.18	.06	64.24	.10	.26	.06	.30
Behavior vs. Intention
Withdrawal Behavior	12,120	23	.47	.16	.59	.20	4.28	.33	.84	.50	.67	19.04^a
Turnover Intention^h	6,878	10	.22	.06	.27	.05	40.55	.20	.34	.23	.31
CWB—Turnover Intentions
Rating Source
Self-Rating (same)	7,651	14	.23	.09	.29	.10	20.43	.16	.42	.23	.35	0.00
Self-Other (different)	398	3	.23	.06	.29	.00	100.00	.29	.29	.21	.38

Indicates significant differences between correlations (α = .05).

Refers to the comparison of CWB-C and Lehman & Simpson (1992).

Refers to the comparison of CWB-C and R. J. Bennett & Robinson (2000).

Refers to the comparison of Lehman & Simpson (1992) and R. J. Bennett & Robinson (2000).

Refers to the comparison of Self-Rating (same) and Supervisor Rating (same).

Refers to the comparison of Self-Rating (same) and Self-Other (different).

Refers to the comparison of Supervisor Rating (same) and Self-Other (different).

Because all samples measured turnover intentions as a form of withdrawal intentions, this moderator test represents a comparison of withdrawal behavior and turnover intentions.

Next, we found that relationships between broad CWB and withdrawal were largest when either self-ratings (ρ = .59) or supervisor ratings (ρ = .54) were used for both behaviors; the CWB-withdrawal correlation was weakest when different sources rated the behaviors (ρ = .24). Similarly, the CWBI-withdrawal relationship was significantly larger for same-source self-ratings than for different rating sources (z = 5.16, p < .05), as was the CWBO-withdrawal relationship (z = 7.93, p < .05). However, we note that all of the different source meta-analyses were based on very few samples.

Finally, we examined whether withdrawal being measured as a behavior or turnover intentions moderated CWB-withdrawal relationships. In general, relationships between broad CWB and withdrawal were strongest when withdrawal was measured as behavior (ρ = .61) rather than as turnover intentions (ρ = .30, z = 20.36, p < .05). This pattern was also shown for CWBI (z = 7.87, p < .05) and CWBO (z = 19.04, p < .05). Thus, the results of this moderator test demonstrated that the relationship between CWB and withdrawal was strongest for withdrawal behaviors compared to turnover intentions.⁵

Comparison of Relationships Among CWB, Withdrawal, and Their Common Correlates

The second line of evidence we examined was whether CWB and withdrawal had similar or different relationships with a broad range of correlates (e.g., dispositions, organizational characteristics/perceptions, work performance behaviors, job attitudes, and demographics) in their nomological networks. The meta-analytic results are presented in Tables 6 (withdrawal) and 7 (CWB). In Table 8, we provide a side-by-side comparison of CWB and withdrawal’s corrected correlations with the common correlates.

Table 6

Meta-Analytic Results: Relationships Between Withdrawal Measures and Correlates

Correlate	N	k	r _m	SD_r	ρ	SD _ρ	% Var	CV₁₀	CV₉₀	CI_L	CI_U
Dispositions
Conscientiousness	2,774	11	−.21	.19	−.27	.23	9.77	−.57	.03	−.42	−.12
Extraversion	1,140	5	−.16	.10	−.20	.10	42.11	−.33	−.08	−.31	−.09
Emotional Stability	1,470	6	−.17	.11	−.22	.12	29.34	−.38	−.07	−.34	−.10
Agreeableness	863	4	−.12	.06	−.16	.00	100.00	−.16	−.16	−.23	−.08
Openness	506	3	−.08	.04	−.11	.00	100.00	−.11	−.11	−.18	−.05
Negative Affect (Withdrawal)	2,178	6	.29	.07	.37	.07	45.77	.29	.46	.30	.45
Negative Affect (Turnover Intentions)	1,073	4	.32	.05	.40	.00	100.00	.40	.40	.34	.46
Positive Affect	1,401	3	−.26	.02	−.32	.00	100.00	−.33	−.32	−.36	−.29
Self-Control	960	3	.05	.34	.06	.43	2.65	−.49	.61	−.44	.55
Social Desirability	1,322	5	−.13	.20	−.17	.24	9.41	−.47	.14	−.39	.06
Organization Characteristics
Stress	1,785	3	.23	.10	.28	.10	16.44	.15	.40	.15	.41
Conflict	2,029	4	.23	.11	.30	.13	14.66	.13	.47	.15	.44
Constraints	2,182	5	.34	.17	.42	.19	6.53	.17	.67	.24	.61
Procedural Justice	2,298	5	−.03	.30	−.03	.35	2.42	−.48	.41	−.35	.28
Distributive Justice	1,683	4	.07	.27	.08	.31	3.15	−.32	.49	−.23	.40
Interpersonal Justice	1,185	3	−.20	.07	−.24	.05	59.99	−.30	−.18	−.33	−.16
Organizational Support	750	3	−.41	.07	−.49	.06	49.76	−.57	−.41	−.59	−.39
Work Performance
OCB (Withdrawal)	4,411	15	−.13	.14	−.17	.17	15.89	−.38	.04	−.26	−.08
OCB (Turnover Intentions)	2,146	9	−.21	.13	−.26	.15	21.31	−.45	−.07	−.37	−.15
OCB (Absenteeism)	790	3	−.04	.05	−.05	.00	100.00	−.05	−.05	−.13	.02
Task Performance	1,784	7	.01	.26	.02	.30	5.93	−.37	.40	−.22	.25
Job Attitudes
Job Satisfaction (Withdrawal)	12,033	20	−.40	.20	−.49	.23	3.07	−.79	−.19	−.60	−.38
Job Satisfaction (Absenteeism)	6,410	4	−.21	.10	−.26	.11	6.27	−.41	−.12	−.38	−.14
Job Satisfaction (Turnover Intentions)	6,830	10	−.56	.14	−.68	.16	3.70	−.88	−.47	−.78	−.58
Commitment	2,860	8	−.24	.25	−.30	.30	3.95	−.68	.08	−.51	−.09
Demographic
Age (Withdrawal)	2,212	6	−.07	.09	−.08	.08	36.75	−.18	.02	−.16	.00
Age (Turnover Intentions)	741	4	−.06	.12	−.06	.11	35.46	−.20	.08	−.20	.07

Table 7

Meta-Analytic Results: Relationships Between CWB Measures and Correlates

Variable	N	k	r _m	SD_r	ρ	SD _ρ	% Var	CV₁₀	CV₉₀	CI_L	CI_U
Dispositions
Conscientiousness (CWB)	2,755	11	−.20	.18	−.26	.21	11.80	−.53	.01	−.40	−.13
Conscientiousness (CWBI)	1,051	4	−.20	.07	−.25	.04	75.46	−.31	−.20	−.34	−.17
Conscientiousness (CWBO)	1,582	6	−.30	.09	−.37	.08	40.30	−.48	−.27	−.46	−.29
Extraversion	1,140	5	−.16	.09	−.20	.08	53.02	−.30	−.10	−.30	−.10
Emotional Stability	1,470	6	−.19	.10	−.25	.10	37.60	−.38	−.12	−.35	−.14
Agreeableness	863	4	−.18	.11	−.24	.11	38.94	−.37	.10	−.37	−.10
Openness	506	3	−.09	.17	−.12	.20	19.65	−.38	.15	−.38	.15
Negative Affect (CWB)	4,284	7	.25	.13	.33	.16	7.92	.12	.54	.20	.45
Negative Affect (CWBI)	883	3	.27	.03	.34	.00	100.00	.34	.34	.29	.38
Negative Affect (CWBO)	883	3	.25	.03	.31	.00	100.00	.31	.31	.27	.36
Positive Affect	1,401	3	−.24	.04	−.30	.00	100.00	−.30	−.30	−.36	−.24
Self-Control	941	3	.10	.37	.13	.47	2.24	−.48	.73	−.42	.67
Social Desirability	1,190	4	−.06	.41	−.08	.52	1.99	−.75	.59	−.60	.45
Organization Characteristics
Stress	1,785	3	.28	.06	.34	.07	30.86	.26	.43	.25	.44
Conflict	2,029	4	.41	.04	.54	.02	82.57	.51	.57	.49	.59
Constraints	2,182	5	.42	.13	.52	.15	9.18	.33	.72	.38	.66
Procedural Justice	2,238	5	−.15	.09	−.18	.09	25.86	−.30	−.06	−.28	−.09
Distributive Justice	1,683	4	−.10	.07	−.12	.06	50.39	−.19	−.05	−.20	−.04
Interpersonal Justice	1,146	3	−.29	.05	−.35	.03	77.22	−.39	−.32	−.43	−.28
Organizational Support	750	3	−.25	.08	−.30	.06	55.93	−.38	−.22	−.41	−.19
Work Performance
OCB (CWB)	4,514	15	−.21	.19	−.27	.23	8.71	−.56	.02	−.39	−.15
OCB (CWBI)	2,016	7	−.03	.18	−.04	.21	10.28	−.31	.24	−.20	.13
OCB (CWBO)	2,226	8	−.12	.25	−.15	.30	5.65	−.53	.23	−.37	.06
Task Performance	1,774	7	.08	.23	.09	.27	7.24	−.26	.44	−.12	.31
Job Attitudes
Job Satisfaction (CWB)	14,139	21	−.24	.08	−.30	.09	18.37	−.42	−.18	−.34	−.25
Job Satisfaction (CWBI)	3,063	8	−.20	.08	−.24	.07	41.62	−.33	−.15	−.30	−.18
Job Satisfaction (CWBO)	8,542	12	−.21	.04	−.26	.03	69.32	−.29	−.22	−.29	−.23
Commitment	4,998	9	−.20	.09	−.25	.10	20.58	−.38	−.13	−.32	−.18
Demographic
Age (CWB)	4,465	8	−.18	.04	−.21	.00	100.00	−.21	−.21	−.24	−.17
Age (CWBI)	828	3	−.13	.05	−.15	.00	100.00	−.15	−.15	−.21	−.08
Age (CWBO)	1,055	4	−.16	.06	−.17	.00	100.00	−.17	−.17	−.23	−.11

Table 8

Meta-Analytic Results: Comparison of Relationships Among Counterproductive Work Behavior (CWB), Withdrawal, and Correlates

Common Correlates	CWB Correlation (ρ_c)	Withdrawal Correlation (ρ_w)	Difference \|ρ_c – ρ_w\|	z
Dispositions
Conscientiousness	−.26	−.27	.01	−0.30^a
Extraversion	−.20	−.20	.00	0.00^a
Emotional Stability	−.25	−.22	.03	0.57^a
Agreeableness	−.24	−.16	.08	1.30^a
Openness	−.12	−.11	.01	0.04^a
Negative Affect	.33	.37	.04	−1.29^a
Positive Affect	−.30	−.32	.02	−0.54^a
Self-Control	.13	.06	.07	1.17^a
Social Desirability	−.08	−.17	.09	−1.73^a
Organization Characteristics
Stress	.34	.28	.06	1.60^a
Conflict	.54	.30	.24	6.63^a
Constraints	.52	.42	.10	3.21^a
Procedural Justice	−.18	−.04	.14	4.21^a
Distributive Justice	−.12	.08	.04	0.85^a
Interpersonal Justice	−.35	−.24	.11	2.36^a
Organizational Support	−.30	−.49	.19	−3.57^a
Work Performance
OCB	−.27	−.17	.10	3.70^a
Task Performance	.09	.02	.07	1.87^a
Job Attitudes
Job Satisfaction	−.30	−.49	.19	−14.00^a
Commitment	−.25	−.30	.05	−1.80^a
Demographic
Age	−.21	−.08	.13	4.48^a

Note: ρ_c = corrected correlation between CWB and correlate; ρ_w = corrected correlation between withdrawal and correlate; |ρ_c – ρ_w | = absolute difference between the CWB-correlate and withdrawal-correlate correlations; z = the z statistic reflecting the significance of the difference between the corrected correlations within each moderator category; OCB = organizational citizenship behavior.

Indicates significant differences (zs ≥ ±1.96) between correlations (α = .05).

The results in Table 8 are revealing. First, there are some cases where CWB and withdrawal have significantly different relationships with the same correlate. Interestingly, the majority of these differences were for organizational characteristic variables (i.e., conflict, constraints, procedural and interpersonal justice, and organizational support) and job attitudes (e.g., job satisfaction), whereas none of them were for dispositional variables. We caution against overinterpretation of these differences, particularly because some of the sample sizes are small. Additionally, many of these differences are statistically significant but relatively small in practical terms (e.g., although statistically significantly different, CWB and withdrawal’s relationships with constraints are .52 and .42, respectively). Thus, perhaps most notable in Table 8 is the similar pattern and magnitudes of the relationships CWB and withdrawal have with the wide range of correlates. Even when CWB and withdrawal have significantly different relationships with a common correlate, the relationships are always in the same direction and somewhat similar in magnitude (e.g., although statistically significantly different, the relationships that CWB and withdrawal have with constraints, justice, organizational support, OCB, and job satisfaction are all in the moderate-to-strong range). Further illustrating this point, the average difference between CWB and withdrawal correlation magnitudes was only .08, which suggests there is generally a small difference across correlates (Berry, Carpenter, & Barratt, 2012). Additionally, the vector of CWB-correlate relationships (in Table 8) was correlated r = .91 with the vector of withdrawal-correlate relationships. Therefore, in answer to Research Question 2, CWB and withdrawal generally have very similar patterns and magnitudes of relationships with the variables in their nomological networks.

Discussion

Summary of Findings

Although it may have been assumed that CWB and withdrawal referred to redundant classes of employee behavior, this assumption had yet to be formally and empirically tested. As such, the present study provides much needed clarification of the empirical relationship between CWB and withdrawal. First, the analysis of the CWB-withdrawal relationship indicated that, when corrected for unreliability, overall CWB and withdrawal were strongly related, with corrected correlations ranging from .50 to .59. When additionally corrected for indirect range restriction, CWB and withdrawal were generally correlated about .70 or higher across most moderator categories. CWBO and withdrawal were especially strongly related, with the overall fully corrected CWBO-withdrawal relationship being .80; this relationship was even greater if the CWB-C (Spector et al., 2004) was used but was nonetheless substantial when other measures (e.g., R. J. Bennett & Robinson, 2000; Lehman & Simpson, 1992) were used. Supporting Hypothesis 1, these results mean that CWB and withdrawal are strongly related, but their relationship does not reach unity. Regardless, the fully corrected correlations are in the range thought of as suitable for a test-retest reliability of a single construct and are above the general range of correlations that previous meta-analyses have used to conclude that variables reflected a common construct (e.g., Berry et al., 2007; LePine et al., 2002). Thus, CWB and withdrawal likely reflect a common construct. Further, these strong bivariate correlations indicate that employees who engage in withdrawal behaviors are also likely to enact CWBs and vice versa. Thus, managers should be aware that employees’ enactment of withdrawal behaviors, such as absence and withholding effort, may signal that CWBs are also likely being enacted.

Second, we compared the relationships CWB and withdrawal had with a broad range of nomological correlates representing job attitudes, dispositions, organizational perceptions, and performance behaviors. We found that, in general, CWB and withdrawal had relationships with common correlates that were very similar (r = .91). Thus, the combination of the two lines of evidence examined in this study reveal that CWB and withdrawal are strongly related to each other and they have nearly identical patterns of relationships with theoretical correlates. As such, this evidence indicates CWB and withdrawal reflect a common construct and highlights a need to integrate the understanding of both CWB and withdrawal behavior.

Our moderator results also direct attention to the manner in which CWB and withdrawal are measured. For example, when supervisors or employees rated both behaviors, the CWB-withdrawal relationship was typically significantly larger than the use of different rating sources (e.g., self-supervisor). Plausible explanations for this finding could be either the presence of halo error or that raters, whether employees or supervisors, simply do not make the precise distinctions between CWB and withdrawal behaviors in the same way that some researchers or theories do. However, Berry, Carpenter, and Barratt (2012) also suggested that supervisors and coworkers may not be great judges of the enactment of CWB, particularly CWBO; this argument could also extend to withdrawal (e.g., Does a supervisor always know when an employee takes an extralong break or when an employee is psychologically withdrawing from work?). Indeed, supervisors and coworkers may not observe the extent of an employee’s enactment of different CWB or withdrawal behaviors and may base their judgments of employees’ enactment of these behaviors on other factors, such as employee traits (e.g., employees perceived as less agreeable or conscientious may be assumed to engage in more CWB and withdrawal).

As described above, the CWB-withdrawal relationship was also significantly influenced by the specific scale used to measure the behaviors. Importantly, the CWB-withdrawal relationship was, nevertheless, still strong across all scales. Since the most commonly used scales each contain items with overlapping content (see Table 2), future research that attempts to more cleanly measure CWB and withdrawal behaviors may be valuable in further understanding the CWB-withdrawal relationship. However, we note that the overlapping content may reflect the actual construct overlap between CWB and withdrawal. For example, if withdrawal behaviors should be considered CWBOs (organizationally targeted negative employee behaviors), then there should be content overlap between the measures.

Finally, we also found that CWB and withdrawal were much more strongly related when withdrawal was measured as a behavior (e.g., absence) than as turnover intentions. This suggests that turnover intentions may simply not be aligned with a broad, negative work behavior construct representing CWB and withdrawal. In the section that follows, we discuss pointed conceptual and operational reasons why turnover intentions are not likely to fit in an integrated conceptualization of CWB and withdrawal.

Integration of CWB and Withdrawal Structure

On the basis of the results of the present meta-analysis, we propose that the structure of CWB and withdrawal be integrated in a manner aligned with the hierarchical model that is depicted in Figure 1. This model is consistent with but also extends the hierarchical model of CWB proposed by Sackett and DeVore (2002) as well as the Spector et al. (2004) and Spector, Fox, Penney, Bruursema, Goh, and Kessler (2006) conceptualization of withdrawal as a facet of CWB and CWBO. In hierarchical construct models, each of the lower-order factors contains (a) shared variance that is driven by the higher-order factor(s) and (b) unique variance that is not related to the higher-order factor but is uniquely related to other variables. Additionally, the lower-order factor is most strongly related to the higher-order factor directly above it while (sometimes) being less related to the other lower-order or higher-order factors in the model. Such hierarchical models have been proposed for other constructs, such as cognitive ability (Carroll, 1993; e.g., verbal ability, which is a lower-order factor [facet] of the higher-order general cognitive ability factor, is uniquely predictive of performance on tasks with high verbal demands, but much of the variance in verbal ability is still accounted for by general cognitive ability). Importantly, a hierarchical model of CWB and withdrawal helps to explain why (a) there is such a strong intercorrelation between withdrawal and CWB (especially CWBO), (b) withdrawal and CWB have an overall similar pattern of relationships with their common correlates, while at the same time (c) withdrawal and CWB do exhibit some unique variance (i.e., have different relationships with some correlates).

Figure 1

Hierarchical Model Integrating Counterproductive Work Behavior and Withdrawal

Similar to Robinson and Bennett (1995), we conceptualize CWBI as composed of two lower-order factors: political deviance (encompassing relatively minor CWBIs) and personal aggression (encompassing relatively major, severe CWBIs). Per Spector et al. (2006), CWBO comprises a set of lower-order factors representing relatively active organizational-target CWBs (sabotage, production deviance, theft) as well as a lower-order factor representing withdrawal. Thus, withdrawal is a facet of CWBO. This greater connection of withdrawal to CWBO than to CWBI is partly based on and supported by their bivariate empirical relationships but is also based on the delineations of CWB and withdrawal’s theoretical similarities and differences (see Table 1 and O’Leary-Kelly et al., 2000). That is, both CWBOs and withdrawal behaviors (a) can be active or passive behaviors, (b) include both attempted and completed behaviors, (c) are enacted by organizational members, (d) include mostly physical behaviors, and (e) are directed toward the organization. Only the first three attributes above are shared by CWBI and withdrawal, which supports withdrawal as having a stronger theoretical link to, and being a facet of, CWBO.

Given the strong correlation between CWBO and withdrawal, another possibility could be that CWBO is a lower-order facet of withdrawal. However, we believe this is unlikely because CWBOs are negative employee behaviors that have their most direct effect on (i.e., target) the organization as opposed to individuals in the organization, while withdrawal behaviors are a narrower set of actions enacted to avoid work duties or the organization. Thus, it follows that all withdrawal behaviors are CWBOs in that they are negative employee behaviors that have their most direct effect on the organization, but not all CWBOs are withdrawal behaviors, as many CWBOs are relatively active behaviors that are unlikely to share withdrawal’s possible intent to avoid motive (e.g., theft, sabotage).

As mentioned above, and as should be noted in Figure 1, turnover intentions are not included in this hierarchical model. This is based on not only the empirical findings of this meta-analysis but also conceptual grounds. First, because turnover intentions are cognitions about whether to remain an organizational member, this is conceptually at odds with the remaining forms of CWB and withdrawal that are explicit behaviors. Turnover intentions are furthermore disconnected from both (a) CWB being defined as negative behaviors that are harmful to others or to the organization and (b) withdrawal as avoidance and disengagement from work. Though turnover intentions certainly may lead to actual turnover that is costly to organizations, the intention alone is neither harming the organization nor inherently signaling avoidance or disengagement. Therefore, we do not recommend regarding turnover intentions as part of the hierarchical model of CWB.

In this model, there is value in investigating any level of the hierarchy. For example, if a researcher or practitioner is interested in predicting only withdrawal, the hierarchical model and our findings suggest that there is variance in withdrawal that will be uniquely predicted by job satisfaction and commitment. However, if researchers or practitioners are more interested in higher levels of the hierarchy, such as understanding the determinants of the broad range of CWBOs, it is useful to know that withdrawal and the other CWBOs are similarly predicted by dispositional correlates. The current study’s findings regarding the CWB-withdrawal relationship as well as their respective nomological relationships provide important evidence in support of the hierarchical model. Further empirical tests will be useful for future refinements to this model as well as for theory building in the CWB and withdrawal domains.

Additional Theoretical and Practical Implications

That CWB and withdrawal reflect a broad negative behavior construct has important implications for future research. This broader level of abstraction is in line with Ajzen’s (1988) compatibility principle, which acknowledges that employees engage in different forms of behavior but that the core of these different behaviors is a common propensity, attitude, or disposition. Similar to the Harrison et al. (2006) attitude-engagement model, for example, our findings suggest that a broad negative behavior construct (i.e., CWB and withdrawal) should be more strongly predicted by a broad conceptualization of attitudes or traits. Thus, future theoretical and empirical work that establishes and examines these broad predictors is important for further advancing the negative work behavior domain.

Although our findings suggest that CWB and withdrawal have similar empirical relationships with important predictors and correlates, we believe that additional theoretical and empirical work is necessary to further understand whether these relationships with CWB and withdrawal have similar moderators and mediators. As just one example, we theoretically distinguished between CWB and withdrawal on the basis of the distinct underlying motives that may underlie each of these behaviors (e.g., intent to harm vs. avoid). Despite conceptual arguments that exist for this distinction (see Dalal et al., 2009), we were unable to empirically examine whether CWB and withdrawal are actually enacted for these distinct reasons. Vadera and Pratt (2013) recently argued that delineating the intentions or motives underlying different aspects of organizational crimes (of which CWB is a narrow facet) is necessary for clarifying and integrating the conceptual overlap across the numerous constructs that exist in these domains. We would add that these motives could be important factors that influence whether employees enact CWB or withdrawal in response to organizational situations or even on the basis of their dispositions. Pursuit of this line of research is likely to yield important practical insights as well. For example, though our results suggest CWB and withdrawal may be similarly affected by managerial practices and workplace interventions, this remains an empirical question. In sum, future research that explicitly conceptualizes and examines the specific motives for when employees enact CWB versus withdrawal will be important for further understanding the similarities and differences between CWB and withdrawal, for refining the hierarchical model described above, and for contributing to a better understanding of how CWB and withdrawal fit in the broader organizational crime literature.

The results of this meta-analysis also provide important theoretical and analytical considerations for further advancing the understanding of CWB and withdrawal. A limitation of the current study is that the primary studies composing the meta-analysis each focused on single “snapshots” of employee CWB and withdrawal (R. J. Bennett & Robinson, 2003), in which individual acts were recalled and reported. Though this is the norm of CWB and withdrawal measurement, such is likely a deficient measure of what is likely a complex relationship, especially considering that the important role of time appears to be ignored (Harrison & Martocchio, 1998; Hulin, 1991). CWB and withdrawal likely reflect a process of sequential behaviors and, therefore, should also be studied longitudinally at the within-person level of analysis (R. J. Bennett & Robinson; Judge, Scott, & Ilies, 2006; Yang & Diefendorff, 2009). Indeed, negative work behavior contains considerable within-person variance (Dalal et al., 2009; Judge et al.; Miner, Glomb, & Hulin, 2005), which indicates the need to consider the dynamic nature of and relationships among dispositional factors, situational variables, attitudes, and even the employee’s previous behaviors that likely influence whether CWB or withdrawal is enacted (Judge et al.). These processes are unlikely to be clearly detected from traditional between-person analyses (Dalal et al.; Yang & Diefendorff) and, thus, it should be richly informative to supplement between-person research on the CWB-withdrawal relationship with examinations of CWB and withdrawal at the within-person level of analysis.

For example, we encourage researchers to include CWB in investigations of the classic progression-of-withdrawal model (e.g., Herzberg, Mausnes, Peterson, & Capwell, 1957; Rosse, 1988; Rosse & Miller, 1984), which specifies sequential causal links from lateness to absence to turnover (Harrison et al., 2006). CWB may certainly reflect an added set of behaviors an unhappy employee may enact before exiting the organization. Process models depicting behavioral spirals (e.g., incivility; Andersson & Pearson, 1999) likely fit here as well. Employees may start with minor forms of withdrawal and/or CWB and progress over time to more severe behaviors intended to relieve or express their distress or dissatisfaction, particularly if the preceding mild behaviors do not sufficiently relieve the distress (Rosse & Miller). In sum, we expect that considering the role of time in the CWB-withdrawal relationship will provide important insights into employees’ enactment of these negative work behaviors.

Conclusion

The present meta-analysis clarifies the empirical relationship between CWB and withdrawal. Our findings demonstrated that CWB and withdrawal (a) are strongly intercorrelated when fully corrected for unreliability and indirect range restriction and also across different measurement moderators; and (b) have, in general, very similar patterns of relationships with important correlates representing dispositions, job attitudes, performance behaviors, and organizational context variables. Altogether, on the basis of the meta-analytic evidence, we propose that CWB and withdrawal be integrated into a common hierarchical model that represents their large degree of shared variance as well as their uniqueness.

Footnotes

Acknowledgements

This article was accepted under the editorship of Deborah E. Rupp. The authors would like to acknowledge Daniel S. Whitman for comments on a previous version of the manuscript and Christen L. Dovalina and Margaret T. Horner for assistance with coding.

Notes

References

Ajzen

1985. From intentions to actions: A theory of planned behavior. In Kuhl

Beckmann

(Eds.), Action-control: From cognition to behavior: 11-39. Heidelberg, Germany: Springer.

Ajzen

1987. Attitudes, traits, and actions: Dispositional prediction of behavior in personality and social psychology. In Berkowitz

(Ed.), Advances in experimental social psychology, vol. 20: l-63. New York: Academic Press.

Ajzen

1988. Attitudes, personality, and behavior. Homewood, IL: Dorsey Press.

Andersson

L. M.

Pearson

C. M.

1999. Tit for tat? The spiraling effect of incivility in the workplace. Academy of Management Review, 24: 452-471.

*Avey

J. B.

Luthans

Youssef

C. M.

2010. The additive value of positive psychological capital in predicting work attitudes and behaviors. Journal of Management, 36: 430-452.

*Bauer

J. A.

2011. The role of discrete emotions in predicting counterproductive work behavior. Unpublished master’s thesis, University of South Florida, Tampa.

*Bayram

Gursakal

Bilgel

2009. Counterproductive work behavior among white-collar employees: A study from Turkey. International Journal of Selection and Assessment, 17: 180-188.

*Bennett

J. B.

Lehman

W. E. K.

1996. Alcohol, antagonism, and witnessing violence in the workplace: Drinking climates and social alienation-integration. In VandenBos

G. R.

Bulatao

E. Q.

(Eds.), Violence on the job: Identifying risks and developing solutions: 105-152. Washington, DC: American Psychological Association.

*Bennett

R. J.

Robinson

S. L.

2000. Development of a measure of workplace deviance. Journal of Applied Psychology, 85: 349-360.

10.

Bennett

R. J.

Robinson

S. L.

2003. The past, present and future of workplace deviance research. In Greenberg

(Ed.), Organizational behavior: The state of the science (2nd ed.): 247-281. Mahwah, NJ: Erlbaum.

11.

*Bernerth

J. M.

Walker

H. J.

2012. Reexamining the workplace justice to outcome relationship: Does frame of reference matter? Journal of Management Studies, 49: 945-969.

12.

Berry

C. M.

Carpenter

N. C.

Barratt

C. L.

2012. Do other-reports of counterproductive work behavior provide an incremental contribution over self-reports? A meta-analytic comparison. Journal of Applied Psychology, 97: 613-636.

13.

Berry

C. M.

Lelchook

A. M.

Clark

M. A.

2012. A meta-analysis of the interrelationships between employee lateness, absenteeism, and turnover: Implications for models of withdrawal behavior. Journal of Organizational Behavior, 33: 678-699.

14.

Berry

C. M.

Ones

D. S.

Sackett

P. R.

2007. Interpersonal deviance, organizational deviance, and their common correlates: A review and meta-analysis. Journal of Applied Psychology, 92: 410-424.

15.

*Bolin

Heatherley

2001. Predictors of employee deviance: The relationship between bad attitudes and bad behavior. Journal of Business and Psychology, 15: 405-418.

16.

*Bolton

L. R.

Becker

L. K.

Barber

L. K.

2010. Big Five trait predictors of differential counterproductive work behavior dimensions. Personality and Individual Differences, 49: 537-541.

17.

*Bowling

N. A.

Burns

G. N.

Beehr

T. A.

2010. Productive and counterproductive attendance behavior: An examination of early and late arrival to and departure from work. Human Performance, 23: 305-322.

18.

Bozeman

D. P.

Perrewé

P. L.

2001. The effect of item content overlap on Organizational Commitment Questionnaire–turnover cognitions relationships. Journal of Applied Psychology, 86: 161-173.

19.

*Bruursema

Kessler

S. R.

Spector

P. E.

2011. Bored employees misbehaving: The relationship between boredom and counterproductive work behavior. Work and Stress, 25: 93-107.

20.

Buss

A. H.

1961. The psychology of aggression. New York: Wiley.

21.

Carpenter

N. C.

Newman

D. A.

Arthur

Jr.

2011. What do work performance items measure? A substantive validity examination. Poster presented at the 26th annual meeting of the Society for Industrial and Organizational Psychology, Chicago.

22.

Carroll

J. B.

1993. Human cognitive abilities: A survey of factor-analytic studies. New York: Cambridge University Press.

23.

*Chen

P. Y.

Spector

P. E.

1992. Relationships of work stressors with aggression, withdrawal, theft and substance use: An exploratory study. Journal of Occupational and Organizational Psychology, 65: 177-184.

24.

Cohen

1992. A power primer. Psychological Bulletin, 112: 155-159.

25.

Cohen-Charash

Spector

P. E.

2001. The role of justice in organizations: A meta-analysis. Organizational Behavior and Human Decision Processes, 86: 278-321.

26.

Colquitt

J. A.

Conlon

D. E.

Wesson

M. J.

Porter

C. O.

K. Y.

2001. Justice at the millennium: A meta-analytic review of 25 years of organizational justice research. Journal of Applied Psychology, 86: 425-445.

27.

Connolly

J. J.

Viswesvaran

2000. The role of affectivity in job satisfaction: A meta-analysis. Personality and Individual Differences, 29: 265-281.

28.

*Crede

Chernyshenko

O. S.

Stark

Dalal

R. S.

Bashshur

2007. Job satisfaction as mediator: An assessment of job satisfaction’s position within the nomological network. Journal of Occupational and Organizational Psychology, 80: 515-538.

29.

*Cropanzano

Howes

J. C.

Grandey

A. A.

Toth

1997. The relationship of organizational politics and support to work behaviors, attitudes, and stress. Journal of Organizational Behavior, 18: 159-180.

30.

Dalal

R. S.

2005. A meta-analysis of the relationship between organizational citizenship behavior and counterproductive work behavior. Journal of Applied Psychology, 90: 1241-1255.

31.

Dalal

R. S.

Lam

Weiss

H. M.

Welch

E. R.

Hulin

C. L.

2009. A within-person approach to work behavior and performance: Concurrent and lagged citizenship-counterproductivity associations, and dynamic relationships with affect and overall job performance. Academy of Management Journal, 52: 1051-1065.

32.

*Demuth

R. L. F.

2007. Antecedents and consequences of value expression in the workplace. Unpublished doctoral dissertation, Northern Illinois University, DeKalb.

33.

*Donovan

M. A.

2000. Cognitive, affective, and satisfaction variables as predictors of organizational behaviors: A structural equation modeling examination of alternative models. Unpublished doctoral dissertation, University of Illinois at Urbana-Champaign.

34.

Edwards

J. R.

Berry

J. W.

2010. The presence of something or the absence of nothing: Increasing theoretical precision in management research. Organizational Research Methods, 13: 668-689.

35.

*Fallon

J. D.

Avis

J. M.

Kudisch

J. D.

Gornet

T. P.

Frost

2000. Conscientiousness as a predictor of productive and counterproductive behaviors. Journal of Business and Psychology, 15: 339-349.

36.

*Ferris

D. L.

Brown

D. J.

Berry

J. W.

Lian

2008. The development and validation of the Workplace Ostracism Scale. Journal of Applied Psychology, 93: 1348-1366.

37.

*Gettinger

S. K.

2009. Sexual harassment: Implications for counterproductive work behavior. Unpublished doctoral dissertation, University of Houston, TX.

38.

Ghiselli

E. E.

Campbell

J. P.

Zedeck

1981. Measurement theory for the behavioral sciences. San Francisco: Freeman.

39.

*Goh

2007. An attributional analysis of counterproductive work behavior (CWB) in response to occupational stress. Unpublished doctoral dissertation, University of South Florida, Tampa.

40.

Gottfredson

M. R.

1990. A general theory of crime. Stanford, CA: Stanford University Press.

41.

*Greenbaum

R. L.

Mawritz

M. B.

Mayer

D. M.

Priesemuth

2013. To act out, to withdraw, or to constructively resist? Employee reactions to supervisor abuse of customers and the moderating role of employee moral identity. Human Relations, 66: 925-950.

42.

*Gruys

M. L.

Sackett

P. R.

2003. Investigating the dimensionality of counterproductive work behavior. International Journal of Selection and Assessment, 11: 30-42.

43.

*Guo

2012. Counterproductive work behaviors, Confucian values, and production deviance: The mediating effect of job satisfaction. Social Behavior and Personality, 40: 1045-1056.

44.

*Hanisch

K. A.

Hulin

C. L.

1990. Job attitudes and organizational withdrawal: An examination of retirement and other voluntary withdrawal behaviors. Journal of Vocational Behavior, 37: 60-78.

45.

Hanisch

K. A.

Hulin

C. L.

1991. General attitudes and organizational withdrawal: An evaluation of a causal model. Journal of Vocational Behavior, 39: 110-128.

46.

Hanisch

K. A.

Hulin

C. L.

Roznowski

1998. The importance of individuals’ repertoires of behaviors: The scientific appropriateness of studying multiple behaviors and general attitudes. Journal of Organizational Behavior, 19: 463-480.

47.

Harrison

D. A.

2002. Meaning and measurement of work role withdrawal: Current controversies and future fallout from changing information technology. In Koslowsky

Krausz

(Eds.), Voluntary employee withdrawal and inattendance: A current perspective: 95-131. New York: Kluwer Academic/Plenum.

48.

Harrison

D. A.

Martocchio

J. J.

1998. Time for absenteeism: A 20-year review of origins, offshoots, and outcomes. Journal of Management, 24: 305-350.

49.

Harrison

D. A.

Newman

D. A.

2013. Absence, lateness, turnover, and retirement: Narrow and broad understandings of withdrawal behavior and behavioral engagement. In Schmitt

Highhouse

(Eds.), Handbook of psychology: Industrial and organizational psychology, vol. 12: 262-291. Hoboken, NJ: Wiley.

50.

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.

51.

Hauenstein

N. M.

McGonigle

Flinder

S. W.

2001. A meta-analysis of the relationship between procedural justice and distributive justice: Implications for justice research. Employee Responsibilities and Rights Journal, 13: 39-56.

52.

Hershcovis

M. S.

Turner

Barling

Arnold

K. A.

Dupre

K. E.

Inness

LeBlanc

M. M.

Sivanathan

2007. Predicting workplace aggression: A meta-analysis. Journal of Applied Psychology, 92: 228-238.

53.

Herzberg

Mausnes

Peterson

R. O.

Capwell

D. F.

1957. Job attitudes: Review of research and opinion. Oxford, England: Psychological Service of Pittsburgh.

54.

Higgins

C. A.

Judge

T. A.

Ferris

G. R.

2003. Influence tactics and work outcomes: A meta-analysis. Journal of Organizational Behavior, 24: 89-106

55.

Hogan

1989. How to measure reliability. Journal of Applied Psychology, 74: 273-279.

56.

Hollinger

R. C.

Clark

J. P.

1983. Theft by employees. Lexington, MA: Lexington Books.

57.

Holtom

B. C.

Burton

J. P.

Crossley

C. D.

2012. How negative affectivity moderates the relationship between shocks, embeddedness, and worker behaviors. Journal of Vocational Behavior, 80: 434-443.

58.

Hom

P. W.

Griffeth

R. W.

1995. Employee turnover. Cincinatti, OH: South-Western.

59.

Hulin

C. L.

1991. Adaptation, persistence, and commitment in organizations. In Dunnette

M. D.

Hough

L. M.

(Eds.), Handbook of industrial and organizational psychology (2nd ed.), vol. 2: 445-505. Palo Alto, CA: Consulting Psychologists Press.

60.

Hulin

C. L.

Roznowski

Hachiya

1985. Alternative opportunities and withdrawal decisions: Empirical and theoretical discrepancies and an integration. Psychological Bulletin, 97: 233-250.

61.

Hunt

S. T.

1996. Generic work behavior: An investigation into the dimensions of entry-level, hourly job performance. Personnel Psychology, 49: 51-83.

62.

Hunter

J. E.

1983. Test validation for 12,000 jobs: An application of job classification and validity generalization analysis to the General Aptitude Test Battery. Test research report no. 45, U.S. Employment Service, U.S. Department of Labor, Washington, DC.

63.

Hunter

J. E.

Schmidt

F. L.

2004. Methods of meta-analysis: Correcting error and bias in research findings (2nd ed.). New York: Russell Sage Foundation.

64.

Hunter

J. E.

Schmidt

F. L.

2006. Implications of direct and indirect range restriction for meta-analysis methods and findings. Journal of Applied Psychology, 91: 594-612.

65.

*Jackson

C. L.

Colquitt

J. A.

Wesson

M. J.

Zapata-Phelan

C. P.

2006. Psychological collectivism: A measurement validation and linkage to group member performance. Journal of Applied Psychology, 91: 884-899.

66.

James

L. R.

McIntyre

M. D.

Glisson

C. A.

Green

P. D.

Patton

T. W.

LeBreton

J. M.

Frost

B. C.

Russell

S. M.

Sablynski

C. J.

Mitchell

T. R.

Williams

L. J.

2005. A conditional reasoning measure for aggression. Organizational Research Methods, 8: 69-99.

67.

*Jensen

J. M.

Opland

R. A.

Ryan

A. M.

2010. Psychological contracts and counterproductive work behaviors: Employee responses to transactional and relational breach. Journal of Business Psychology, 25: 555-568. doi:10.1007/s10869-009-9148-7

68.

Johns

1987. Understanding and managing absence from work. In Dolan

S. L.

Schuler

R. S.

(Eds.), Canadian readings in personnel and human resource management: 324-335. St. Paul, MN: West.

69.

Johns

1997. Contemporary research on absence from work: Correlates, causes, and consequences. International Review of Industrial and Organizational Psychology, 12: 115-174.

70.

*Jones

D. A.

2005. Toward a better understanding of fairness in the workplace: Attitude strength, predictive asymmetry, and the revenge motive. Unpublished doctoral dissertation, University of Calgary, Alberta, Canada.

71.

Judge

T. A.

Scott

B. A.

Ilies

2006. Hostility, job attitudes, and workplace deviance: Test of a multilevel model. Journal of Applied Psychology, 91: 126-138.

72.

*Khan

A. K.

Quratulain

Crawshaw

J. R.

2012. The mediating role of discrete emotions in the relationship between injustice and counterproductive work behaviors: A study in Pakistan. Journal of Business Psychology, 28: 49-61. doi:10.1007/s10869-012-9269-2

73.

*Kinkade

K. M.

Morelli

N. A.

Downey

S. N.

Hoffman

B. J.

Eby

L. T.

2012. What if they don’t leave? Prolonged turnover intentions and CWB. Poster presented at the 26th annual meeting of the Society for Industrial and Organizational Psychology, San Diego, CA.

74.

*Krischer

M. M.

Penney

L. M.

Hunter

E. M.

2010. Can counterproductive work behaviors be productive? CWB as emotion-focused coping. Journal of Occupational Health Psychology, 15: 154-166.

75.

Lee

R. T.

Ashforth

B. E.

1996. A meta-analytic examination of the correlates of the three dimensions of job burnout. Journal of Applied Psychology, 81: 123-133.

76.

*Lehman

W. E. K.

Simpson

D. D.

1992. Employee substance use and on-the-job behaviors. Journal of Applied Psychology, 77: 309-321.

77.

LePine

J. A.

Erez

Johnson

D. E.

2002. The nature of dimensionality of organizational citizenship behavior: A critical review and meta-analysis. Journal of Applied Psychology, 87: 52-65.

78.

Marcus

Schuler

2004. Antecedents of counterproductive behavior at work: A general perspective. Journal of Applied Psychology, 89: 647-660.

79.

*Marcus

Schuler

Quell

Hümpfer

2002. Measuring counterproductivity: Development and initial validation of a German self-report questionnaire. International Journal of Selection and Assessment, 10: 18-35.

80.

*Marrs

M. E. M.

2000. Antecedents and outcomes of verbal aggression in the workplace. Unpublished doctoral dissertation, University of Missouri, Columbia.

81.

Mathieu

J. E.

Zajac

D. M.

1990. A review and meta-analysis of the antecedents, correlates, and consequences of organizational commitment. Psychological Bulletin, 108: 171-194.

82.

Miner

A. G.

Glomb

T. M.

Hulin

2005. Experience sampling mood and its correlates at work. Journal of Occupational and Organizational Psychology, 78: 171-193.

83.

*Mitchell

M. S.

2006. Understanding employees’ behavioral reactions to aggression in organizations. Unpublished doctoral dissertation, University of Central Florida, Orlando.

84.

Murphy

K. R.

1989. Dimensions of job performance. In Dillon

Pellingrino

(Eds.), Testing: Applied and theoretical perspectives: 218-247. New York: Praeger.

85.

Murphy

K. R.

1990. Job performance and productivity. In Murphy

K. M.

Saal

F. E.

(Eds.), Psychology in organizations: Integrating science and practice: 157-176. Hillsdale, NJ: Erlbaum.

86.

Neuman

J. H.

Baron

R. A.

1997. Aggression in the workplace. In Giacalone

Greenberg

(Eds.), Antisocial behavior in organizations: 37-67. Thousand Oaks, CA: Sage.

87.

Neuman

J. H.

Baron

R. A.

2005. Aggression in the workplace: A social psychological perspective. In Fox

Spector

P. E.

(Eds.), Counterproductive work behavior: Investigations of actors and targets: 13-40. Washington, DC: American Psychological Association.

88.

*Ntayi

J. M.

2005. Work ethic, locus of control and sales force task performance. Journal of African Business, 6: 155-176.

89.

O’Leary-Kelly

A. M.

Duffy

M. K.

Griffin

R. W.

2000. Construct confusion in the study of antisocial work behavior. In Ferris

(Ed.), Research in personnel and human resources management, vol. 18: 275-303. Greenwich, CT: JAI Press.

90.

Ones

D. S.

Viswesvaran

Schmidt

F. L.

2003. Personality and absenteeism: A meta-analysis of integrity tests. European Journal of Personality, 17: 519-538.

91.

*Peng

2012. Counterproductive work behavior among Chinese knowledge workers. International Journal of Selection and Assessment, 20: 119-138.

92.

Podsakoff

N. P.

LePine

J. A.

LePine

M. A.

2007. Differential challenge stressor–hindrance stressor relationships with job attitudes, turnover intentions, turnover, and withdrawal behavior: A meta-analysis. Journal of Applied Psychology, 92: 438-454.

93.

Raju

N. S.

Brand

P. A.

2003. Determining the significance of correlations corrected for unreliability and range restriction. Applied Psychological Measurement, 27: 52-71.

94.

Rhoades

Eisenberger

2002. Perceived organizational support: A review of the literature. Journal of Applied Psychology, 87: 698-714.

95.

*Rich

B. L.

2006. Job engagement: Construct validation and relationships with job satisfaction, job involvement, and intrinsic motivation. Unpublished doctoral dissertation, University of Florida, Tampa.

96.

*Richards

D. A.

Schat

A. C. H.

2011. Attachment at (not to) work: Applying attachment theory to explain individual behavior in organizations. Journal of Applied Psychology, 96: 169-182.

97.

Robinson

S. L.

Bennett

R. J.

1995. A typology of deviant workplace behaviors: A multidimensional scaling study. Academy of Management Journal, 38: 555-572.

98.

Robinson

Greenberg

1999. Employees behaving badly: Dimensions, determinants, and dilemmas in the study of workplace deviance. In Rousseau

D. M.

Cooper

(Eds.), Trends in organizational behavior, vol. 5: 1-23. New York: Wiley.

99.

*Robinson

S. L.

O’Leary-Kelly

A. M.

1998. Monkey see, monkey do: The influence of work groups on the antisocial behavior of employees. Academy of Management Journal, 41: 658-672.

100.

Rosse

J. G.

1988. Relations among lateness, absence, and turnover: Is there a progression of withdrawal? Human Relations, 41: 517-531.

101.

Rosse

J. G.

Miller

H. E.

1984. Relationship between absenteeism and other employee behaviors. In Goodman

P. S.

Atkin

R. S.

(Eds.), Absenteeism: New approaches to understanding, measuring, and managing employee absence: 194-228. San Francisco: Jossey-Bass.

102.

Rotundo

Spector

P. E.

2010. Counterproductive work behavior and withdrawal. In Farr

J. L.

Tippins

N. T.

(Eds.), Handbook of employee selection: 489-511. New York: Routledge/Taylor & Francis Group.

103.

Rusbult

C. E.

Farrell

Rogers

Mainous

A. G.

1988. Impact of exchange variables on exit, voice, loyalty, and neglect: An integrative model of responses to declining job satisfaction. Academy of Management Journal, 31: 599-627.

104.

Sackett

P. R.

DeVore

C. J.

2002. Counterproductive behaviors at work. In Anderson

Ones

D. S.

Sinangil

H. K.

Viswesvaran

(Eds.), Handbook of industrial, work, and organizational psychology, vol. 1: 145-164. London: Sage.

105.

*Shoss

M. K.

Shoss

B. L.

2012. Check-up time: A closer look at physical symptoms in occupational health research. Stress and Health, 28: 193-201.

106.

Skarlicki

D. P.

Folger

1997. Retaliation in the workplace: The roles of distributive, procedural, and interactional justice. Journal of Applied Psychology, 82: 434-443.

107.

Spector

P. E.

Bauer

J. A.

Fox

2010. Measurement artifacts in the assessment of counterproductive work behavior and organizational citizenship behavior: Do we know what we think we know? Journal of Applied Psychology, 95: 781-790.

108.

Spector

P. E.

Fox

2002. An emotion-centered model of voluntary work behavior: Some parallels between counterproductive work behavior and organizational citizenship behavior. Human Resource Management Review, 12: 269-292.

109.

Spector

P. E.

Fox

2005. The stressor-emotion model of counterproductive work behavior. In Fox

Spector

P. E.

(Eds.), Counterproductive work behavior: Investigations of actors and targets: 151-174. Washington, DC: American Psychological Association.

110.

Spector

P. E.

Fox

Penney

L. M.

Bruursema

Goh

Kessler

2004. Development of the Counterproductive Work Behavior Checklist (CWB-C). Unpublished manuscript, University of South Florida, Tampa.

111.

*Spector

P. E.

Fox

Penney

L. M.

Bruursema

Goh

Kessler

2006. The dimensionality of counterproductivity: Are all counterproductive behaviors created equal? Journal of Vocational Behavior, 68: 446-460.

112.

Spector

P. E.

Jex

S. M.

1998. Development of four self-report measures of job stressors and strain: Interpersonal Conflict at Work Scale, Organizational Constraints Scale, Quantitative Workload Inventory, and Physical Symptoms Inventory. Journal of Occupational Health Psychology, 3: 356-367.

113.

Stone-Romero

E. F.

Alvarez

Thompson

L. F.

2009. The construct validity of conceptual and operational definitions of contextual performance and related constructs. Human Resource Management Review, 19: 104-116.

114.

*Thau

Crossley

Bennett

R. J.

Sczesny

2007. The relationship between trust, attachment, and antisocial work behaviors. Human Relations, 60: 1155-1179.

115.

*Tobares

2009. An examination of affect and job satisfaction as mediators in the relationship between personal mastery and work behaviors. Unpublished master’s thesis, University of Texas at San Antonio.

116.

Vadera

A. K.

Pratt

M. G.

2013. Love, hate, ambivalence, or indifference? A conceptual examination of workplace crimes and organizational identification. Organization Science, 24: 172-188.

117.

*Villanueva

L. S.

2006. An examination of the role of self-control in the prediction of counterproductive work behaviors: Does cognition matter? Unpublished doctoral dissertation, University of Houston, TX.

118.

Viswesvaran

Ones

D. S.

2000. Measurement error in “Big Five factors” personality assessment: Reliability generalization across studies and measures. Educational and Psychological Measurement, 60: 224-235.

119.

Viswesvaran

Ones

D. S.

Schmidt

F. L.

1996. Comparative analysis of the reliability of job performance ratings. Journal of Applied Psychology, 81: 557-574.

120.

*Wei

2013. Tit for tat? Abusive supervision and counterproductive work behaviors: The moderating effects of locus of control and perceived mobility. Asia Pacific Journal of Management, 30: 281-296. doi:10.1007/s10490-011-9251-y

121.

Weiss

H. M.

Cropanzano

1996. Affective events theory: A theoretical discussion of the structure, causes, and consequences of affective experiences at work. Research in Organizational Behavior, 18: 1-74.

122.

*Whinghter

L. J.

2006. Employed youth: An exploration of the relationship between job quality and counterproductivity. Unpublished doctoral dissertation, Bowling Green State University, Ohio.

123.

Yang

Diefendorff

J. M.

2009. The relations of daily counterproductive workplace behavior with emotions, situational antecedents, and personality moderators: A diary study in Hong Kong. Personnel Psychology, 62: 259-295.