When employees purposefully do work incorrectly or slowly, come to work late, steal from the organization, or behave in rude and disrespectful ways to others, they are said to be engaging in counterproductive work behavior (CWB). More specifically, CWB refers to a broad class of intentional behaviors that run counter to the legitimate interests of the organization or its members (Sackett 2002; Spector et al. 2006a). Financial estimates point to billions of dollars in stolen property and lost productivity resulting from these behaviors (see Mount et al. 2006). Just as important, these behaviors arguably have the potential to erode the ethical and social landscapes of organizations.

What motivates individuals to engage in these potentially harmful acts? Bushman et al. (2001) suggested that individuals engage in harmful behaviors because they view them as a way to feel better in response to provocation. In their experimental study, Bushman et al. (2001) found that participants were less likely to respond to provocation with aggression after taking a pill that they believed “froze” their emotions and therefore eliminated any value associated with engaging in aggressive behaviors for emotion regulation. CWB theorists have likewise proposed that employees respond to acts of provocation at work (i.e., workplace stressors) with CWB because they anticipate emotional benefits (i.e., these behaviors will make them feel better, Bies and Tripp 1996; Krischer et al. 2010; Spector and Fox 2002). In one of the few empirical studies to investigate this claim, Jones (2009) found that a composite measure including desire for revenge and anticipation that revenge would make one feel better predicted CWB in response to workplace injustice (also Hung et al. 2009).

Although both stressor-based models of CWB (Spector and Fox 2002) and theoretical perspectives on coping (Folkman and Lazarus 1985) suggest that stressors trigger the enactment of behavior aimed at coping, few studies have explicitly characterized CWB as a coping strategy or investigated whether (and more importantly when) employees view CWB as such. Examining these beliefs, which we refer to as CWB coping instrumentalities, is critical for developing insight into the factors that influence the volitional processes driving CWB. More generally, instrumentalities refer to beliefs (i.e., anticipations) that a given behavior will result in a (desired) outcome (Pinder 2008). Instrumentalities are a critical component of both rational (e.g., Naylor et al. 1980) and descriptive (e.g., Beach and Mitchell 1987) models of behavioral choice and have been shown in the broader work motivation literature to have generative effects on both intentions and subsequent behavior (Van Eerde and Thierry 1996). Such parallels are important given that CWB is by definition intentional and discretionary and, accordingly, should be enacted because individuals anticipate some benefit in doing so. Moreover, coping research suggests that individuals choose coping strategies that they believe will be effective (Lazarus and Folkman 1984). We suggest that CWB coping instrumentalities serve as a motivational “linking pin” between traits or situational perceptions and subsequent CWB. As such, investigating coping instrumentalities for CWBs may yield theoretical clarity and parsimony (e.g., Le et al. 2010), thus advancing the understanding of their motivational impetus in a manner consistent across the coping, CWB, and motivation literatures.

We aim to address two fundamental questions regarding employees’ beliefs about the coping instrumentality of CWB. First, to what extent are CWB instrumentalities a function of (a) ongoing, situation-specific processes, (b) more stable factors (e.g., personality and beliefs), and (c) the interaction between situation and person factors? Second, and relatedly, what are the personal and situational factors that lead employees to believe that CWBs are instrumental responses to workplace stressors? These questions are inherently multilevel (i.e., within- and between-person) and they extend arguments that employees perform CWB in response to negative work situations (Spector and Fox 2002) by considering particular situational and employee characteristics that influence CWB coping instrumentalities. Stated otherwise, considering within- and between-person variance in individuals’ beliefs regarding the coping value of CWB sheds light on the degree to which certain individuals hold scripts and schemas dictating the general usefulness of CWB in response to negative workplace events, as well as the extent to which instrumentalities are contingent on more specific features of the situation.

In order to address these questions, we leverage Lazarus and Folkman’s (1984) transactional model of coping to identify situational (i.e., within-person) and stable individual (i.e., between-person) predictors of employees’ CWB coping instrumentalities in response to workplace stressors. We tested our predictions by employing a within-person scenario design wherein we asked employees to indicate the likelihood that engaging in three representative CWBs (stealing from one’s employer, decreasing work effort, and being rude to another) would make them feel better in response to various workplace stressors found in prior literature to predict CWB. Presenting employees with multiple scenarios in this manner allowed us to (a) evaluate the extent to which CWB coping instrumentalities are invariant across situations (i.e., vary between persons—e.g., Bowling and Eschleman 2010) or change based on the specifics of the given situation (i.e., vary within person) and (b) examine whether stressor source influences the specific type of CWB that employees will perceive to have instrumental value, both of which help to address calls for researchers to pay greater attention to levels of analysis in CWB research (Yang and Diefendorff 2009). This type of scenario methodology is widely employed in studies involving behavioral judgment, especially those that examine ethically questionable and low base-rate behaviors. Of particular note, it allowed us to standardize the social situation(s) of interest while providing respondents with specific situations from which to make decisions (Alexander and Becker 1978; Schminke et al. 1997).

Examination of the perceived value of CWB as a coping mechanism has important implications not only for understanding CWB as a coping response, but also for the continued use of coping as a theoretical lens through which to generate hypotheses about these behaviors. Quite simply, it is critical to empirically evaluate the fundamental assumption that employees view CWB as a viable emotion-focused coping tactic. Also, given that employees tend to view CWBs as immoral (Cohen et al. 2013), our work contributes to the study of ethical decision-making more broadly by (a) positioning CWB as unethical or ethically questionable behaviors that individuals may judge to have personal value for coping with stressors and (b) suggesting coping theory and its constituent components as a framework through which to understand when such judgments are likely to surface. Finally, our within- and between-person approach has implications for decisions regarding the level(s) at which interventions may be directed. In the following sections, we explicate CWB as an emotion-focused coping strategy and integrate research on coping and CWB to predict that perceptions of control, stressor source, and prior CWB enactment serve as proximal predictors of CWB instrumentalities. For illustrative purposes, the relationships investigated are displayed in Fig. 1.

Fig. 1
figure 1

Hypothesized relationships

Full size image

CWB and Emotion-Focused Coping

CWB has typically been viewed as the manifestation of frustration and negative workplace experiences (Fox and Spector 1999; Neuman and Baron 2005), especially when stressors are attributed to actions of an external actor (i.e., the organization or another employee; Martinko et al. 2002). However, CWB is not simply a reaction to negative workplace experiences; rather, it is a volitional choice in response to these events. Spector and Fox (2002) argued that negative affect arising from stressors stimulates intentions to engage in behaviors that individuals anticipate will be instrumental in reducing this affect and further strain. Similarly, many authors have pointed out parallels between CWB and emotion-focused coping, consistent with the idea that employees’ personal efforts to cope with stressors may be either counterproductive or productive depending on the viewpoint (e.g., Krischer et al. 2010; Penney and Spector 2008; Reynolds et al. 2015). For example, Krischer et al. (2010, p. 155) note that withdrawal (e.g., taking longer breaks than allowed) and passive aspects of production deviance (e.g., working slowly) “may reflect attempts by employees to limit their exposure to stressful situations and prevent subsequent strain” as well as attempts to let initially strong negative emotions subside. Passive interpersonal CWB (e.g., avoiding others, not responding to emails) might allow individuals to combat threats in the same manner. Further, acts of CWB (e.g., theft, sabotage, mistreating others) might also reflect employees’ attempts to regulate negative affect by enacting revenge against the party responsible for causing the stressor (Bies and Tripp 1996; Folger and Skarlicki 2005), which is consistent with social psychological research demonstrating that individuals believe that retaliation against offending others will improve their mood (Bushman et al. 2001).

Within-Person (Situational) Drivers of CWB Coping Instrumentalities

While studies in the CWB literature have used Lazarus and colleagues’ (Folkman and Lazarus 1985; Lazarus 1991) transactional model of stress and coping to predict CWB as a coping response to workplace stressors (e.g., Bowling and Eschleman 2010; Rodell and Judge 2009), less attention has been paid to the process through which individuals choose to engage in CWB and, subsequently, the factors that influence these decisions. These are critical omissions given that transactional theory predicts coping choices are selected based on an evaluative process that is influenced by both person factors and situational factors that shape how one views or appraises the demands facing him/her (Folkman and Lazarus 1985). Because the theory suggests that a necessary condition for the pursuit of coping strategies is the anticipation that these strategies will be effective or instrumental in making one feel better, its depiction of relevant considerations in this process is particularly valuable for our purposes.

Broadly, Lazarus and Folkman (1984; Folkman and Lazarus 1985) treat coping as a dynamic process wherein individuals consider and try to shape the person-environment transaction in a way that reduces threats to their well-being. Because each person-environment transaction (i.e., stressor) is different, individuals evaluate possible coping strategies in the context of the particular situation with which they are coping. Accordingly, transactional theory predicts a significant amount of within-person variation in beliefs that a given strategy will be helpful (i.e., strategy-specific coping instrumentalities) depending on the specific nature of the situation(s) (Folkman and Lazarus 1980; Folkman et al. 1986; Lazarus 1991; Lazarus 1999).

Control

Transactional theory argues that individuals’ evaluations of potential coping strategies occur during the process of secondary appraisal, wherein they consider whether (and how) they can manage stressors. The secondary appraisal process directs individuals towards primarily problem-focused or emotion-focused strategies. When the stressor is deemed uncontrollable, individuals tend to direct their coping efforts towards managing their emotions. Thus, they tend to anticipate emotion-focused strategies will be instrumental for coping (Folkman et al. 1986).

Given this, we expect employees to view CWB as more instrumental for coping when they perceive stressors as uncontrollable. This is consistent with theory and research findings that a person experiencing threats to control “becomes angry and hostile, and actively attempts to maintain or regain control” (Dupré et al. 2005, p. 385). Control, once lost, is often restored through symbolic means that help to provide emotional placation without necessarily addressing the stressor at hand (Allen and Greenberger 1980). For example, deciding to work slowly, to spread a rumor about someone, or to steal something all constitute acts of control (Bennett 1998). Moreover, at least some of these behaviors (e.g., withdrawal) may be seen as having the added coping advantage of allowing individuals to avoid uncontrollable and negative work situations (e.g., interactions with an abusive supervisor) and therefore be anticipated to limit further emotional strain (Whitman et al. 2014; Krischer et al. 2010).

Hypothesis 1

Perceptions of control over workplace situations are negatively associated with CWB coping instrumentalities at the within-person level.

Stressor Source

Folkman, Lazarus, and colleagues (Folkman and Lazarus 1985; Folkman et al. 1986) suggest that there may be other factors unique to the particular situation that influence the strategies that come to mind and are expected to be successful. We suggest that one such factor is the source of the stressor (e.g., boss, coworker, organization). In particular, we expect that the source of the stressor influences the specific type of CWB that employees anticipate will be effective. Acts of CWB are generally distinguished by whether they target a person within the organization (i.e., person-directed CWB; being rude to others, sabotaging others’ work) or the organization itself (i.e., organization-directed CWB; theft, decreasing contributions to the organization via production deviance or withdrawal; Bennett and Robinson 2000). If employees are approaching CWBs as a way to feel better by retaliating and/or avoiding mistreatment, it stands to reason that these behaviors will be linked to the source of the mistreatment (e.g., either retaliating against/avoiding the organization or the person(s) associated with it). This targeting hypothesis is a feature of retaliatory-based perspectives on CWB (e.g., Jones 2009; Skarlicki and Folger 1997; Spector and Fox 2002). It has been supported by research findings that organization-caused stressors (e.g., procedural injustice) engender more organization-directed CWB than person-directed CWB and that individual-caused stressors (e.g., informational injustice, incivility) engender more person-directed CWB than organization-directed CWB (Jones 2009). Furthermore, CWB targeting provides additional evidence of the instrumental use of CWB for coping, given that strictly hostile models would predict more diffuse acts of CWB.

Hypothesis 2a

Stressors stemming from the organization are associated with greater coping instrumentalities regarding organization-directed CWB than are stressors stemming from a specific person, at the within-person level.

Hypothesis 2b

Stressors stemming from a specific person are associated with greater coping instrumentalities regarding person-directed CWB than are stressors associated with the organization, at the within-person level.

In addition to direct effects, the source of the stressor (e.g., boss, coworker, organization) may signal the amount of control employees have in any given situation (Elfering et al. 2005). Although exceptions exist, it is generally easier to influence the actions of coworkers or a supervisor (i.e., specific individuals) than to influence the actions of organizational decision-makers (perceived by employees as “the organization”; Levinson 1965) who are further removed from the employee. Consequently, employees are likely to perceive stressors caused by organizational decisions or policies as less controllable than stressors associated with the actions of a single supervisor or coworker. In addition, control can be thought of as being negotiated between employees and organizations (see Alge et al. 2010). Much of the stressors emanating from organizations (e.g., bureaucratic policies, personnel decisions; Campbell 2012; Ouchi 1979) reflect types of organizational control and thus are likely to be perceived as particularly uncontrollable by the employee.

This expectation leads us to predict a partial mediation relationship involving stressor source and CWB instrumentalities at the within-person level. Specifically, we expect stressor source to impact perceptions of control, which in turn should shape CWB instrumentality judgments. As stated in Hypotheses 2a and b, we also expect direct relationships between stressor source and instrumentalities consistent with the targeting of CWBs.

Hypothesis 3

Perceptions of control partially mediate the relationship between stressor source and CWB coping instrumentalities, at the within-person level. Stressors stemming from the organization are associated with lower levels of perceived control (and thus higher coping instrumentalities) than stressors stemming from coworkers or one’s supervisor.

Between-Person Drivers of CWB Coping Instrumentalities

Although Folkman and Lazarus (1985) emphasized the role of situational appraisals in determining coping instrumentalities, they also pointed out that there might be stable individual differences in how individuals appraise both situations and coping capacity and, subsequently, evaluate the instrumentalities of various coping strategies.

Work Locus of Control

In particular, individuals differ in the extent to which they generally view events as controllable and believe that problem-solving efforts will successfully help them reach their goals, reflected in the construct of locus of control (Rotter 1966). In this study, we examine a related personality variable, work locus of control, which reflects the degree to which employees view workplace events as mutable and expect that effort will be met with reward. Work locus of control focuses on generally stable domain-specific beliefs (Rotter 1966) that individuals form regarding whether they have control over important outcomes at work or if outcomes are generally driven by luck or more powerful forces (i.e., whether luck or hard work is needed to get ahead at work; Perlow and Latham 1993). Consistent with arguments that employees with a more external work locus of control perceive fewer “productive” coping options and therefore see greater advantages in counterproductive ways to manage stressors, empirical evidence has linked locus of control and work locus of control to both self- and other- reports of CWB (Fox and Spector 1999; Perlow and Latham 1993; Sprung and Jex 2012). Since an external work locus of control is associated with reliance on emotion-focused strategies (Parkes 1984), including destructive behavior (Mitchell and Ambrose 2012), external work locus of control may also serve as a general shortcut (Edwards 1988) that leads employees to positively evaluate the coping instrumentality of CWBs in light of workplace stressors.

Hypothesis 4

External work locus of control is positively associated with CWB instrumentalities, at the between-person level.

Prior Engagement in CWB

Although Lazarus and colleagues focused on stable individual differences that may influence coping via the appraisal process, other coping researchers point out that there may be stable individual differences in coping preferences that occur outside of this process. For example, Edwards (1988) and Connor-Smith and Flachsbart (2007) have posited that there may be a degree of habit inherent in how people evaluate coping strategies. The basic idea is that people develop associations between situational triggers, emotions, and coping strategies. Given the depleting nature of stress, judgments regarding potential effectiveness of a given strategy are likely at least in part to be influenced by past use of that strategy. Research argues that enactment of past behavior influences an individual’s attitudes toward that behavior and whether they might be likely to engage in that behavior in the future (e.g., intentions and actual engagement in future behavior) in a manner partially independent of previous consequences of that behavior (Albarracín and Wyer 2000). Consequently, CWB might become a learned response to stressors, and individuals who have engaged in CWB in the past may be more likely to anticipate that these behaviors will be effective (Edwards 1988). It is worth noting that CWB strategies may work, at least in the short term. For example, Krischer et al. (2010) found that withdrawal and production deviance help reduce emotional exhaustion associated with injustice. Past use of CWB for coping might also become reinforced as individuals justify potentially negative or unethical behaviors to themselves (Bandura et al. 1996).

Hypothesis 5

Prior frequency of engaging in CWB is positively associated with CWB coping instrumentalities, at the between-person level.

Interactions of Within- and Between-Person Factors

Thus far, we have suggested that between-person variables (i.e., past CWB, external work locus of control) influence stable, or average levels of CWB instrumentalities associated with stressful situations. We have also argued that control, as a salient feature of situations, will be associated with CWB coping instrumentalities. Consistent with theory that individuals match their coping strategies to the particular situation in which they are faced (DeLongis and Holtzman 2005), it may be that these person-level variables take on added relevance when the stressor is deemed uncontrollable.Footnote 1 That is, individual tendencies towards viewing CWB as an instrumental response to stressors may be activated or enhanced when the situation also lends itself towards viewing CWB as such. Consequently, uncontrollable situations may be particularly likely to elicit positive CWB coping instrumentalities when individuals are pre-disposed towards viewing CWB in that manner due to past experience or an external work locus of control. This reasoning is consistent with associationist models of personality and aggression that suggest that individuals hold mental representations of beliefs, events, and behavioral scripts that become differentially activated depending on the situational context (Berkowitz 1990; Mischel and Shoda 2008). Thus, perceiving a situation as uncontrollable may activate schemas and scripts related to past behavior and beliefs, leading to a person by situation interaction between controllability (a situational factor) and external work locus of control and previous enactment of CWB (two person-level factors) when a negative or stressful event has occurred. Given this, we hypothesize interactions between perceptions of situation control (within-person) and external work locus of control and prior frequency of CWB (both between-person).

Hypothesis 6a

The negative effect of perceptions of control on CWB coping instrumentalities is greater to the extent that external work locus of control is high.

Hypothesis 6b

The negative effect of perceptions of control on CWB coping instrumentalities is greater to the extent that prior frequency of engaging in CWB is high.

Furthermore, given our expectation that situation control mediates the relationship between stressor source and CWB coping instrumentalities (see Hypothesis 3), we investigate the following:

Hypothesis 7a & b

The conditional indirect effect of stressor source of CWB instrumentality via perceptions of control is greater to the extent that (a) external work locus of control is high or (b) prior frequency of engaging in CWB is high.

Method

Participants

In order to sample across a wide variety of industries and positions, we recruited 400 participants via an online forum (Amazon’s Mechanical Turk). Numerous authors have encouraged the use of online recruitment in psychological research due to evidence of comparable validity and reliability of results (i.e., comparable data quality) and greater diversity of these samples as compared to more traditional sources (e.g., employees in a single organization; college students; Behrend et al. 2011; Buhrmeister et al. 2011). We retained 297 participants (74.25 %) after removing individuals who missed checks for careless responding (i.e., “this item is for key purposes, please select strongly agree”). Past research has shown that in large samples it is possible that up to 50 % of respondents may have to be removed for inadequate responding (Meade and Craig 2012).

The average age of participants was 32.5 years (SD = 10.63, range = 18–63). Over half of the participants were female (62.5 %). The largest number of participants stated college as their highest level of education (61.7 %), while 24.5 % reported a high school degree, 1.9 % reported no high school degree, and 11.5 % had earned a graduate degree. Participants reported an average tenure of 4.5 years (SD = 4.68, range = 0–30). Participants worked in a wide variety of industries including education, health care, customer service, and engineering. On average, participants reported that they work 37.4 h per week (SD = 10.72, range 15–80). Due to missing data, we had data for 294–297 participants and between 3527 and 3538 data points for the models predicting each of the three CWB outcomes. Our sample’s demographics are similar to those reported in other recent studies using online recruitment strategies (cf. Behrend et al. 2011).

Procedure

We employed a scenario-based approach to testing our hypotheses. As previously noted, this approach was advantageous for a number of reasons that were central to the purpose of this study. In particular, using scenarios allowed us to better understand variability in the formation of instrumentalities at both the between- and within-person (situational) levels. In addition, this approach allowed us to precisely select the types of stressors and CWBs of interest.

Scenarios were presented in random order. Similar to previous scenario research on unethical behavior at work, while the 12 scenarios were randomized between participants, each participant responded to the subsequent questions in the same order (cf. Maher and Bailey 1999). After answering questions about CWB coping instrumentalities for each scenario, participants were asked questions about controllability of each scenario, prior frequency of engagement in each CWB, and work locus of control. Although we theorized that these variables serve as antecedents of CWB instrumentalities, we assessed them afterwards to ensure that we did not risk the possibility of priming participants to our hypotheses. However, we note that there are trade-offs to this decision, and asking participants to consider the likelihood that each CWB would make them feel better before asking these questions may have also influenced the results.

Measures

Scenarios

Participants responded to 12 1-sentence scenarios involving a range of workplace stressors previously linked to engagement in CWB (Table 1). We attempted to maintain an appropriate balance between length and number of scenarios (cf. Weber 1992); thus the scenarios were kept as short as possible in order to avoid participant fatigue.

Table 1 Study scenarios and scenario source
Full size table

The scenarios were developed with the assistance of five graduate students in I/O psychology who were blind to the study hypotheses and served as subject matter experts. Specifically, the aim was to develop 1–2 sentence statements that would capture instances of stressors that had been linked to CWB in prior research. For example, organizational constraints have been linked to CWB (Hershcovis et al. 2007). Organizational constraints reflect conditions in the work environment that interfere with one’s ability to perform one’s job, including poor equipment, red tape, conflicting job demands, and inadequate information and training (Peters and O’Connor 1980; Spector and Jex 1998). Consequently, to portray a situation wherein the employee is faced with constraints, our scenario read “You have run into red-tape (e.g., organizational rules and procedures that make it difficult to accomplish things) for what seems to be the hundredth time this week.” As another example, one item from Cortina et al.’s (2001) incivility scale reads “paid little attention to your statement or showed little interest in your opinion” (see similar items on Einarsen et al. 2009 work-related bullying measure). This was translated into the scenario “You spent a lot of time making edits and suggestions on a business proposal. Your coworker sends the proposal without including any of them.” Working from measures, definitions, and examples of stressors used in prior research linking them to CWB, the SMEs narrowed the potential scenarios down from 22 to 12. These scenarios reflect the wide range of stressors associated with CWB, including incivility, interpersonal conflict, supervisor undermining, supervisory over control, distributive justice, procedural justice, job autonomy, and role ambiguity (e.g., Chen and Spector 1992; Colquitt et al. 2001; Dupré and Barling 2006; Fox et al. 2001; Hershcovis et al. 2007; Lau et al. 2003; Penney and Spector 2005; Rodell and Judge 2009; Yang and Diefendorff 2009).

As previously noted, CWB researchers have argued that organization-directed CWB is associated with organizational stressors whereas person-directed CWB is associated with person stressors (Bennett and Robinson 2000; Jones 2009). We coded stressor source to be consistent with this literature. However, while some stressors are associated with a clear source (e.g., incivility from coworkers, supervisory over control, organizational constraints), others (e.g., work-family conflict, job ambiguity) are not. Although we developed explicit hypotheses about the impact of person- and organization-caused stressors, we also examined stressors that are ambiguous in source because past research has linked these stressors to CWB as well.

Stressor source was dummy coded with three dummy variables—boss, coworker, ambiguous—with organization being coded as 0 in all three. The organization was coded as the reference group so that the coefficient estimates of the dummy variables reflect the difference between the organization scenarios and those stemming from coworker, supervisor, and ambiguous, respectively. This allows for testing Hypothesis 2a, 2b, and 3, which predicted differences between stressors stemming from a particular person and stressors stemming from the organization. After the study, we contacted participants to provide an additional verification for our coding scheme. Thirty participants provided information on whom (coworker, supervisor, organization, and other) they viewed as the cause of each scenario, who was to blame for each scenario, and whether they thought each scenario would be stressful. The majority (on average, 92 %) classified the coworker, supervisor, and organization scenarios in accordance with our coding scheme. Participants attributed blame and causality across a number of sources for the ambiguous scenarios. On average, 88.3 % agreed that the scenarios would be stressful.

CWB Coping Instrumentalities

We assessed CWB coping instrumentalities by asking participants “To what extent do you think that doing the following will make you feel better in this situation?” with regard to three types of CWB: (1) Decreasing your work effort (e.g., working slowly, wasting resources, taking long breaks); (2) Taking something belonging to the company; (3) Being rude to this person. These CWBs correspond to Spector et al.’s (2006b) categories of production deviance and withdrawal, theft, and abuse against others, respectively, and are referred to as such in our description of the results. For organization and ambiguous scenarios, we replaced “this person” with “a coworker.” We instructed participants to assume that they would have the opportunity to do all CWBs. Participants responded using a 5-point scale (1 = Definitely will not, 5 = Definitely will). Single-item measures are appropriate when the construct is simple and comprised of only one main element (rather than confusing or frustrating respondents with seemingly identical items; e.g., Robins et al. 2001).

Participants made instrumentality judgments for each of the three CWBs independently, consistent with theory on how individuals judge coping strategies. In particular, Edwards (1988) argued that individuals under stress tend to consider strategies in a sequential manner rather than evaluating multiple coping strategies simultaneously. In doing so, they evaluate each strategy based on the anticipated likelihood that it will make them feel better (e.g., what is the likelihood this will make me feel better?) rather than magnitude of improvement (e.g., how much better will I feel?) and stop searching for strategies when one is deemed acceptable.

Control

For each of the 12 scenarios, participants answered, “How much control do you think you would have in the following situations?” using a 5-point scale (1 = None, 5 = A lot).

Frequency of Prior Engagement in CWB

Participants indicated how often they had engaged in each of the three CWBs in their present job using a 7-point scale (1 = Never, 2 = Less than once a month, 3 = Once a month, 4 = 2–3 times a month, 5 = Once a week, 6 = 2–3 times a week, 7 = Daily).

External Work Locus of Control

To minimize study duration, we used the five highest loading items from Spector’s (1988) work locus of control scale (factor analytic results from Oliver et al. 2006) (α = .89). The items were “getting the job you want is mostly a matter of luck,” “making money is primarily a matter of good fortune,” “to get a really good job, you need to have family members or friends in high places,” “promotions are usually a matter of good fortune,” and “the main difference between people who make a lot of money and people who make a little money is luck.” Participants responded using a 5-point scale (1 = Strongly disagree, 5 = Strongly agree). Higher scores indicate more external work locus of control.

Results

Variance Components

We estimated null random intercept models using SAS Proc Mixed to assess the degree of within- and between-person variance in control and instrumentalities. These models provide estimates of τ2 and σ2, which reflect between-person (level-2) and within-person (level-1) variance, respectively. The percentage of variance associated with each level can be obtained by dividing the relevant estimate by the total variance (τ2 + σ2) (Bliese, 2000). Between 47.37 and 67.98 % of the variance in these variables was within-person (see Table 2), consistent with our argument that perceptions of CWB as a viable coping strategy differ both across situations and between individuals. Descriptive statistics and variable intercorrelations are displayed in Table 3.

Table 2 Variance components for instrumentalities and control
Full size table
Table 3 Descriptive statistics and variable intercorrelations
Full size table

Analytic Strategy

We used multilevel modeling in SAS to test our hypotheses (Singer 1998). Following recommendations by Snijders and Bosker (1999) we calculated average control perceived across all scenarios as a between-person variable. We calculated within-person variation in control by subtracting participants’ reports of control over each situation (i.e., person-mean centering). Including both in the regression models allowed us to separate within-person and between-person variance in control perceptions. All between-person variables were grand-mean centered.

Main Effects (Hypotheses 1–5)

We estimated separate models for each of the three CWBs—production deviance, theft, and abuse. The results of these analyses are displayed in Table 4.

Table 4 Multilevel regression results for production deviance coping instrumentality, theft coping instrumentality, abuse coping instrumentality, and control models
Full size table

Hypothesis 1 concerned the within-person influence of appraisals of control on CWB coping instrumentalities. As expected, participants anticipated that each of the three CWBs was likely to make them feel better in situations where they perceived lower levels of control [production deviance, B = −.12, SE = .02, p < .01; theft, B = −.07, SE = .01, p < .01; abuse, B = −.07, SE = .02, p < .01]. Hypotheses 2a and 2b predicted that stressor source would predict the specific forms of CWB individuals viewed as instrumental for coping. As hypothesized (2a), stressors stemming from the organization were associated with greater coping instrumentalities for production deviance and theft than were stressors stemming from a person [production deviance: boss dummy, B = −.39, SE = .06, p < .01, coworker dummy, B = −.68, SE = .06, p < .01; theft: boss dummy, B = −.21, SE = .04, p < .01, coworker dummy, B = −.38, SE = .04, p < .01]. Hypothesis 2b predicted that stressors stemming from a person are associated with greater coping instrumentalities regarding person-directed CWB than are stressors stemming from the organization. Consistent with this expectation, stressors stemming from either the boss or a coworker were associated with greater coping instrumentalities for abuse than stressors stemming from the organization [boss dummy, B = .80, SE = .05, p < .01, coworker dummy, B = 1.64, SE = .05, p < .01]. We also note that employees anticipated that theft was less likely and abuse was more likely to make them feel better when the stressor was ambiguous as opposed to organizational in source [theft model: ambiguous dummy, B = −.13, SE = .03, p < .01; abuse model: ambiguous dummy, B = .39, SE = .04, p < .01], but there was no difference for the production deviance coping instrumentality [ambiguous dummy, B = .02, SE = .05, ns]. In preliminary support of Hypothesis 3, we found that stressors stemming from the boss or a coworker were considered more controllable than those stemming from the organization (boss dummy: B = .58, SE = .05; coworker dummy: B = .88, SE = .05, ps < .01), as were ambiguous stressors (B = .67, SE = .04, p < .01). The formal test of Hypothesis 3 (i.e., whether control mediates the relationship between stressor source and CWB coping instrumentalities) is discussed below when we report models that also include the between-and within-person interactions (i.e., moderated-mediation models).

Hypothesis 4 stated that external work locus of control is positively related to CWB coping instrumentalities for all three CWBs. Supporting Hypothesis 4, those with a more external locus of control were more likely to endorse that each of the CWBs would make them feel better compared to those with a more internal locus of control [production deviance, B = .25, SE = .05, p < .01; theft, B = .13, SE = .04, p < .01; abuse, B = .17, SE = .04, p < .01]. Likewise, and in support of Hypothesis 5, we found that past engagement in production deviance [B = .21, SE = .03, p < .01], theft [B = .43, SE = .04, p < .01], and being rude to another at work [B = .38, SE = .03, p < .01] all significantly and positively predicted their respective coping instrumentalities.

Interactions (Hypotheses 6a & 6b, 7a & 7b)

Hypotheses 6a and 6b predicted interactions between control (a within-person variable) and external work locus of control and frequency of prior CWB (between-person variables). We ran additional models (displayed in Table 5) wherein we estimated these interactions as predictors of the CWB instrumentalities.Footnote 2 Before doing so, we re-ran the models predicting the three instrumentalities in order to assess whether there were significant random slopes for control, which would indicate the possible presence of a cross-level moderator. There were significant random slopes for control in all of the models.

Table 5 Multilevel regression results for cross level interactions between control and person variables in predicting production deviance, theft, and abuse coping instrumentalities
Full size table

As seen in Table 5, there were significant interactions between external work locus of control and situational (i.e., within-person) perceptions of control in predicting production deviance and theft coping instrumentalities, but not abuse coping instrumentalities. Simple slopes analyses revealed that the relationship between control and the instrumentalities was stronger for employees with a high (+1 SD) external work locus of control (production deviance B = −.16, SE = .03; theft B = −.11, SE = .02, ps < .01) than for employees with a low (−1 SD) external work locus of control (production deviance B = −.07, SE = .03, p = .02; theft B = −.04, SE = .02, p = .05). In other words, the positive impact of decreasing control on production deviance and theft instrumentalities was strengthened when external work locus of control was high. Thus, Hypothesis 6a was partially supported. The interaction between frequency and situational control was also significant in the models predicting production deviance and theft coping instrumentalities, but it did not significantly predict abuse instrumentalities. Simple slopes analyses revealed that the relationship between control and production deviance and theft instrumentalities was stronger for employees reporting a high frequency (production deviance B = −.16, SE = .03, p < .01; theft B = −.13, SE = .02, p < .01) than for employees reporting a low frequency of past CWB enactment (production deviance B = −.07, SE = .03, p = .03; theft B = −.02, SE = .02, ns). Thus, Hypothesis 6b was also partially supported. As these four interactions were graphically similar, we present only the interaction between control and external work locus of control predicting theft instrumentalities in Fig. 2 for illustrative purposes.

Fig. 2
figure 2

Interaction between situational control and external work locus of control predicting theft coping instrumentality. The label ‘low’ reflects −1 SD; the label ‘high’ reflects +1 SD. EWLOC refers to external work locus of control

Full size image

Finally, given that we anticipated that situational/within-person perceptions of control would mediate the relationship between stressor source and the CWB instrumentalities (Hypotheses 3), and that the link between situational/within-person perceptions of control would be moderated by the between-person variables in our study (Hypotheses 7a & 7b), we ran Sobel tests to examine the conditional indirect effects of stressor source on the CWB coping instrumentalities (MacKinnon and Dwyer 1993).Footnote 3 There were significant indirect effects of stressor source on CWB coping instrumentality for all levels of the moderators (noted above) where the simple slope between situational/within-person perceptions of control and CWB coping instrumentality was significant (supporting Hypothesis 7). In the case that the moderation results described above did not demonstrate a significant second-stage moderation (Edwards and Lambert 2007), we simply estimated the indirect effect of stressor source on CWB coping instrumentality. In all of these cases, situational/within-person perceptions of control mediated the relationship between stressor source and CWB coping instrumentality. Note that this mediation is partial given the evidence of direct effects of stressor source on CWB coping instrumentalities noted above (supporting Hypothesis 3). The full results of these analyses are available from the first author.

Discussion

The current study examined factors that lead employees to anticipate that CWBs will help them feel better when experiencing workplace stressors. Coping theory suggests that individuals pursue certain coping strategies because they expect those strategies to improve their well-being, and that these expectations vary between individuals as well as within individuals based on the nature of the situation (Edwards 1988; Folkman et al. 1986; Lazarus and Folkman 1984). Consistent with these ideas, we found that coping instrumentalities vary significantly both within- and between-persons for all three types of CWB we investigated. These findings suggest that some individuals are generally more likely to view CWB as an effective coping strategy than others, but they also indicate that individuals take situational characteristics into account when evaluating coping instrumentalities for CWBs.

In particular, our results point to perceptions of situational control as an important predictor of coping instrumentalities. For each of the three CWBs, the less control individuals perceived over a situation, the more they thought engaging in these behaviors would make them feel better. It is important to note that the impact of control was at the situational and not at the person level; between-person variance in perceptions of control was not related to instrumentalities. That is, those individuals who tended to view the scenarios as less controllable on average did not believe CWBs were more likely to make them feel better than those who tended to view the scenarios as more controllable. This finding is consistent with transactional theory’s argument that individuals appraise each situation with regard to control and, in turn, perceptions of control influence how coping strategies are evaluated. It is also consistent with existing theory that has posited a key role for situational control in reducing potentially dysfunctional coping responses to workplace stressors (Fox and Spector, 2006). Although individual differences in perceptions of control across the scenarios did not predict instrumentalities, individual differences in work locus of control predicted instrumentalities in all three models. Those with an external work locus of control were more likely to anticipate that CWB will be an effective coping strategy than those with an internal locus of control. External work locus of control also appeared to exacerbate the effects of low situational control on production deviance and theft instrumentalities.

Consistent with our expectations, stressors associated with one’s boss or coworkers were viewed as more controllable than those associated with the organization more broadly. Also as expected, stressors stemming from one’s boss or coworkers were more strongly associated with person-focused CWB instrumentalities than were stressors stemming from the organization (and vice versa for organization-focused CWB instrumentalities). Interestingly, ambiguous stressors were not different than organizational stressors for instrumentalities for production deviance. However, instrumentalities for theft were lower for ambiguous stressors than for organizational stressors and instrumentalities for abuse were higher for ambiguous stressors than for organizational stressors. Future research is needed focusing on how individuals appraise and respond to these types of stressors (cf. Martinko et al. 2002), especially given that many of the stressors that individuals face in the workplace likely arise from a confluence of different factors.

Our results also support arguments that frequency of past engagement in CWB influences how individuals evaluate these behaviors as coping strategies. This suggests the existence of a learned process whereby having engaged in CWB as a coping strategy in the past makes one more likely to anticipate that it will be successful in the future (Albarracín and Wyer 2000; Edwards 1988). The interaction results suggest that past experience engaging in production deviance and theft is particularly relevant for instrumentalities when individuals perceive low situational control. Perceptions of low control may, therefore, activate schemas developed regarding the coping instrumentality of CWB from past experience.

By finding that a number of constructs that have been theoretically or empirically linked to the enactment of CWB (i.e., control, work locus of control, stressor source) also predict CWB coping instrumentalities, our study provides further support for the idea that attempts to cope underlie the enactment of CWB. This is important because coping instrumentalities for CWB have largely been assumed rather than directly investigated, and therefore it has not been entirely clear why (or when) the experience of stressors should necessarily result in CWB. If CWB, at least in part, is motivated by employees’ attempts to cope, we would expect variation in employees’ expectations that CWB will make them feel better. This variation, then, should be predicted by constructs explicated in coping theory. Our findings provide more compelling evidence that CWB reflects employees’ efforts to cope than do correlations between stressors and CWB (cf. Martocchio and Judge 1994). Our findings also extend the limited literature on between-person variance in instrumentalities by considering both within-person and between-person variance in these judgments (Bushman et al. 2001; Jones 2009).

Practical Implications

Our findings point to a process wherein individuals form appraisals regarding the value of CWB for coping on an ongoing basis. Thus, in addition to the oft-noted recommendation for organizations to address stressors to decrease triggers of CWB, our results suggest that organizations looking to reduce CWB might do so by intervening to influence the extent to which employees perceive CWB as instrumental for meeting coping aims. For example, organizations may implement interventions that seek to increase employees’ perceived control over workplace stressors, thus helping employees to adopt more desirable, productive coping strategies. One potential strategy might be to increase employee participation in organizational decision-making. Organizations might also encourage climates where employees feel comfortable bringing problems to the supervisor or organization’s attention. This may help employees to feel they have greater control over stressors and, consequently, reduce the extent to which they choose CWBs as coping responses. Organizations faced with instances of CWB may also use the knowledge that (a) coping instrumentalities for CWB increase as situational control decreases and (b) that CWB coping instrumentalities are differentially impacted given the source of the stressor in order to ascertain the particular stressors that are giving rise to CWB.

While our findings suggest that individuals evaluate CWB coping instrumentalities depending on the situation, they also point to the existence of stability in these beliefs based on external work locus of control and frequency of prior CWB. The selection process may manage external work locus of control. Additionally, organizations may also be able to screen for past engagement in deviant behaviors in the selection process via integrity testing (Ones et al. 1993).

The fact that we found (a) approximately half of the variance in CWB coping instrumentalities occurred at the within-person level and half occurred at the between-person level, and (b) significant within- between-person interactions, suggests that interventions need to be directed at both levels to be maximally effective. Indeed, interventions directed at both levels may be mutually reinforcing. For example, interventions to increase situational control may reduce enactment of CWB via reducing CWB coping instrumentalities. At the same time, the reduction of CWB may serve to further reduce CWB coping instrumentalities. Because prior engagement in CWB is associated with CWB instrumentalities, efforts to reduce CWB in the present may have carry-over effects in reducing the likelihood of CWB in the future.

Limitations and Additional Directions for Future Research

A particular strength of this study is the use of scenarios to distinguish between within- and between-person variance in instrumentalities regarding the use of CWB as coping strategies. As previously noted, this nuanced view provides greater insight into the formation of important beliefs that may motivate CWB in stressful work situations. However, given the nature of the research question and design, we did not examine explicit behaviors. While formal expectancy models suggest that the motivational force for a specific course of action (an intention in its own right; see Pinder 2008) can be calculated as a multiplicative function of instrumentality beliefs, valence beliefs (expected satisfaction of an outcome, in this case reduced strain), and expectancy beliefs (beliefs regarding whether effort leads to successfully engaging in a specific behavior, in this case whether individuals could engage in CWB if they tried), research has provided evidence that instrumentality perceptions in isolation serve as reliable predictors of intentions and behavior and that the multiplicative models do not necessarily lead to higher effect sizes (Van Eerde and Thierry 1996). Thus, we measured only CWB instrumentalities. However, we instructed participants to assume that they would have the opportunity to engage in each of these behaviors in order to avoid the possibility that variation in expectancies would add noise and possibly suppress relationships between instrumentalities and intentions. Further, we assumed that the valence beliefs would be similar across participants and situations (i.e., feeling better). However, future research might explicitly examine these other motivational factors as well. Future research that examines the interplay between the factors mentioned here and the generation, evaluation, selection, and enactment of CWBs as coping tactics would be valuable.

As previously noted, we asked about interpersonal CWBs directed towards a coworker when the stressor was associated with the organization or was ambiguous, in order to maintain the same level of specificity in the CWBs across the scenarios. An implication of this decision is that this study does not examine the extent to which employees perceive engaging in CWB directed towards a supervisor will make them feel better when the stressor stems from the organization, a coworker, or is ambiguous (i.e., displaced aggression). It is possible that interpersonal CWB in the context of a stressor associated with the organization might involve actions directed against the supervisor, given the supervisor is a representative of the organization. However, given that judgments of supervisors’ representativeness vary depending on the supervisor (Shoss et al. 2013), we did not consider this to be a likely possibility in this study. Additionally, meta-analytic evidence by Hershcovis et al. (2007) suggests that stressors associated with the organization relate similarly to supervisor- and coworker-targeted aggression. Thus, we viewed the value of being specific in the CWB options to outweigh limitations of this approach. Nonetheless, future research might examine this possibility when studying coping motivations for CWB in field studies.

In order to more closely examine the decision-making processes underlying the pursuit of CWB for coping purposes, we created scenarios rather than asking employees about their specific work situations. While there are certain strengths associated with using a scenario methodology, our findings are limited by the extent to which employees engaged the same decision-making processes they would when actually experiencing stressful situations in the workplace. As recommended by Weber (1992), it is important to be aware of the amount of effort required of participants as they may become fatigued if there are too many scenarios to read or if the scenarios are too long. Consequently, we elected to keep the scenarios as short as possible given that we had twelve of them. However, future research might replicate the current study’s findings using more detailed scenarios. Additionally, we point out that we do not expect CWB to be generated as a response to every stressful workplace situation. The scenarios in our study involved situations that individuals could attribute to their organization, supervisor, coworkers, or some combination thereof. We did not include any scenarios where individuals could attribute causes of stressors to themselves. Martinko et al. (2002) suggested that internal and stable attributions (e.g., lack of ability) would lead to internal CWB (e.g., alcohol use) by triggering emotions such as guilt instead of anger. It would be interesting to see if and when employees believe that these internal CWBs would likewise make them feel better.

Moreover, we acknowledge that the data were collected at a single point of time. In any study that relies primarily on a given methodology, in our case self-report surveys, common method variance (CMV) is a potential issue. From a broad standpoint, Spector (2006) argues that if CMV is an omnibus biasing factor in a given study, one would expect to see all measures using the same method to be correlated. Even given the substantial statistical power in our study due to the large number of participants (and corresponding observations at the within-person level), there are a number of non-significant correlations amongst self-reported variables (see Table 2) which provides preliminary evidence that CMV may not be a primary concern. In addition, it is important to note that our choice of response format (which may be considered a method factor in its own right) varied substantially across the different self-report measures. Given this, along with the numerous observed negative correlations, we believe that a case can be made against general biasing effects of unidirectional response-sets or acquiescence. However, future research is needed to examine these hypotheses in field settings and in non cross-sectional designs.

Additionally, although research has shown that the quality of data collected through mTurk meets, and may even exceed, the psychometric standards found in published research (Behrend et al. 2011; Buhrmeister et al. 2011), it is important to note the limitations that are associated with using online recruitment tools such as mTurk. First, mTurk attracts a limited demographic of participants (e.g., only internet savvy individuals). Further, research has shown that mTurk participants are less extraverted and emotionally stable than participants recruited using traditional means; however, they still produce reliable results that are similar to previous research findings (Goodman et al. 2013). Research is needed to validate the use of mTurk for psychological research and examine the value of screening proceeds (e.g., attention checks). Finally, although the quality of data is similar, it is possible that mTurk participants may vary from traditional participants; thus, researchers should use caution when generalizing findings to a broader population of individuals (Behrend et al. 2011).

In conclusion, the current study provides evidence that employees view CWB as a viable mechanism for coping with workplace stressors, in particular to the extent that they have engaged in CWB in the past and in situations that they perceive as uncontrollable. We view this coping perspective as particularly fruitful for understanding and ultimately managing CWBs, and we encourage future research in this area.