1. Introduction
Counterproductive work behavior (CWB) refers to employees’ discretionary behavior that violates major norms or harms organizational well-being [
1]. These deviant acts can vary from stealing from coworkers or the employer, withholding effort while at work, showing up late (or not at all), spreading rumors, to outright interpersonal aggression toward others [
1,
2]. Not only has research found that CWBs result in billions of dollars lost each year [
3,
4,
5], but CWBs also occur at meaningful rates [
6]. In brief, CWBs are of major concern to organizations, and this concern has resulted in a great deal of research attention focused on understanding, predicting, and ultimately preventing CWB.
One major stream of research on CWB focuses on important situational factors that influence the likelihood of CWB occurring. These contextual variables have included various features of the organization, such as leadership [
7], climate [
8], and ethical codes of conduct [
9]. Other research focuses extensively on individual differences that may predispose an employee to engage in CWB. Broad personality factors such as agreeableness and conscientiousness [
10] as well as dark personality traits such as Machiavellianism, psychopathy, and narcissism [
11,
12] are linked to CWB; however, substantial levels of variance have yet to be explained through either situational factors or individual differences alone [
13]. Given the complex nature of unethical behaviors, such as CWB, this finding suggests that simple main effect relationships (e.g., trait–CWB or situation–CWB) are inadequate to capture the totality of CWB.
One reason for the lack of precision in predicting CWB may stem from the reliance of cross-sectional studies that do not consider potential changes in an individual’s likelihood to commit CWB [
13]. Several popular examples depict a “slippery slope” of unethical behavior gradually increasing over time, highlighting the limitation of examining CWB cross-sectionally. For example, Bernie Madoff’s Ponzi scheme began as small levels of fraud that eventually grew into an elaborate scheme with a size of USD 64.8 billion [
14]. Madoff himself was quoted saying “It starts out with you taking a little bit, maybe a few hundred, a few thousand. You get comfortable with that, and before you know it, it snowballs into something big” [
15]. The Enron scandal is another case of major unethical decision making resulting from a “steady accumulation of habits and values and actions that began years before and finally spiraled out of control” [
16] (p. 132).
Although these examples show a gradual eroding of ethicality over years, research has demonstrated that the slippery slope effect can unfold over much shorter periods of time. For example, the infamous series of obedience studies by Milgram also demonstrates the extent to which small behaviors, such as delivering a mild shock, can escalate into delivering seemingly lethal levels of electricity to an unresponsive subject within a relatively short period of time [
17]. Indeed, research has demonstrated this effect across only three to ten math problems [
18].
2. The Role of Cognitive Dissonance and Dissonance Reduction
Despite the amount of mainstream attention focused on gradual declines in ethical behavior, relatively little research has examined explanatory factors that contribute to such changes. One major component in understanding the gradual progression of unethical behavior, such as CWB, may be cognitive dissonance and how cognitive dissonance is reduced. For example, cognitive dissonance, which is the discomfort a person experiences when he or she makes choices or has thoughts that are inconsistent with each other [
19], was evident during the Milgram studies. Videos of participants reveal several instances of this psychological tension when asked to deliver increasingly severe shocks to the “learner”. These participants likely knew they were delivering shocks that might seriously harm the learner; however, this thought conflicted with being told by an authority figure that “the study must continue”. Many participants appeared to be able to reduce this dissonance by diffusing their responsibility to be an ethical person onto the researcher.
Reducing dissonance, such as the through diffusion of responsibility, is rooted in Bandura’s social cognitive theory [
20]. According to social cognitive theory, unethical behaviors are deterred through anticipated self-condemnation from engaging in some possible deviant act. However, this self-condemnation only occurs when self-regulatory capacities are successfully in place. Unfortunately, there are cognitive processes that can cause this self-regulation to fail, resulting in an increased likelihood of unethical behavior. For example, research indicates that individuals tend to make rationalizations for small levels of unethical behaviors to avoid tainting an otherwise positive, moral self-concept [
21,
22,
23]. By mentally distancing themselves from the moral implications of a deviant act, individuals can avoid breaching their own personal ethics that would otherwise deter them from unethical acts.
The social cognitive process of distancing oneself from morals and personal ethics, known as moral disengagement, occurs through a cognitive distortion that involves viewing one’s wrongful acts as being acceptable while preventing a breach of ethical standards [
24,
25,
26]. Not only does moral disengagement increase the probability of exhibiting wrongful or aggressive behavior [
27,
28], it also can induce the motivated forgetting of personal ethics during a simple contemplation of unethical behaviors [
29]. Research has examined how this process can unfold, leading to increasingly severe levels of unethical behaviors over time [
30] and the propensity to morally disengage acts as an explanatory mechanism for gradually eroding levels of ethicality [
18]. Bandura described this “slippery slope” effect as a form of “ethical numbing” that occurs through repeated exposure to rationalizations [
20].
3. Study 1: The Role of Dark Triad Traits in Cognitive Dissonance Reduction
Researchers note the need to understand the role of individual differences, such as the dark triad traits of Machiavellianism, psychopathy, and narcissism, in the susceptibility of falling into a slippery slope of unethical behavior [
18]. Individual differences are likely to play a major role whether unethical behavior becomes increasingly likely over time because of the influence personality has on the initial occurrence of unethical behavior such as CWB [
31]. The dark triad, which refers to the personality traits of narcissism, Machiavellianism, and psychopathy, is especially relevant to understanding traits that predispose an individual to engage in an unethical behavior [
32].
Perhaps the most heavily researched of the dark triad traits is narcissism [
33]. The underlying characteristics of narcissism involve a sense of entitlement, superiority, vanity, arrogance, and self-sufficiency [
34]. Machiavellianism is a trait based on the literature written by Niccolo Machiavelli, who was an Italian philosopher and political writer. In some of his writings, he endorsed the use of immoral behavior and ruthlessness as being effective and acceptable in politics. A lack of morality, manipulativeness, and general suspicion of others’ intentions are primary characteristics of Machiavellianism [
35]. Finally, psychopathy is characterized by lack of remorse, callousness, impulsivity, and egocentricity [
36].
Research has also suggested that individuals with high levels of dark triad traits may be predisposed to morally disengage. For example, Moore and colleagues’ [
37] series of studies on moral disengagement in work settings found Machiavellianism to be associated with a greater propensity to morally disengage. Similarly, Egan et al. [
38] found that moral disengagement mediated the relationship between the dark triad personality traits (i.e., narcissism, Machiavellianism, and psychopathy) and unethical consumer attitudes.
While there is limited research examining the role of moral disengagement in explaining how dark triad traits are translated into unethical attitudes and behaviors, there is no research (to the authors’ knowledge) that examines how dark triad traits and their indirect effect through moral disengagement influence changes in the likelihood of engaging in unethical behaviors across subsequent opportunities. The present study attempts to fill this gap in the research literature by examining the influence of individual differences, specifically, the dark personality traits, on the emergence of a “slippery slope” effect where unethical behavior increases past some initial level. While past research has examined the propensity to morally disengage, the present studies provide an opportunity to actually engage in cognitive dissonance reduction (i.e., moral disengagement) by presenting a manipulation to half of the sample that asks participants to explain why they did or did not engage in unethical behaviors presented to them prior to the manipulation.
Inducing participants to explicitly justify their actions should make participants’ reasoning more salient to themselves. Individuals high in dark personality traits (dark triad traits and the propensity to morally disengage) are more likely to engage in unethical behaviors and subsequently reduce their dissonance through moral disengagement when asked to explain their decisions. The self-serving bias and lack of empathy that characterizes individuals high in dark personality traits enables them to justify unethical actions more easily. Conversely, individuals that are low in dark personality traits are guided by strong personal ethics, which deter them from initial unethical decisions and make them even less likely to engage in further unethical actions after reflecting on their moral reasoning in the manipulation condition. Specifically, we predict the following:
H1. After controlling for initial unethical decision making, the opportunity to morally disengage will moderate the effect of dark personality traits on subsequent unethical decision making, such that, in the manipulation condition, those high on dark personality traits will be more likely to engage in unethical decisions and those low on dark personality traits will be less likely to engage in unethical decisions.
3.1. Method
Participants
Working adults within the United States (
n = 411) were recruited via Amazon’s Mechanical Turk and compensated USD 0.50 for completing the experiment. Participants worked in a variety of occupations and were pre-screened based on working at least part-time (i.e., 20 h a week). In addition, participants were required to be 18 years of age or older, living within the United States, and to report proficiency in the English language. A total of 52 participants did not complete the survey and were removed from the sample. An additional 42 participants failed a careless responding attention check (e.g., “please select ‘agree’ for this item”) [
39] and were also removed from the sample. The final sample of 317 participants was 53.9% female and were, on average, 36 years old (ranging from 21 to 69). The racial representation of the sample was 70.7% White/Caucasian, 8.5% Black/African-American, 7.9% Hispanic/Latino, 9.8% Asian-Pacific Islander, 0.6% Native-American/American-Indian, and 2.5% other.
3.2. Procedure and Manipulation
After providing informed consent and completing the demographic section of the survey, participants were given measures assessing levels of the dark triad traits along with the propensity to morally disengage. Next, participants were given a random subset of four items from the eight-item unethical decision-making scale that asked participants to describe the likelihood they would engage in unethical behaviors across various scenarios. Following the unethical decision-making items, half of the participants were randomly assigned to provide a text response to the following prompt: “Take a little time to think back to the situations and work behaviors you just read about. In the space below, explain why you would choose to do (or not do) the behaviors described in the situation”. Participants were required to type at least a 25-character response.
The goal of this manipulation was to make the participants’ reasoning for engaging (or not engaging) in behaviors described in the prior set of items more salient to themselves. For individuals that indicated they were likely to engage in the behaviors, this manipulation may have provided an opportunity to “practice” moral disengagement by justifying their decisions. On the other hand, those who indicated they would not be likely to engage in the behaviors were provided an opportunity to “practice” or “morally engage” by describing any reasons they did not make unethical decisions.
Following the manipulation, all participants were given a distractor task involving simple cognitive ability test items. This distractor task took approximately three to four minutes for the participants to complete. After the distractor task, each participant was given the four unethical decision-making items that they did not answer during the first administration of the items. Following these final four items, all participants were thanked for their participation and debriefed.
3.3. Measures
The Short Dark Triad (SD3) [
40]: The 27-item Short Dark Triad scale was used to assess Machiavellianism (e.g., “It’s not wise to tell your secrets”;
α = 0.83), narcissism (e.g., “Many group activities tend to be dull without me”;
α = 0.79), subclinical psychopathy (e.g., “Payback needs to be quick and nasty”;
α = 0.84). The response format was a 7-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = somewhat disagree, 4 = neither agree nor disagree, 5 = somewhat agree, 6 = agree, 7 = strongly agree).
Unethical Decision Making [
41]: Unethical decision making was assessed using an eight-item measure that involved different scenarios that each involved some sort of unethical behavior (e.g., “You work in a fast-food restaurant downtown, it’s against policy to eat food without paying for it. You came straight from class and are therefore hungry. Your supervisor isn’t around so you make something for yourself and eat it without paying”;
α = 0.84). Participants were instructed to indicate if they would engage in the behavior described (1 = I would definitely NOT do this, 2 = I would probably not do this, 3 = unsure, 4 = I would probably do this, 5 = I would definitely do this). Four items were presented to each participant across two time points, as described in the study procedure. To reduce item-level effects across the two administrations of the scale, each participant received items in a random order.
Moral disengagement [
37]: Eight items from Moore et al. [
37] were used to assess the propensity to morally disengage (e.g., “Taking something without the owner’s permission is okay as long as you’re just borrowing it”;
α = 0.91). The response format was a 7-point Likert scale (anchored from 1 = strongly disagree to 7 = strongly agree).
Distractor Task: Raven’s Progressive Matrices [
42] were used to introduce time between measures of unethical decision making. Each item comprised a three-by-three matrix with all but one block containing a black and white pattern that changes based on relationships and rules from other columns and rows. Participants chose one of the eight options to fill in the missing block. This section was not scored.
3.4. Results
Descriptive statistics and bivariate relationships between variables of interest are detailed in
Table A1. Before testing our hypothesis for study 1, a series of t-tests were conducted to examine differences between mean levels of unethical decision-making scores and conditions during the first and second administrations of the unethical decision-making scale. There were no significant differences found between conditions for either administration (first administration:
t = −0.61,
df = 314,
p = 0.54; second administration:
t = 0.93,
df = 303,
p = 0.35). In addition, paired samples t-tests were also conducted to examine within the condition equality of means across the two administrations of the unethical decision-making scale. There were no significant differences found through these t-tests either (manipulation condition:
t = 1.72,
df = 153,
p = 0.09; control condition:
t = −0.92,
df = 162,
p = 0.36).
To test our prediction that the opportunity to morally disengage would moderate the relationship between dark personality traits and subsequent unethical decision making, a series of hierarchical regressions were conducted with the mean level of the second administration of the unethical decision-making scale as the outcome. Three models were examined for each dark personality trait. Since we were interested in changes in unethical decision making beyond participants’ initial levels, the first model controlled for the mean score for the first administration of the unethical decision-making scale. The second model added the main effects for the dark personality trait and condition. The third model added the interaction term for dark personality interacting with the condition. The results from these regressions are detailed in
Table A2.
As shown in
Table A2, full support for our hypothesis was found for Machiavellianism and psychopathy, which both interacted significantly with the condition in predicting subsequent unethical decisions. This interaction was such that, in the manipulation condition, where participants were asked to explain their choices, Machiavellianism and psychopathy both had significant effects on subsequent unethical decision making beyond their initial levels. Participants’ Machiavellianism and psychopathy scores had no relationship on subsequent unethical decision making beyond initial levels in the control group. These interaction terms also explained significant incremental variance in the scores for the second administration of unethical decision making beyond the first two models. Partial support was found for the propensity to morally disengage, with a marginally significant interaction with the condition (B
interaction = 0.13,
p < 0.10). Effects were in the expected direction and followed the same pattern as Machiavellianism and psychopathy. However, scores for the moral disengagement scale were positively skewed, and, after conducting a log10 transformation on moral disengagement scale scores, the interaction with the condition was significant and also explained significant incremental variance beyond the first two models (B
interaction = 0.34,
p < 0.05; Δ
R2 = 0.01,
F1, 312 = 4.04,
p < 0.05). No support was found for our hypothesis for narcissism. Despite effects in the expected direction for narcissism and a simple slopes analysis revealing a significant effect of narcissism on subsequent unethical decision making only in the manipulation condition, the interaction term was nonsignificant and did not explain any incremental variance beyond the first two models. Simple slopes plots for each interaction are shown in
Figure A1,
Figure A2,
Figure A3 and
Figure A4.
3.5. Study 1 Discussion
Study 1 examined whether the likelihood of subsequent unethical decisions would change as a function of moral disengagement and individual differences. Results demonstrate the likelihood of subsequent unethical decisions not only changed beyond initial reported levels, but that this change only occurred when participants’ reasoning for initial decisions was made salient to themselves. Although this study had no manipulation check, examining the text responses provided in the manipulation condition revealed rationalizations consistent with moral disengagement. For example, one participant who had an elevated Machiavellianism score reported the following rationale for choosing to engage in the unethical behaviors described:
“Let the people pay for their mistakes. When I make mistakes, I suffer. I would expect the others to do so, so that they may be careful next time.”
Consistent with past research that demonstrates moral disengagement can lead to high levels of unethical decisions [
18], individuals high on all but narcissism tended to report a higher likelihood of making unethical decisions beyond the level they reported in the first administration of unethical decision-making items. Interestingly, the opposite effect appeared to occur for individuals with low levels of dark personality traits. For many of these individuals, their perceived future self-condemnation [
20] was able to be seen in text responses:
“Most of these are morally wrong to do, and I would feel guilty about having chosen wrong and not gone with my brain and heart on what I felt was right.”
Expected effects were only found in the manipulation condition where reasoning was made explicit by the participants. However, in “real world” settings where the potential for making unethical decisions exist, such as with CWBs in organizations, individuals would likely be able to eventually make sense of their actions, even if this sense-making and justification is not explicitly described as it was in study 1.
One factor that research has already found to play a role in the sense making and attribution process for CWB is perceptions of injustice [
43,
44]. Perceptions of injustice can be a major stressor for employees at work and play a significant role in several major models of predicting CWB, such as the widely supported stressor-emotion model [
45]. However, there has been a lack of research investigating how individual differences, such as dark personality traits, influence the way people attribute injustice to their actions at work [
44]. Thus, study 2 was conducted to further examine the influence of dark personality traits and moral disengagement on the role of justice perceptions.
4. Study 2: The Influence of Dark Personality and Moral Disengagement on the Role of Justice Perceptions
Organizational justice, or the perception of fairness in the workplace, is an important attitudinal variable in management research, particularly given its relation to performance, compliance with management decisions, organizational citizenship behaviors, and counterproductive work behaviors [
46]. Nevertheless, observed justice–CWB relationships are consistent but modest (
r = 0.22–0.28) [
46]. The idea that perceptions of equity and fairness drive human responses to events (e.g., perceived injustice) is as old as the golden rule, and this idea has guided research to presume that, when CWB and justice are found to be related, CWB is a consequence of perceived injustice.
Bandura and colleagues [
25] posited that individuals may use cognitive mechanisms that allow themselves to behave in deviant ways by separating from their self-determined morals, thereby protecting positive (i.e., moral) evaluations of the self. Further, Tenbrunsel and Messick [
47] introduced their concept of “ethical fading”, by explaining that individuals have the capacity to engage in self-deception in order to behave in a manner that benefits themselves, while still holding the (false) belief that they are upholding their moral principles. Self-deception allows individuals to use “biases in perceived causation that erroneously reduce moral responsibility” in order to make decisions that are optimal for the individual [
47] (p. 233). Following this logic, one could use distorted perceptions of justice to deceive themselves and engage in unethical behavior, all while believing their behavior is justified and in line with their morals. If justice perceptions are distorted and manipulated to be used to facilitate unethical behavior, there is ambiguity in whether it is organizational injustice or the personality of individuals who are more susceptible to ethical fading that is predictive of CWB. Following this logic, distorted standards of justice via self-deception may be used to justify unethical behaviors such that the behaviors are perceived to be in-line with personal moral standards. Due to the intertwining of distorted justice perceptions and dark personality traits, it may be difficult to attribute CWB to actual organizational injustice vs. an individual susceptibility to “ethical fading”.
Moral self-regulation occurs through both proactive and inhibitive mechanisms [
20], and justice perceptions are likely involved in broader self-regulation at work, but perceived injustice may serve as a justification for moral disengagement. Following the concept of a “slippery slope” [
18], providing an opportunity to justify unethical decisions (study 1) or counterproductive work behavior (study 2) may allow those lower on dark triad traits to morally engage and those higher on dark triad traits to morally disengage.
Participants without the opportunity to explicitly engage or disengage should draw more from justice perceptions while participants experiencing the intervention will form justifications for their (im)moral behavior that make justice perceptions less influential. For participants without the opportunity to explicitly engage or disengage, justice perceptions are more influential and salient in the decision-making process to engage in unethical behavior. Conversely, participants experiencing the intervention and given an opportunity to morally “engage” or disengage are likely to form justifications for their (im)moral behavior, thereby diminishing the influence of justice perceptions.
While Welsh et al. [
18] have suggested that individuals high on dark triad traits may be more likely to morally disengage over time, we are not aware of any empirical tests of that proposition. Thus, study 2 investigates whether a dissonance reduction strategy that allows participants to rationalize ethical or unethical behavior reduces the mediating role of justice perceptions in dark personality–CWB relationships. Specifically, we predict the following:
H2. The opportunity to rationalize moderates the indirect effect of dark personality traits (Machiavellianism, narcissism, psychopathy, and the propensity to morally disengage) on CWB through justice perceptions, such that the mediating effect of justice is weaker among employees in the manipulation condition.
4.1. Methods
Participants and Procedure
Working adults within the United States (
n = 292) were recruited via Amazon’s Mechanical Turk system and compensated USD 0.50 for completing the experiment. Participants worked in a variety of occupations and were pre-screened on the basis of working at least part-time (i.e., 20 h a week). In addition, participants were required to be 18 years or older, living within the United States, and report to be proficient in the English language. A total of 35 participants failed a careless responding attention check (e.g., “please select ‘agree’ for this item”) [
39] and were removed from the sample. After the removal of careless responders, the sample consisted of 257 individuals, with the average participant being female (52.3%) and age 35 (18–74). The racial representation of the sample was 74.3% White/Caucasian, 7.4% Black/African-American, 6.6% Hispanic/Latino, 8.9% Asian-Pacific Islander, 1.2% Native-American/American-Indian, and 1.6% identified a race not listed.
The procedure for study 2 followed the experimental between-subject design of study 1; however, organizational justice perceptions and counterproductive work behaviors were included in lieu of the unethical decision-making scale. First, participants completed the short dark triad, organizational justice, and the past CWB scale. Then, participants were randomly assigned to a condition. In the experimental condition, they were given an opportunity to explain their own CWB, or lack thereof, and then administered a brief form of the Raven’s progressive matrices as a filler task. The control condition involved only the Raven’s task. After completing the Raven’s task, both groups then completed an identical organizational justice measure and a CWB intention measure that assessed their future intentions to engage in CWBs.
4.2. Measures
Study 2 used the same scales to measure narcissism (
α = 0.77), Machiavellianism (
α = 0.86), psychopathy (
α = 0.82) [
40], and moral disengagement (
α = 0.91) [
37], and it also used the same set of distractor task problems (Raven’s Progressive Matrices) [
42], as did Study 1. Two additional scales were added to examine perceptions of justice and self-reported counterproductive workplace behaviors.
Perceived overall justice: Overall justice perceptions were assessed with the 12 item Perceived Overall Justice scales [
48] split into two sections (e.g., “For the most part, my organization treats its employees fairly”;
α = 0.89; 0.92) adapted in accordance with Holtz and Harold [
49] to target both the organization and supervisors. The response format was a 7-point Likert scale (anchored from 1 = strongly disagree to 7 = strongly agree).
Counterproductive work behaviors: Past counterproductive work behavior was assessed with the short form 10 item CWB-C-10 checklist [
50] (
α = 0.94; 0.96). Participants were prompted, “how often have you done each of the following things on your present job?” Then, participants were instructed to rate each item on a 5-point Likert scale (1 = never, 2 = once or twice, 3 = once or twice per month, 4 = once or twice per week, 7 = everyday). To measure counterproductive work behavior intentions, the CWB-C-10 scale was adapted by changing the prompt to “how often will you do each of the following things on your present job?”, and, then, participants were instructed to rate each item on a 5-point Likert scale (1 = never, 2 = once or twice, 3 = once or twice per month, 4 = once or twice per week, 7 = everyday).
4.3. Results
Table A6 shows the means, standard deviations, and bivariate correlations between the main variables in study 2. All variables were significantly correlated with each other except for justice perceptions at both administrations with narcissism (
r = −0.01 and −0.03). Before testing our hypothesis for study 2, a series of t-tests were conducted to examine differences between mean levels of justice perception scores between conditions during the first and second administrations of the justice perception scale. There were no significant differences found between conditions for either administration (first administration:
t = 0.39,
df = 254,
p = 0.69; second administration:
t = 0.61,
df = 254,
p = 0.54). In addition, paired sample t-tests were conducted to examine within the condition equality of means across the two administrations of the justice perception scale within the condition. There were no significant differences found through these t-tests either (manipulation condition:
t = −0.72,
df = 256,
p =
0.47; control condition:
t = −89,
df = 127,
p = 0.38).
Our hypothesis for study 2 predicted that the opportunity to morally disengage would moderate the mediating effect of justice perceptions. This hypothesis was tested by examining the extent to which the indirect effects of justice perceptions varied as a function of the condition using a moderated mediation model with the PROCESS macro for SPSS [
51], whereby all variables were mean-centered.
As detailed in
Table A7, justice perceptions did not mediate the role between narcissism and CWB; however, justice perceptions partially mediated the effects of Machiavellianism, psychopathy, and moral disengagement on CWB. The magnitudes of the indirect effects on CWB through justice perceptions varied significantly based on the condition for Machiavellianism and moral disengagement, such that the influence of justice perceptions as a mediator was weakest when the condition was justified. Marginal support was found for psychopathy in the expected direction; however, the bootstrapped confidence interval for the index of moderated mediation crossed zero (
Table A7). Congruent with study 1, no support was shown for the moderation of narcissism’s indirect effect on CWB by moral disengagement practice.
4.4. Discussion
Findings from study 2 partially support our hypothesized model of moral disengagement. Justice perceptions mediated the effects of Machiavellianism, psychopathy, and moral disengagement on CWB, indicating that perceptions of fairness in the workplace explain a significant proportion of the dark personality–CWB relationship. However, the mediating effect of justice perceptions is qualified by a significant interaction, such that the mediating effect of justice perceptions was significantly weaker in the experimental condition than in the control condition.
Dark personality traits and justice perceptions are both studied as antecedents of CWB; however, research on their interaction is notably sparse [
44]. Our results indicate that justice explained less of the personality–CWB relationship when individuals had the opportunity to explicitly rationalize their ethical or unethical behavior. Additionally, our findings suggest that personality traits significantly influence perceptions of justice, with individuals high in dark personality traits perceiving lower levels of organizational justice. This could be due to individuals making non-justice related appraisals salient (e.g., personal values or judgments) that activate a larger pool of rationalizations for engaging in unethical behavior. That is, consciously explaining ethical or unethical behavior (past CWB) may attenuate the role of justice as participants pull from a variety of possible causes outside of injustice.
Indeed, few participants relied on rationalizations expanding on organizational justice as the “cause” for their behavior. Participants commonly pointed to their morals, values, lack thereof, or situational factors (e.g., opportunity for CWB, perceived consequences of CWB). To illustrate, two examples of text responses in the experimental condition are provided below:
“I was taught early in my career that you need to go after things aggressively if you want them bad enough. Sometimes that means stepping on a few people’s toes or doing things that others may think are wrong.”
“For some of the situations, I felt like my morals were being challenged. In other situations, I just felt like being an asshole.”
5. General Discussion
The relationships between dark personality traits and counterproductive work behavior are notably complex, and scientists have called for more nuanced approaches that parse the underlying mechanisms of personality–CWB relations [
44,
52]. The two studies presented in this paper offer investigations of how rationalizations may operate using a moral disengagement framework. In study 1, we demonstrated how individuals’ likelihood of making unethical decisions can change based on how their personality translates into moral disengagement or self-condemnation [
20]. The findings from study 1 contribute to the research literature by showing how dark personality can predispose an individual to morally disengage, thereby leading to an increased (or decreased) likelihood of making an unethical decision. This finding clarifies associations found between personality and moral disengagement in past research [
37] while also demonstrating the effect of personality on changes in levels of unethical behaviors consistent with research on the “slippery slope” effect [
18]. Study 2 extended these findings with results suggesting justice plays a lesser explanatory role when personality-driven cognitive processes are considered.
These studies provide several implications for practice. Given that individuals’ intentions to do unethical behaviors can increase when they rationalize their past decisions, disrupting the rationalization process could be a useful countermeasure for organizations. Establishing a clear and unambiguous code of ethics in organizations can reduce rationalizations by explicitly defining acceptable behaviors and consequences. This clarity fosters a “prevention focus”, which helps employees avoid engaging in risky behaviors. That is, providing salient cues for expectations and/or normative behavior may facilitate employees’ framing of behavior within the expected boundaries of acceptable behavior. Past research has shown that a prevention focus can reduce the extent to which some individuals fall into a “slippery slope” of unethical behavior [
18].
Additionally, our findings also shed light on the impact of both dispositional and situational causes of unethical decision making, as well as their interaction. Study 2 particularly points toward the importance of individual differences when individuals have opportunities to rationalize their behavior. This finding is notable as employee communication channels have increased over the past few decades, with interpersonal interactions occurring in both in-person and virtual means (e.g., office messaging applications). If employees are frequently justifying their behavior (whether good or bad), the predictive role of justice perceptions on CWB may be less powerful, suggesting a stronger need to bring in applicants with(out) particular values and personality traits.
However, this research provides direction for organizations to focus their attention for these types of interventions by demonstrating that individuals possessing high levels of Machiavellianism and subclinical psychopathy are particularly susceptible to showing changes in unethical decision likelihood. Relatedly, this research highlights the importance of leveraging selection to screen out applicants with individual differences that may act as a major driver that predisposes individuals to morally disengage. While not explicitly focused on the decision-making element of unethical behavior, the large literature on manager derailment and personality traits also supports the importance of recruitment and selection on undesirable organizational behaviors [
53].
In addition, there are several implications of our research on future endeavors into moral disengagement and the “slippery slope”. First, this study demonstrates the likelihood of engaging in an unethical decision, such as a CWB, can change across short amounts of time with a simple manipulation that involves asking participants to explain their actions. Cross-sectional research that does not examine multiple timepoints may be limited in fully assessing the extent to which someone may engage in unethical behaviors. Second, these studies demonstrate how the explanatory power of major variables used to predict CWB, such as justice perceptions, can change based on small manipulations to the study design (e.g., being given the chance to explain past behaviors).
These research implications can help guide future research on unethical decision making and CWB. First, it is interesting that the hypothesized role of narcissism was not supported in either study 1 or study 2. Dark triad traits are narrow personality traits in that they target a very particular subclinical personality characteristic but are each a constellation of smaller characteristics. While Machiavellianism and psychopathy have straightforward connections to morality and deviance, narcissism is more closely aligned with self-enhancement, and its role in moral disengagement is less obvious.
Future research should also strive to study models that not only consider individual differences and situational factors but how these factors interact with cognitive processes such as moral disengagement. Furthermore, research should aim to examine the occurrence of CWB over time to fully capture within-person variation. This type of longitudinal data could attempt to examine multiple time points to document how the sense making and rationalizing that took place in our manipulation condition unfold across a longer period even for individuals who were not prompted to explain themselves. As our studies did not provide the potential for employees to perform alternating rationalizations for both desirable and undesirable behaviors, future research that explores the possibly dynamic nature of rationalizations could be particularly valuable.
Finally, future research should attempt to replicate the results from these studies while also addressing our limitations. For example, we primarily examined intentions to engage in CWB or unethical behaviors. Although actual behaviors may be less frequently reported or enacted, the rationalizations for both actual and intended behaviors may not differ significantly [
29]. In addition, longitudinal research should attempt to examine more time points than the “initial” and “subsequent” time points examined in this study to examine the full unfolding of a “slippery slope” effect. The sense making and rationalizing that took place in our manipulation condition may unfold across a longer period even for individuals who were not prompted to explain themselves. In a higher fidelity setting, future research may also be able to document what are the particular vehicles of justification and rationalizations that occur in job experiences, such that an understanding of typical events or processes that prompt those rationalizations may be developed.