The study of organizational fairness has become prominent during the past fifteen years, although it largely began in the 1960s with Adam’s equity theory (Adams 1963). The focus of Adam’s equity theory was on the perceived fairness of the distribution of outcomes (i.e., distributive justice). However, subsequent research focused on the fairness of the process used to determine the outcome (i.e., procedural justice) as a way to further explain individuals’ reactions to perceptions of unfairness (e.g., Folger and Konovsky 1989). While some researchers focused on these forms of organizational justice as predictors of important outcomes, others pursued studies from the opposite perspective as they examined organizational politics (e.g., Ferris and Kacmar 1992; Kacmar and Ferris 1991).

Both distributive and procedural justice and perceptions of organizational politics (POP) have been linked with a variety of organizational and attitudinal outcomes. For example, POP have been linked with reduced job satisfaction (Kacmar et al. 1999; Zhou and Ferris 1995), increased job stress (Cropanzano et al. 1997; Ferris et al. 1996; Kacmar et al. 1999), and turnover intentions (Kacmar et al. 1999). Similarly, distributive and procedural justice have been linked with job and pay satisfaction (Cohen-Charash and Spector 2001), organizational commitment (Konovsky and Cropanzano 1991; Tepper 2001), and reduced intentions to turnover (Masterson et al. 2000). These studies often relied on theories of social exchange to explain responses to organizational justice (e.g., Aryee et al. 2004; Rupp and Cropanzano 2002).

While research has demonstrated the independent influences of politics, distributive, and procedural justice on employee turnover intentions and job satisfaction (Cropanzano et al. 1997; Ferris et al. 2002; McFarlin and Sweeney 1992), little research has examined how these elements of fairness influence each other in predicting outcomes (e.g., Tepper 2001). This is an important topic to study as justice and POP are ubiquitous in organizations and rarely occur in isolation from one another. In the current study, we examine the direct relationships between POP and turnover intentions and job satisfaction and propose that the negative effects from POP may be buffered by both distributive and procedural justice. This study extends past research efforts and fills a gap in the extant literature by simultaneously investigating the effects of POP, distributive justice, and procedural justice on intentions to turnover and job satisfaction.

Social Exchange Theory

The foundations of social exchange theory can be seen in Gouldner’s (1960) discussion of reciprocity. According to Gouldner, individuals are motivated by rational self interest and engage in behaviors that help others in order to create feelings of reciprocity. Extending this idea, Blau (1964) argued that these feelings of reciprocity are the basis of the formation of interpersonal relationships. Individuals respond positively and will reciprocate out of a sense of obligation to the providers. Eisenberger et al. (1986) later applied social exchange theory to organizations and suggested people form relationships with organizations as well as individuals. A number of researchers have since applied social exchange theory as an underlying framework in explaining individual attitudes and behaviors in organizations (e.g., Masterson et al. 2000).

Some researchers have maintained the creation of social exchange relationships can be prompted by organizational justice (Moorman et al. 1998; Rupp and Cropanzano 2002). Research indicates those who perceive high levels of organizational justice reciprocate with OCBs, increased job performance (Aryee et al. 2004), increased job satisfaction, and reduced intentions to turnover (Aryee et al. 2004). Different forms of justice serve as varying sources of the exchange relationship. For example, the supervisor, through fair interpersonal treatment, has been considered the source of interactional justice (Rupp and Cropanzano 2002). The organization, through formal policies and procedures, has been found to be the source of procedural justice. The organization also can be considered the source of distributive justice as outcomes are often determined by policies and procedures. Further, the organization has been considered a political arena in which decisions are based on political behaviors (Pfeffer 1981) and thus also may serve as a source of unfairness. While distributive justice and procedural justice serve to strengthen the social exchange relationship, politics can have the opposite effect and weaken the relationship. In the current study, we view the social exchange as the relationship formed between the individual and the organization and examine the interaction of three aspects of the work environment: procedural justice, distributive justice, and organizational politics, for their effects on job satisfaction and turnover intentions.

Perceptions of Organizational Politics

Organizational politics have been defined as “actions by individuals which are directed toward the goal of furthering their own self-interests without regard for the well-being of others or their organization” (Kacmar and Baron 1999, p. 4). Examples of political behaviors include not following proper procedures, going around the boss, and lobbying high level managers with the intent of gaining rewards such as special work assignments and/or promotions. While these behaviors are not condoned by the organization, they are generally not expressly forbidden (Ferris et al. 2002). Thus, when behaviors occur that are not fully consistent with established organizational rules and policies, yet are not prohibited by the organization, POP can be enhanced. Given that individuals react to their perceptions of reality rather than reality itself, organizational politics often has been studied by asking individuals about their perceptions of these behaviors in their organizations (e.g., Andrews et al. 2003; Ferris et al. 2002).

Politics has often been conceptualized as a workplace stressor because it leads to increased stress/strain reactions (e.g., Ferris et al. 1996, 2002; Harris and Kacmar 2005). Specifically, from an occupational stress perspective, individuals react both psychologically and physically to a political environment that they perceive to be threatening (Ferris et al. 1996). Physical outcomes include fatigue and somatic tension (Cropanzano et al. 1997). Psychological reactions to organizational politics are numerous and have included reduced commitment (Vigoda 2000) and job satisfaction (Bozeman et al. 2001) and increased turnover intentions (Ferris et al. 2002). Based on these arguments and consistent with past research, we predict that:

Hypothesis 1:

POP are positively related to turnover intentions.

Hypothesis 2:

POP are negatively related to job satisfaction.

Moderators of POP––Outcome Relationships

Although POP have been shown to be positively associated with a number of undesired stress-related outcomes, research has shown that the strength of these relationships can be minimized or buffered by different moderators. Early conceptualizations of POP suggested feelings of control and understanding may weaken the relationships between POP and these negative outcomes (Ferris et al. 2002). Ferris and colleagues have argued that those who have feelings of control and/or understanding of the work environment would be more likely to view politics as an opportunity rather than a threat and would react less negatively to organizational politics. Subsequent research provided empirical support for these two variables (e.g., Bozeman et al. 2001; Ferris et al. 1996, 2002). For example, Ferris et al., found that understanding weakens the politics-job anxiety relationship when examined as a moderating variable.

Other research efforts have examined different personality, demographic, and situational moderators of POP-outcome relationships (Andrews et al. 2003; Treadway et al. 2005; Witt et al. 2002). For example, Witt et al. found that the Big 5 dimension of agreeableness moderated the relationship between POP and the contextual performance dimension of interpersonal facilitation such that agreeable individuals maintained a higher level of interpersonal facilitation even when high levels of politics existed. Several studies also have examined situational moderators including perceptions of teamwork (Valle and Witt 2001) and supervisor behaviors (Harris and Kacmar 2005). Specifically, individuals who engage in participative decision making and have high quality relationships with their supervisors experience reduced levels of job strain that results from politics. A number of these predictions were based on the ability of these variables to buffer (minimize) the negative impacts of POP, and cumulatively, these studies have helped to establish boundary conditions related to the negative impact of POP. Similar to these studies, we posit that procedural and distributive justice also may serve to reduce the negative effects of POP as they may provide feelings of control while strengthening an employee’s social exchange with the organization (Folger 1977).

Distributive and Procedural Justice

Distributive and procedural justice have their roots in equity theory (Adams 1963) and are typically subsumed under organizational justice theory (Greenberg 1987). Distributive justice refers to the perceived fairness of the outcomes received by employees, whereas procedural justice refers to the perceived fairness of the processes used to determine the outcomes received (Colquitt et al. 2001). When examined independently, both forms have been found to positively predict job satisfaction (McFarlin and Sweeney 1992) and negatively predict turnover intentions (Cohen-Charash and Spector 2001).

However, in studies where distributive and procedural justice have been examined simultaneously, researchers have noted the differences in the strengths of their predictive abilities (Folger and Konovsky 1989; Harvey and Haines 2005; McFarlin and Sweeney 1992). McFarlin and Sweeney suggested procedural justice is a more potent predictor of organizational outcomes, while distributive justice is a stronger predictor of individual outcomes. Unlike Rupp and Cropanzano (2002), they considered turnover and satisfaction to be individual outcomes rather than organizational outcomes. When examining the interactive effects of procedural and distributive justice, McFarlin and Sweeney found that distributive justice was more strongly related to job satisfaction than procedural justice. The results when both forms of justice have been examined in the same study have been informative, and the few research efforts investigating their joint impact have provided additional insights. Thus, with the knowledge that POP (individual actions that are a workplace stressor) lead to negative outcomes, a logical question is how do these forms of justice, which are organizational actions, interact with the established politics––consequences relationships.

Politics in the Face of Justice

Earlier we stated the creation or demise of social exchange relationships in the workplace can be initiated in several ways. They can be strengthened due to justice or weakened due to POP. However, when simultaneously examining the three workplace variables, the question arises as to whether POP, distributive, or procedural justice is the most salient in forming attitudes. If all three conditions are favorable (i.e., low POP, high distributive and procedural justice) or unfavorable (high POP, low distributive and procedural justice), the point is moot. The issue, however, becomes intriguing when the conditions oppose each other. For example, how do individuals respond when POP are high, but so are procedural and distributive justice? Do they have strong intentions of quitting because they are surrounded by political manipulations, or do they overlook these behaviors since the work environment is generally fair in terms of policies and the distribution of rewards?

We believe it is the latter. Specifically, it is our contention that individuals will respond less negatively to politics when justice exists. The theoretical perspectives of stress buffering (Sutton and Kahn 1986) and social exchange theory guide our predictions. From a stress buffering perspective, the high levels of turnover intentions and low levels of job satisfaction resulting from POP are buffered when procedural and distributive justice are present. Specifically, while those who engage in politics may be disliked by others, the general work environment is still relatively fair in terms of pay and policies.

Further, based on findings by Cohen-Charash and Spector (2001) that distributive justice is a stronger predictor of job satisfaction than procedural justice, we propose POP will have the weakest relationship with job satisfaction when distributive justice is the highest. Based on the norm of reciprocity (Gouldner 1960), and its assumption that individuals act according to their own rational self interest, POP should be relatively unimportant when there is no threat to the fairness of distributed rewards (i.e., high distributive justice). The actions that occur in the organization may be unfair, and run counter to organizational policy, but if these actions do not directly affect the individual, then his/her level of job satisfaction and intent to turnover will not be as strongly impacted. An individualistic mentality of “I got my raise/bonus, so it (i.e., politics) doesn’t bother me” may prevail. Thus, although we expect that either form of justice alone can reduce the negative effects of politics, we believe that distributive justice will exert the most influence and that the existence of both forms of justice together will provide a much stronger buffer. Hence, we predict:

Hypothesis 3:

The relationship of POP to turnover intentions varies as a function of the levels of procedural and distributive justice. Specifically, when both forms of justice are higher, the positive relationship between POP and turnover intentions is weaker.

Hypothesis 4:

The relationship of POP to job satisfaction varies as a function of the levels of procedural and distributive justice. Specifically, when both forms of justice are higher, the negative relationship between POP and job satisfaction is weaker.

Method

Sample

The sample consisted of 311 full time employees (response rate of 46%) from a water management district in the southeastern United States. The respondents were primarily male (57%) and Caucasian (83%). The ages of the respondents ranged from 22 to 76 with a mean age of 41 years. Due to concerns about respondent anonymity, tenure was measured categorically with responses being grouped in three year blocks (i.e., 0–2 years, 3–5 years, 6–8 years, and so on). Respondents indicated that their organizational tenures ranged from 0 to more than 18 years, with the largest majority of subjects reporting tenure between 6 and 8 years.

Procedure

The survey was conducted at the request of the company’s top management who were interested in determining employees’ feelings about the organization. All employees were notified by company memo of the opportunity to participate in the survey. Participation in the survey was voluntary and respondents were assured of the confidentiality and anonymity of their responses. To adhere to these promises, two of the authors held multiple sessions on site in which the surveys were administered. Surveys were administered during regular working hours in groups of up to 50 over a period of three days. Upon arrival the employees were told the purpose of the study and given a survey to complete. To protect those with unique characteristics, respondents were asked not to supply their names and we collected limited demographics. Upon completion respondents placed completed surveys in a collection box. At the conclusion of the study, two of the authors held multiple feedback sessions with employees to describe the findings and explain how they would be used.

Measures

All of the survey items were responded to on a 5-point Likert scale. The anchors for the scale were strongly disagree (1) to strongly agree (5). The items in the scales were averaged to create an overall mean for each variable. The items were coded such that high values represent high levels of the constructs.

Perceptions of Politics

Perceptions of organizational politics were measured with the fifteen-item (α = .87) Perceptions of Organizational Politics Scale (POPS) (Kacmar and Carlson 1997). A sample item from this scale was “People in this organization attempt to build themselves up by tearing others down.”

Procedural Justice

Procedural justice was measured with Moorman’s (1991) seven-item scale (α = .82). A sample item from this scale was “There are procedures in this organization that allow an opportunity to appeal or challenge a decision.”

Distributive Justice

Distributive justice was measured with Price and Mueller’s (1986) six-item scale (α = .93). A sample item from this scale was “I am fairly rewarded considering the responsibilities I have.”

Intent to Turnover

Intent to turnover was measured with Seashore et al. (1982) three-item scale (α = .84). A sample item from this scale was “I will probably look for a new job in the near future.”

Job Satisfaction

Job satisfaction was measured with Cammann et al. (1979) three-item scale (α = .71) from the Organizational Assessment Questionnaire. A sample item from this scale was “All in all, I am satisfied with my job.”

Control Variables

In this study we controlled for tenure, gender, race, age, and department. We included tenure, gender, race, and age as biographical control variables in this study because previous research has shown these variables to be related to turnover intentions and job satisfaction (Hom and Griffeth 1995; Rhodes 1983; Spector 1997). As previously mentioned, tenure was measured categorically in three year blocks. Gender was coded with men as 1 and women as 2. Age was measured in number of years. Race was coded with non-Caucasians as 1 and Caucasians as 2. We controlled for department (compliance, environmental issues, and permits), via dummy codes, to eliminate any differences that may be due to respondents working in different departments of the organization.

Analyses

We conducted hierarchical moderated regression analyses to assess the form and magnitude of the relationships between POP, distributive justice, and procedural justice, and the two dependent variables, turnover intentions and job satisfaction. In the first step, we entered the control variables of tenure, gender, race, age, and department. In the second step, we entered the centered main effects for POP, distributive justice, and procedural justice. In the third step, we entered the three different two-way interactions between the centered variables entered in the second step. Finally, in the fourth step, we entered the three-way interaction term between the centered variables for POP, distributive justice, and procedural justice.

Results

Means, standard deviations, and intercorrelations among the variables of interest are provided in Table 1. Results from the hierarchical moderated regression analyses are provided in Table 2. We begin with the results for turnover intentions. In the first step, gender and age were both negatively and significantly related to turnover intentions indicating males and those individuals who were older were less likely to turnover. Tenure, race, and department were not significantly related to turnover intentions. In step 2, POP were positively and significantly related to turnover intentions while both distributive and procedural justice were negatively and significantly related to intentions to turnover. This finding provides support for hypothesis 1. In the third step, all three of the 2-way interaction terms created between POP, distributive justice, and procedural justice were not significantly related to turnover intentions.

Table 1 Correlations, means, and standard deviationsa
Full size table
Table 2 Hierarchical moderated regression analyses of the 3-way interactions among politics, distributive justice, and procedural justicea
Full size table

Finally, results from the fourth step show that the three-way interaction term was negatively and significantly related (ΔR2 = 1.0%) to turnover intentions. To graphically illustrate the 3-way interaction, we plotted the interaction of POP and procedural justice at high (+1 SD) and low (−1 SD) levels of distributive justice (see Fig. 1a, b). The simple slopes tests (Aiken and West 1991) for each line were significantly different from zero (low distributive justice: t = 4.36, p < .01 for low procedural justice and t = 4.81, p < .01 for high procedural justice; high distributive justice: t = 6.34, p < .01 for low procedural justice and t = 5.86, p < .01 for high procedural justice). As shown in Fig. 1a and b, turnover intentions are lower under higher levels of distributive justice for three of the four points (from the interaction of POP and procedural justice). In particular, turnover intentions for individuals higher in POP, higher in distributive justice, and lower in procedural justice were slightly higher than the turnover intentions for individuals lower in distributive justice, lower in POP, and higher in procedural justice, but lower than the turnover intentions for all other individuals who perceived lower levels of distributive justice. When distributive justice was higher, turnover intentions rose when politics were present, but the rise was higher under lower procedural justice conditions. Finally, when distributive justice was lower, turnover intentions fell when politics were lower and procedural justice was higher, but remain fairly stable when both POP and procedural justice were lower. These findings provide support for Hypothesis 3 as the positive relationship between turnover intentions and POP is weaker when procedural justice is high as long as distributive justice is also high.

Fig. 1
figure 1

(a, b) Interaction between perceptions of politics and procedural justice on turnover intentions at high and low levels of distributive justice

Full size image

The results for the three-way interaction predicting job satisfaction also are provided in Table 2. In step 2, POP were negatively and significantly related to job satisfaction, providing support for hypothesis 2, whereas distributive justice was positively and significantly related to job satisfaction and procedural justice was not significantly related to job satisfaction. In the third step, all three of the two-way interaction terms were significantly related to job satisfaction. Finally, in the fourth step, the three-way interaction term was positively and significantly related (ΔR2 = 1.4%) to job satisfaction. Figures 2a and b provides the graphical representation of the three-way interaction for job satisfaction. The simple slopes tests (Aiken and West 1991) for each line were significantly different from zero (low distributive justice: t = −3.58, p < .01 for low procedural justice and t = −4.11, p < .01 for high procedural justice; high distributive justice: t = −3.38, p < .01 for low procedural justice and t = −3.46, p < .01 for high procedural justice). As Figure 2a and b show, satisfaction levels are higher under higher levels of distributive justice regardless of the levels of politics and procedural justice and lower under low distributive justice conditions regardless of the levels of POP and procedural justice. When distributive justice is higher, satisfaction levels drop when politics are present. Finally, when distributive justice is lower, satisfaction levels rise when POP are lower and procedural justice is higher, but remain fairly stable when both politics and procedural justice are lower. These findings provide support for Hypothesis 4 as the negative relationship between job satisfaction and POP is weaker when procedural justice is high as long as distributive justice is also high.

Fig. 2
figure 2

(a, b) Interaction between perceptions of politics and procedural justice on job satisfaction at high and low levels of distributive justice

Full size image

Discussion

Results of the present study, which used social exchange theory as an overarching guide, as well as aspects from the stress-buffering framework, provide support for the importance of examining justice at the same time as POP (Folger and Cropanzano 2001). Supporting our first two hypotheses, POP were significantly related to turnover intentions and job satisfaction. These findings are in line with established research on POP as a workplace stressor, and make the examination of moderating (potentially buffering) variables all the more necessary.

In terms of our 3-way interactions, as predicted, working conditions that define the most fair work environment, when justice is high and politics are low, produced the lowest levels of turnover intentions and the highest levels of job satisfaction. Further, working conditions that define the most unfair (i.e., low justice and high politics) environment produced the highest levels of turnover intentions but not the lowest levels of job satisfaction. The lowest levels of job satisfaction occurred under low distributive justice and high POP and procedural justice. A possible reason for this unexpected finding relates to the match of high politics and high procedural justice without the buffer of a fair distribution of outcomes. Essentially, political actions may be prohibiting fair procedures from being followed. Thus, individuals may feel frustrated and dissatisfied with their jobs as they have little hope that things will improve.

Although it is interesting and informative to identify the extreme conditions, a more intriguing aspect of this study was its ability to determine how organizational justice interacts and either neutralizes and or increases the unwanted outcomes that result from POP (Ferris et al. 2002). Further, we wanted to determine whether distributive or procedural justice had a stronger impact on the politics––turnover intentions or politics––job satisfaction relationships. For both outcomes, distributive justice exerted the strongest influence. These results suggest that people are more satisfied and more willing to remain in an organization that has a political environment, even when the process by which rewards are determined are unfair, as long as they receive rewards. These same individuals are more likely to be dissatisfied and to leave a political organization when they are not rewarded regardless of how fairly the rewards are determined. Apparently, our respondents care more about the fairness of the actual distribution of outcomes they receive than the processes used to determine the outcomes (Greenberg 2001).

With respect to the results when distributive and procedural justice are moving in different directions, the outcome of job satisfaction is primarily in line with previous research. In particular, numerous researchers have noted that personal evaluations (e.g., job and pay satisfaction) are influenced more strongly by distributive justice rather than procedural justice (e.g., Sweeney and McFarlin 1993). However, the general notion is that procedural justice plays a stronger role in organizational outcomes (Brockner 2002; Rupp and Cropanzano 2002). Our study showed that in predicting turnover intentions the opposite was true as distributive justice played a larger role. Although, our findings may seem to contradict the extant literature, recent research has reexamined this assumption and often arrived at a different conclusion. More specifically, Cohen-Charash and Spector (2001) performed a meta-analysis and found that both forms of justice were equally related to turnover intentions. Additionally, Skitka et al. (2003) concluded that distributive justice is just as, if not more important than procedural justice, and that researchers need to try to determine which forms of justice relate stronger to which outcomes. Thus, our results for the interactions of these variables show that distributive justice is a potentially more powerful buffer than procedural justice.

Strengths and Weaknesses of the Present Study

Our study possesses a number of strengths. First, we extended research on organizational justice (Greenberg 1987) by including POP. This is important as these are related constructs (Nye and Witt 1993) that need to be investigated in the same study. Further, we used organizational justice as a means of determining how bad the negative outcomes will be from POP. Second, we investigated the interactive effects of both distributive and procedural justice, an analysis that has been called for by researchers (e.g., Brockner 2002; Harvey and Haines 2005; Tepper 2001), but rarely performed. Third, we found the results in this study even after controlling for tenure, gender, race, age, and department. This adds strength to our findings and likely provides a more conservative estimate of the relationships. Finally, we investigated a three-way interaction, which is more representative of the real world, as variables do not truly exist only as main effects, and found support for our hypotheses.

Although there are definite strengths in this study, there are a few limitations that must be acknowledged to properly interpret the study’s results. First, our data were cross sectional which could introduce problems related to common method variance (CMV) (Podsakoff et al. 2003). However, previous researchers have noted that it is unlikely that CMV would explain the data fitting multiple a priori interaction hypotheses (e.g., Duffy et al. 1998). Additionally, a number of researchers have noted that CMV actually reduces the likelihood of detecting interaction effects (Wall et al. 1996). Thus, CMV is most likely not a serious limitation of this study. Second, the sample was primarily Caucasian, thereby limiting the generalizability of these findings to other more diverse samples. Finally, as previously mentioned, the two three-way interactions explained relatively small amounts of variance, although not less than the percentage expected from interactions in organizational research (Champoux and Peters 1987).

Practical Implications

A number of practical implications for managers emerged from this study. First, managers need to make every attempt to minimize POP. This is important, as regardless of the levels of distributive or procedural justice, the most desirable outcomes (low turnover intentions or high job satisfaction) occurred when politics were low and the most undesirable when politics were high. With respect to minimizing politics, research has shown that POP are more likely to occur in ambiguous and uncertain work environments. Thus, by implementing, communicating, and following clear, unambiguous policies and incentive structures, managers can reduce political perceptions. In addition, political behaviors can be extinguished by ensuring that employees who engage in self-interested actions that characterize politics are not rewarded.

When focusing on the two aspects of justice examined in the present study the practical applications are easy to see. Similar to politics, managers need to try to maximize perceptions of both distributive and procedural justice. This can be accomplished by encouraging and enforcing clear policies and procedures in determining organizational rewards. If, however, organizational or resource limitations placed on managers do not allow them to maximize both types of justice, our results offer some guidance for the type of justice on which to focus. In particular, our results suggest that organizations may experience better outcomes (i.e., decreased turnover intentions and higher levels of job satisfaction) when the distribution of rewards is fair (high distributive justice) even if the procedures are somewhat unfair (low procedural justice). Our findings, though not consistent with previous research on social exchange relationships, provide tentative evidence that there are situations (i.e., highly political environments) where distributive justice may be more important in leading to desired outcomes than procedural justice.

Directions for Future Research

The results of this study point to a number of directions for future research. The first recommendation is to investigate not only distributive and procedural justice, but also interactional justice (Byrne and Cropanzano 2001). Interactional justice, which pertains to the human side of organizational practices and relates to the components (i.e., honesty, respect, and politeness) of the communication process between the source and the recipient of justice (Byrne and Cropanzano 2001), needs to be included to more accurately discriminate between the different types of justice and hopefully provide new insights as to their importance.

Second, there is a need to expand the outcome variables investigated. Some outcomes that could be candidates for future research include job performance, stress, job involvement, and organizational citizenship behaviors. Finally, we suggest that future researchers extend the findings from this study. We found that overall, employees are more likely to react negatively to perceived politics when it affects them personally (i.e., such as an unfair distribution of rewards). Although this is an interesting finding, it only raises additional questions. In particular, it would be interesting to learn whether negative outcomes result only from self-interest or the unfairness of outcomes that personally affect an employee, or if injustice towards others leads to similar negative consequences.