1
Anger and Sadness as mediators of the relationship between organizational
injustice and deviant behaviours
Robin Hulsman 10034366 Master Thesis Final version: 26-08-2015 MSc. in Business Administration Leadership and Management Track
University of Amsterdam The Amsterdam Business School
Supervised by Stefan Mol Renske van Geffen
2
Statement of originality
This document is written by Student Robin Hulsman who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Contents 1. Introduction 5 2. Literature Review 7 2.1 Injustice 8 2.2 Workplace Deviance 12
2.3 Injustice and Deviance 13
2.3.1 Distributive Injustice and Deviance 14
2.3.2 Procedural Injustice and Deviance 14
2.3.3 Interactional Injustice and Deviance 16
2.3.4 Interaction among Injustice Components and Deviance 16
2.4 Anger and Sadness 18
2.5 Workplace Deviance over Time 20
2.6 Research Model 23
3. Research Methodology 25
3.1 Measures of variables 25
3.2 Sample and Statistical Procedure 26
4. Results 28
4.1 Analytical strategy 28
4.2 Results linked to the hypotheses 29
5. Discussion 38
5.1 Influence organizational injustice on anger and sadness 38
5.2 Mediating effects of anger and sadness 41
5.3 Workplace deviance over time 44
5.4 Theoretical and practical limitations 45
5.5 Limitations 46
5.6 Further research 47
6. Conclusion 47
7. References 50
4 Abstract
This study investigated whether employees’ anger and sadness mediated the relationship between distributive, procedural and interactional injustice on deviant workplace behaviour over time. This is important as employers and employees will often depend on each other. Having a meaningful employer-employee relationship can lead to greater employee happiness, increased productivity, and efficiency of employees, less conflict and more loyalty. In order to reap these benefits, one always has to keep in mind the dynamics of this relationship. This study can give organizations some insight in these processes. By means of a longitudinal research design, the change in (deviant) behaviour over time of the employees is measured. In this way the potential positive impact of justice and the (negative) consequences of mistreatment on employee attitudes and behaviour, will be elucidated. This study was done by means of Latent Growth Structural Equation Modelling. Data were gathered for six weeks by weekly online surveys. The study started with 215 respondents but because of the high attrition rate ended with 89 respondents. The three separate forms of injustice together, as well as separate, turned out to be directly and significantly related to the workplace deviance slope, as well as to anger and sadness in wave 1. Higher levels of anger and sadness in wave 1 and 5 lead to higher levels of deviant behaviours over time. Anger and sadness however turned out not to be significant mediators of the relationship between the three forms of injustice (together as well as separate) and deviant workplace behaviour over time.
5 1. Introduction
For an employer, hiring a new employee is not just about bringing a new person aboard, it is starting a new relationship with someone. Employers and employees will often depend on each other. Having a meaningful employer-employee relationship can for example lead to greater employee happiness, increased productivity, and efficiency of employees, less conflict and more loyalty (O'Malley, 2000; Boxall, 2013). In order to reap these benefits, one always has to keep in mind the dynamics of this relationship.
In other words, one can say that the value that an employee has for the organization derives in part from his or her relationship with his or her employer. The success of an organization depends upon the energy and efforts of its employees. Barney (1991) stated that in order for a resource to be a source of sustainable competitive advantage, it has to possess four characteristics. First, the resource must be valuable for the organization (this could be by differentiating the product or service or through reducing costs). Second, it has to be rare. Third, it must be almost inimitable and fourth, the firm needs an appropriate structure and systems that can effectively and efficiently utilize the resource (Barney, 1991; Barney, 1995). Human resources can be the root of sustained competitive advantage because they meet these four criteria. However, human resources are characterized by social complexity, causal ambiguity, and unique historical conditions. Therefore not all firms can successfully develop human resources as a sustained competitive advantage by imitating the HR practices of firms that have successfully developed human resources (Wernerfelt, 1984; Barney 1991; Conner 1991). Human resources need to be integrated into the formulation stage of the strategy of a firm. The resource-based approach can be seen as a framework for examining the potential for a given pool of human capital resources to carry out a given strategy. This is why the relationship of the organization with its employees can make the difference for an organization to obtain and sustain a competitive advantage (Wernerfelt, 1984; Barney 1991; Conner 1991).
This relationship is reciprocal; on the one hand employee behaviour depends on how much attention an organization pays towards its employees’ welfare and interests. On the other hand, when employees feel their needs are fulfilled, they will show serious efforts to achieve organizational objectives (Blau, 1964; Gouldner, 1960; Holmes, 1981).
Following from what is stated above, there is evidence to suggest that this reciprocal behaviour of employees can be negative and even counter-productive when employees
6 experience feelings of injustice. This means that they have the feeling that the organization is not giving them what they deserve or need (Douglas et al., 2002; Sackett, 2002). When employees keep showing negative and counter-productive work behaviour (CWB) for a longer period of time, it can result in the deterioration of organizations. Specifically, these deviant behaviours are deliberate acts that damage the organization and have a pernicious effect on how the organization is functioning (Aquino, Lewis & Bradfield, 1999; Aquino, Galperin & Bennett 2004; Bordia, Restubog & Tang, 2008; Farhadi, Fatimah, Nasir, & Wan Shahrazad, 2012; Henle, 2005).
There are different factors that can cause employees to exhibit deviant behaviours. The focus in most studies has been on distinguishing environmental antecedents of deviant behaviours and identifying personality traits that may increase an employees’ propensity to engage in such behaviours (Aquino et al., 1999; Skarlicki, Folger & Tesluk, 1999). The social cognitive view of individuals underlines the interaction between persons and specific situations (Rosenthal & Zimmerman, 2014; Breithaupt, 2012). The cognitive appraisal model can help explain why different individuals react with different emotions to the same event. Lazarus (1991) describes some core relational themes that relate to particular emotional responses. Anger is said to be a natural reaction to a situation in which one feels neglected, harmed or threatened. Sadness, on the other hand, appears to be one’s natural reaction to events in which someone feels helpless and lost. A contribution of this study is to look at the emotions anger and sadness as mediators between the organizational injustice and workplace deviance relationship.
It is showed that organizational injustice leads to organizational deviance (Aquino, Lewis & Bradfield, 1999; Ambrose, Seabright & Schminke, 2002), but to have a closer and more precise look at this relationship, theoretically it is important to capture people’s within-person changes of deviant behaviour over time. By measuring employees’ behaviour over time, hopefully changes or developments in the characteristics of the target population at both the group and individual level will be detected. This longitudinal study extends beyond a single moment in time by which sequences of events can be established and which will hopefully result in more reliable results.
7 For organizations it is interesting and important to see how the emotions anger and sadness influence the relationship between organizational injustice and deviant behaviour of their employees. In this way the potential positive impact of justice and the (negative) consequences of mistreatment on employee attitudes and behaviour, will be elucidated. The ‘damage’, exhibited in the form of deviant behaviours on the part of employees as a response to experiencing mistreatment, can vary from absenteeism to sabotage depending upon their personality dispositions (Paunonen & Ashoton, 2001; Ones & Viswesvaran, 1996). This study, by examining the link between people’s emotions and deviant behaviour over time, could help organizations manage the behaviours expected of their employees.
Within this study, these mediated effects will be examined by means of Latent Growth Curve Modelling analysis. By means of a longitudinal research design, the change in behaviour over time of the employees will be measured (Duncan, 2004).
In order to reach a comprehensive conclusion, the rest of this study is structured as follows: the next chapter describes the current state of the art with respect to perceived injustice, workplace deviance and individual dispositions. Subsequently, in chapter three the data collection procedure and research method are outlined. Results based on the collected data are discussed in chapter four. Next the most important conclusions and implications of the results of this study are discussed in chapter five, together with a number of suggestions for further research and the most important limitations. Finally, this study ends with a short overall conclusion.
2. Literature Review
This chapter discusses the most relevant findings from the current literature about deviant behaviours in the workplace, organizational injustice, and individual dispositions of employees. Following from the literature, it also presents the hypotheses of this study. The most relevant key concepts related to deviant behaviours are discussed. First injustice, then workplace deviance, and finally individual dispositions will be described. It will be outlined how the mediators anger and sadness transmit the effect of injustice perceived by the employees on employees’ deviant behaviours. The chapter ends with a conceptual model, which illustrates the stated hypotheses graphically.
8 2.1 Injustice
There is evidence to suggest that organizational injustice is positively related to employees’ deviant behaviours, which in turn hamper organizational performance (Fox et al., 2001). It sounds obvious that organizations strive to avoid the occurrence of incidents that lead to employees having negative emotions as much as possible, however organizations do and cannot always succeed to avoid these situations. As a result, organizations need to put effort and time in settling these situations.
Organizational environments that foster justice and a fair treatment are more likely to nurture success. Justice can be defined as a system of standards and beliefs, which describe appropriate relationships between people and their fates (Wenz, 1988). Justice influences the relationship between individuals and the society or organization to which they belong, by encouraging people to regulate their own behaviour instead of being constrained by others (Wenz, 1988). Creating a just environment can be operationalized in terms of legal procedures but also less formally through shared values (e.g. equality) and norms (e.g. reciprocity) (Wenz, 1988). Therefore organizations should understand the positive impact of justice and the (negative) consequences of mistreatment on employees’ attitude and behaviour. This study aims to show organizations, and help them understand, the significance of fair treatment of employees and how mistreatment may be avoided.
There are different forms of organizational injustice; distributive, procedural and interactional injustice. Distributive injustice is related to the unfairness associated with decision outcomes and the distribution of resources. The outcomes of resources distributed can be tangible (e.g., pay) or intangible (e.g., praise). When outcomes are perceived to be equally applied, perceptions of distributive justice can be fostered (Adams, 1965). Procedural injustice is about the unfairness of the processes that lead to certain outcomes. When individuals have the feeling that they have a voice in the process or the process involves characteristics as ethicality and lack of bias then perceptions of procedural justice are enhanced (Leventhal, 1980).
Interactional justice involves the treatment that an individual receives as decisions are made (Bies & Moag, 1986). This form of justice can be promoted by facilitating explanations for decisions made and bringing the news with predisposition and respect (Bies & Moag, 1986). Colquitt (2001), in his construct validation study, suggests that interactional justice should be split in two components. These would be interpersonal and informational justice.
9 Interpersonal justice should reflect the degree to which people are treated with dignity, respect, and politeness by third parties or authorities involved in carrying out procedures or defining outcomes. The second, informational justice, focuses on the explanation provided to people that convey information about why procedures were used in a specific way or why outcomes were distributed in a certain way.
There is evidence to suggest that organizational justice impacts employees’ actual deviant behaviours (Goldman, 2003; Aquino et al., 2004). While on the one hand for example having a voice would overall influence the perception of (in)justice of the employees as some sort of characteristic of the organization, on the other hand personality dispositions will also result in variation in justice perceptions across employees. Jones (2009) showed that procedural justice explains unique variance in organizational deviant behaviours. Adams, already in 1963, in his equity theory, propounded that people do not simply become dissatisfied because they are disgruntled with the injustice in their organization; instead they react in predictable ways. As a result of perceived distributive injustice, employees might desire to get back at the malefactor. Indeed, Hershcovis et al. (2007) found that interpersonal mistreatment by supervisors has a strong correlation with supervisor-targeted aggression. It has also been shown that mistreatment by the organization or colleagues, enhances the desire for revenge to this colleague or the organization by the mistreated person (Bordia et al., 2008; Jones, 2009). It has been showed that organizational injustice is associated with deviant behaviours, but a comprehensive insight in the process by which this relationship develops is lacking in literature.
Lazarus (1991), states that the cognitive appraisal model defines emotions as “ways of apprehending states of the world that have significance for personal well-being” (p.89). This means that emotions are our response to things we experience. Emotions govern our actions to the losses or gains we experience in meaningful events. The cognitive appraisal has two stages; first, one assesses the core meaning of the experience, which means that one thinks about the gains and losses emerging from the event. Second, one thinks about the appropriate actions required to respond and cope with what one experienced (Lazarus, 1991). When an event does not have any personal meaning, no emotional response will emerge independent of the outcome of the event. The second stage pertains to how to restore personal well-being by expressing a certain emotion and enacting a certain concomitant behaviour. This can for
10 example be getting angry and seeking retaliation. The emotion and behaviour a person shows, is influenced by the losses a person experiences in combination with the gains.
The cognitive appraisal model can help explain why different individuals react with different emotions to the same event. Lazarus (1991), describes some core relational themes that relate to particular emotional responses. Anger is said to be a natural reaction to a situation in which one feels neglected, harmed, or threatened. In anger as an emotion, assigning blame to a specific agent is a critical defining feature (Lazatus, 1991). Sadness, on the other hand, appears to be one’s natural reaction to events in which someone feels helpless and lost. For Sadness, no specific agent is perceived to be responsible for the loss.
Lazarus proposes that it is, “possible for attributions and control, therefore appraisals of blame, to change from moment to moment as the person grieves over the loss” (Lazarus, 1991: 829-830). As a consequence, the emotional response would also change, either from anger to sadness, or vice versa (Lazarus, 1991). Applications of Weiss and Cropanzano’s (1996) affective events theory (AET), have found support for an interaction effect between an individuals’ perceived outcome favourability, regarding a specific event, and procedural or interactional justice when predicting a variety of discrete emotions.
AET (Weiss & Cropanzano, 1996) provides a theoretical basis for the relationship between emotions, workplace events, and subsequent behaviours. Building on this theory, this study will investigate the relationship between organizational injustice and deviant behaviours. Not only the Injustice-Deviant behaviours relationship will be tested, but also the underlying mechanisms and processes (i.e., anger and sadness, through which perceptions of injustice leads towards employees’ deviant behaviours) will be explained to answer the why’s and how’s of these relationships.
As said earlier, distributive injustice is related to fairness associated with tangible and intangible decision outcomes, and how resources inside an organisation are distributed (Adams, 1965). Anger is a natural reaction to a situation in which one felt neglected, harmed or threatened in which a specific agent is blamed. Sadness, on the other hand, is a natural reaction to a situation in which someone feels helpless and lost, with no specific agent responsible for this.
The distribution of tangible resources, e.g. pay, or intangible resources, e.g. praise, can in most cases be directly assigned to someone, ‘an agent’. This agent is the person who
11 distributes the resource. As there is a ‘visible’ specific agent, distributive injustice will probably have a more positive influence on employees’ anger than sadness.
In a company, for example, there is almost always a specific supervisor who is responsible for an employees’ pay. When the employee is dissatisfied about his or her salary, it is easy for the employee to see the supervisor as the villain, and become angry with him or her. At the same time, when a colleague gives you a compliment, it is easy to appreciate this and to thank the colleague and to see him or her as the person who is on your side.
H1: The positive relationship between Distributive Injustice and employees’ Anger is stronger than the positive relationship between Distributive Injustice and Sadness.
Procedural injustice is about the unfairness of the processes that lead to outcomes. When individuals have the feeling that they have a voice in the process or the process involves characteristics as ethicality and lack of bias then procedural justice is enhanced (Leventhal 1980). As this is about processes, in which many people are involved and which goes on over the time, no specific agent can be blamed. A process can be seen as a series of changes, actions, or functions bringing about a result. Processes mostly have been around for years and people have no recollection of who designed these processes in the first place. Processes consist of a sequence of interdependent and linked procedures, which at different stages consume different resources as employee time and energy. This is done to convert inputs, for example data, into outputs. These outputs serve as inputs for the next stage until the desired result is reached. This means that a process will probably consist of different stages, which ‘use’ the energy and time of multiple employees. So to conclude, as there is not one clear agent, procedural injustice will probably have a stronger positive influence on employees’ sadness compared to anger.
H2: The positive relationship between Procedural Injustice and employees’ Sadness is stronger than the positive relationship between Procedural Injustice and Anger.
Interactional justice is about the treatment that an individual receives as decisions are made (Bies & Moag, 1986). This form of justice can be promoted by facilitating explanations for decisions made and bringing the news with predisposition and respect (Bies & Moag, 1986).
12 Within this the two components of Colquitt (2001) are important. The first component, interpersonal justice, reflects the level to which people are treated with dignity, respect, and politeness by authorities, or third parties, involved in carrying out procedures or determining outcomes. The second, informational justice, focuses on the explanation given to people that conveys information about why procedures were used in a specific way or why outcomes were distributed in a certain way.
As someone has to facilitate this explanation and people can clearly see and experience who is interacting with them in what way, it is in this case easy to blame an agent for breaches of interactional justice. Therefore interactional injustice will have a stronger influence on employees’ Anger compared to Sadness.
H3: The positive relationship between Interactional Injustice and Anger is stronger than the positive relationship between Interactional Injustice and Sadness.
2.2 Workplace Deviance
Empirical studies have provided support for the effect of injustice on deviant behaviours that target both the organization and its members (Robinson & Benett, 1995; Ambrose, Seabright & Schminke, 2002). For example, in a study that explored the unique effects of justice components on different forms of deviant behaviour, Ambrose et al. (2002), showed that when employees perceive procedural injustice, they were inclined to sabotage their organization. In another study, procedural injustice led employees to show aggression against their supervisors (Greenberg & Barling, 1999).
Different theories of interactional, procedural, and distributive injustice show that unfair treatment of employees does not only evoke negative work attitudes and emotions, but can also lead to deviant workplace behaviour. The literature is full of empirical evidence supporting this line of thinking (Ambrose et al., 2002; Appelbaum, Deguire & Lay, 2005; Greenberg & Alge, 1998; Mitchell & Ambrose, 2007).
Deviant behaviour has been studied from several perspectives by using different, partially overlapping, terms to explain these detrimental behaviours. This includes deviance (Hirschi & Gottfredson, 1994; Allen & Lee, 2002), aggression (Douglas et al., 2001) and revenge (Bies, Tripp, & Kramer, 1997; Bies & Tripp, 2005). The extent to which employees show deviant
13 behaviour, can differ from less damaging to very intensive deviant behaviours. Two primary types of workplace deviance can be distinguished, namely Interpersonal Deviance, which is targeted at organizational members and Organizational Deviance, which is targeted at the organization itself (Robinson and Bennett, 1995). Interpersonal Deviance includes behaviours such as being rude to a colleague, or saying something hurtful, whereas Organizational Deviance includes acts such as stealing and withdrawal (Robinson and Bennett, 1995). These two primary types of workplace deviance are assessed separately in this study and will together be part of the research model in this study.
Different studies and theories differ in the elements of justice they emphasize, as this depends on the research context. Investigations range from just one component or measure of justice as predictor in combination with a specific form of deviant behaviour as the outcome variable (Aquino et al., 2004; Berry, Ones, & Sackett, 2007), to multiple justice components as predictors, and a combination of different forms of workplace deviance as the outcome variable (Skarlicki & Folger, 1997;McCardle, 2007).
But what is this deviant behaviour exactly? Deviant behaviour ranges from physical violence against people to milder forms of aggression such as verbal aggression and other forms of mistreatment towards other people. Deviant behaviour also includes acts directed towards the organization instead of its employees. This organizational form of deviant behaviour includes performing work inaccurately, withdrawal, misusing organizational property, or not notifying superiors about organizational problems. Deviant behaviour, therefore, can be separated into behaviour that is directed towards other employees and behaviour that is directed toward the organization as a whole (Bennett and Stamper’s, 2001).
Below, the hypotheses pertaining to the relationship between the different forms of injustice in combination with workplace deviance will be developed and explained.
2.3 Injustice and Deviance
Having explained the theories on injustice and on deviant behaviour, below these different perspectives will be linked. Consecutively distributive, procedural, and interactional injustice will be linked to theories of deviant behaviour.
14 2.3.1 Distributive Injustice and Deviance
Distributive (in)justice has primarily been studied from the perspective of Equity theory. Adams’ (1963) equity theory, proposes that every individual needs to maintain a view of their ‘organizational’ and social worlds as equitable and predictable places. People evaluate the fairness of every outcome distribution compared to their contributions to an event or situation, against that of a referent (Adams, 1965; Cropanzano & Greenberg, 1997). When people see the outcome allocation as unfair, perceptions of injustice are provoked. Besides psychological distress, this can also evoke a behavioural response among individuals. This means that people do not only ‘express’ their dissatisfaction when they have the feeling that outcomes are not allocated fairly, but they may also have a certain behavioural reaction. Deviant behaviour is one of the reactions one can have.
These deviant behaviours can be ‘direct’ (e.g., withdrawal) or more ‘symbolic’ (e.g., attacking a colleague personal) (Robinson & Bennett, 1995). Early research on the distributive form of injustice shows that when people have the feeling that outcomes are not fair, it is a primary motivation for people to show different kinds of deviant behaviours. In the case of distributive injustice, it is expected that these deviant behaviours are carried out because people want to ‘restore equity’ (Adams, 1963). Research has shown that distributive injustice was an essential cause for employees to sabotage or commit theft (Ambrose et al., 2002). This because they had the feeling that the organization ‘owed them’.
Equity theory has contributed to our understanding of justice and deviance research but the theory only focuses on how people form fairness evaluations, i.e., the economic part of fairness. The theory does not address the effects of the interpersonal and procedural part of the evaluation of fairness. Distributive injustice is a necessary but not sufficient condition to motivate deviant behaviour (Alge, 1998). One can conclude that equity theory is maybe too narrow as it only considers the material or economic outcomes people receive.
2.3.2 Procedural Injustice and Deviance
When procedural injustice was introduced, a lot of people started to do research on justice. This form of justice prompts that people do not only judge fairness on the outcomes they receive, but that the procedures to get to these outcomes also matter. Two models of procedural injustice were introduced (Lind & Tyler, 1988). First, process control was introduced as an important influence to get to the desired outcomes: the so-called, self-interest
15 or instrumental model. When individuals can control procedures they maximize the advantages of the outcomes in the long term. Second, the group-value or relational model was introduced (Lind & Tyler, 1988; Lind & Tyler, 1992). This model is about the acceptance and support for change of individuals inside a group. It explains why procedural justice shapes cooperation in organizations, groups, and societies. It states that within groups procedures are important because they shape people’s social identity, and social identity influences attitudes, values, and behaviours. Individuals judge a procedure as unfair if it indicates a negative relationship with authority or low status group membership (Lind & Tuler, 1992).
Research has shown that, apart from distributive injustice, procedural injustice can have a large impact on different behavioural outcomes and attitudes (Sweeney & McFarlin, 1993). Unfair decision-making processes can lead to higher turnover intentions, theft, low organizational commitment and lower performance (Cropanzano & Greenberg, 1997; Folger & Skarlicki, 1998). Procedural injustice, more than distributive injustice, can be seen as a predictor of behaviours in response to judgements about the organization (Materson, Lewis, Goldman, & Taylor, 2000).
In the case of procedural injustice, employees can also see specific people inside the organization as the main source of (in)justice because they establish formal rules and policies to regulate people’s behaviour. Furthermore, people inside the organization are responsible for the rules about the allocation of outcomes. When employees have the feeling that the regulations and rules are unfair, they can have the feeling that it is impossible to get fair outcomes for their performance input. Actions employees take towards these people inside the organization but also towards the organization itself, when they experience procedural injustice, are disloyalty, low organizational commitment and physical property destruction (Aquino, Lewis, & Bradfield, 1999). It looks like these actions are more against the organization than the people who are responsible for the procedures inside the organization, but when focussing on the instrumentality of the emotions, anger at an organization will be less likely to invoke a change on the part of the organization than anger directed at an individual. Therefore in the end, people will mostly direct their anger at an individual inside the organization than at the organization itself.
16 2.3.3 Interactional Injustice and Deviance
Interactional injustice is about how individuals feel that they are treated at the interpersonal level when organizational decisions are made. It would be rational to think that this kind of injustice would mainly be an important predictor of responses to judgements about the supervisor. However, research showed that besides the person-focused outcomes, such as conflict, low performance, and poor attitudes, it also predicts behavioural outcomes such as organizational citizenship behaviour (OCB), withdrawal, and counterproductive behaviour (Colquitt, Conlon, Wesson, Porter, & Ng, 2001). Bies and Moag (1986) showed that impersonal treatments, more than other types of injustice, are likely to provoke intense emotional and behavioural responses. When people have the feeling that they face interpersonal injustice, the strongest emotional responses can be evoked. This can range from anger to moral outrage (Bies, 1987), whereby revenge usually is accompanied by intense anger (Buss, 1961).
Petri and Tanzer (1990) investigated the difference between how individuals who cause a negative incident and the individuals who suffer from the negative incident, evaluate these negative incidents. Their results showed that people find it more important when interactional justice is violated than when procedural or distributive injustice is violated. When people experience interactional injustice, they will often show behaviours to ‘get even’ with the offender (Folger & Cropanzano, 1998). Bensimon (1994) showed that an authoritarian workplace could often contribute to workplace deviance. In her study, employees reported that they showed violent and deviant behaviour because they felt treated in a dehumanized way, rather than the fact that they were demoted or terminated (Bensimon, 1994).
2.3.4 Interaction among Injustice Components and Deviance
Research has not only shown a direct effect between injustice and employees’ behaviour, but also that the different components of injustice interact to affect how individuals react to their perceptions of organizational justice. Of the different forms of injustice, procedural justice has the strongest relationship with employees showing deviant behaviour when the outcome of a procedure is unfair. Distributive injustice has the strongest relationship with employees showing deviant behaviour when the procedure itself is unfair. The interactions between the
17 outcome of a process and the process itself have also been well documented in the literature (Brockner & Wiesenfield, 1996).
The deviant behaviour literature shows that people’s motivation to show deviant behaviour is likely to be the strongest when people perceive multiple unfair events (Tripp & Bies, 1997). Also, emotions such as anger or frustration will be most intense when employees believe that not only the outcome, but also the procedure is unfair and are carried out in a disrespectful way (Folger & Cropanzano, 1998). While multiple facets of injustice can lead to the most intense responses in behaviour, the effect of one type of injustice can be made less intense when people identify and recognize other aspects of justice. When, for example, high perceptions of interactional justice are recognized, these can dissolve the effect of distributive injustice. Folger and Konosku (1989) reported that when employees found their supervisors used fair performance appraisal procedures, they more often had higher levels of pay satisfaction and trust in their supervisor, regardless of the perceived fairness of the pay or amount of pay. Also, when people perceive outcomes as fair, they are less affected by interpersonal injustice (Greenberg, 1993). When employees perceived the outcomes as fair, they did not respond to insensitive and disrespectful personal treatment (Greenberg, 1993).
On the one hand there are many studies, which have examined the direct relationship between organizational injustice and deviant behaviours (Dalal, 2005). But on the other hand there are also studies that include moderators and/or mediators in the justice-deviance relationship. For example different aspects of the Big Five personality dimensions as agreeableness and conscientiousness, as well as organizational citizenship behaviour, are aspects which underlie the justice-deviance relationship (Berry et al., 2007).
In this study the emotions sadness and anger are included as a mediator between the justice-deviance relationship. Anger and sadness are considered as basic discrete emotions that may underlie this relationship. Anger has been shown to occur as a reaction of perceived injustice in early justice theories (Adams, 1965; Homans, 1961). Therefore it seemed plausible to use anger as mediator between perceived injustice and deviant workplace behaviour. Anger as mediator, for example, has been found to partially mediate and as well moderate the relationship between three-way justice interaction and litigation (Goldman, 2003). Sadness, as being an inward focused emotion, was added as mediator in this study to balance the outward
18 focused emotion of anger. In the next section anger and sadness as mediators between the justice-workplace deviance relationship, will be reviewed.
2.4 Anger and Sadness
There are different studies, in different contexts, that have also looked at how sadness and anger changed over time. Sbarra and Emery (2005) demonstrated that anger and sadness changed differently over time. In their study, which is about predicting the onset of emotional recovery following non-marital relationship dissolution, they found that self-reported sadness declined relatively slowly, whereas reports of anger dampened more quickly (Sbarra & Emery, 2005). They also looked at the predictors of the speed of emotional change, hereby demonstrating that self-reported attachment style influences rates of change. Faster rates of sadness decline were evidenced for individual high in attachment security and fearful avoidance. Individuals high in dismissing avoidance showed slower rates of decline for love and anger.
Smith, Cronin & Kessler (2008), looked at anger and sadness as emotional reactions of faculty members to collective pay disadvantage. The results show that anger as well as sadness are distinct emotional responses to a collective disadvantage. For example, anger mediated the relationship between collective disadvantage and willingness to protest whereas group-based sadness mediated the relationship between collective disadvantage and organizational loyalty (Smith et al., 2008).
Weiss and Cropanzano (1996) have studied the role of work events, emotions, and discretionary behaviours in the framework of Affective Events Theory (AET). Organizational injustice can be seen as a factor which causes negative emotions and as a consequence leads to deviant behaviours. Based on AET, in this study the relationship between injustice and deviant behaviours will be investigated. Hereby, anger and sadness are considered as basic discrete emotions that may underlie this relationship. Anger and sadness are intense and short-lived emotions caused by specific precipitous events (Fox & Spector, 1999). Affect research has mostly focused on two broad mood variables (Watson & Clark, 1991, 1994, 1997), but more discrete emotions as sadness and anger are giving little attention.
Negative emotions as anger and sadness can lead to a number of deviant workplace behaviours. Inside an organization employees will now and then behave in ways that
19 ‘threaten’ the organization, other employees, or both. Also, established norms will be violated. Many of these deviant behaviours can be traced to negative emotions (Bedeian, 1995; Ben-Ze’ev, 2000; Lee & Allen, 2002). When employees become sad or angry because, for example, they do not receive the salary they expected, or a better work assignment, this can lead to malicious deviant behaviours (Lee & Allen, 2002; Judge, Scott & Ilies, 2006). Angry or sad employees, for example, will backstab a colleague or distort others’ successes. Evidence suggests that people who feel negative emotions, are more likely than people who do not feel, or less, negative emotions, to engage in workplace deviant behaviour (Lee & Allen, 2002; Judge, Scott & Ilies, 2006).
Anger arises when events are seen as threatening (e.g. unfair procedures) and the individual attributes the responsibility to others (Parrott & Smith, 1993). Homans (1961) states that anger is an emotion people feel when they are under rewarded. Low procedural justice, in its own right, can be a negative outcome and can trigger the perception of unfairness, independent of the outcome favourability (Cropanzano & Ambrose, 2001; Folger & Cropanzano, 2001). When perceptions of norm of justice are violated, this leads to anger due to unfavourable outcomes in respect of economic and socioeconomic terms.
As sadness is a more inward focused emotion (Tangney & Salovey, 1999), this emotion is expressed by for example avoidance, timidity and submissiveness (Madera & Smith, 2009). The relationship between justice perceptions and sadness has been examined by previous studies. These studies showed that sadness occurred when individuals perceived outcomes as unfavourable. Sadness has been seen as one of the emotional reactions towards perceived injustice (Mikula, Scherer & Athenstaedt (1998). Krehbiel and Cropanzano (2000) also showed that when individuals experience unfavourable economic outcomes, the sadness was greater. So when socio-economical outcomes are violated, sadness will occur (Krehbiel & Cropanzano, 2000).
As explained earlier, distributive injustice will probably be more strongly related to anger than sadness. Employees who feel treated unfairly in a ‘distributive’ way, will report higher levels of anger than those who feel treated fairly and do not really experience distributive injustice. Anger in turn, will lead employees who feel unfairly treated, to engage in higher levels of deviant workplace behaviour compared to those who feel distributively fairly
20 treated. Sadness will still have a significant, but smaller, mediating effect of distributive injustice on workplace deviance.
H4: The significant positive Anger mediated indirect effect of distributive injustice on workplace deviance is larger than the significant positive sadness mediated indirect effect of distributive injustice on workplace deviance.
As said above, procedural injustice will probably be more strongly related to sadness, than to anger. Employees who feel treated unfairly in a ‘procedural’ way, will report higher levels of sadness than those who feel treated fairly and do not really experience procedural injustice. Sadness in turn, will lead employees who feel unfairly treated, to engage in higher levels of deviant workplace behaviour compared to those who feel procedurally fairly treated.
H5: The significant positive Sadness mediated indirect effect of procedural injustice on workplace deviance is larger than the significant positive Anger mediated indirect effect of procedural injustice on workplace deviance.
As mentioned earlier, interactional injustice will probably be more strongly related to anger than sadness. Employees who feel treated unfairly in an ‘interactional’ way, will report higher levels of anger than those who feel treated fairly and do not really experience interactional injustice. Anger in turn, will cause employees who feel unfairly treated, to engage in higher levels of deviant workplace behaviour compared to those who feel interactionally treated fairly.
H6: The significant positive Anger mediated indirect effect of interactional injustice on workplace deviance is larger than the significant positive Sadness mediated indirect effect of interactional injustice on workplace deviance.
2.5 Workplace Deviance over Time
As already introduced, workplace deviance in this study is measured over time. The perception and behaviour of people change over time and as emotions change rapidly, it is important to measure these over a period of time. Diary designs are suitable for studying the role of emotions, in this case sadness and anger, in predicting behaviour as they bring the level of analysis closer to the causative conditions leading to the emotions (Bolger et al.,
21 2003). Thereby in this study it is important to have the possibility to do research on an individual level and look at the within person reactions to the three types of injustice.
Depending on the constraints of a situation, people will show certain (deviant) behaviours (Robinson & Bennett, 1997). As in an organization situations can change quickly, it is important to use multiple measures of deviant behaviour over time. If an employee had a difficult or ‘bad’ week, this can have a big impact on his or her behaviour. Therefore it is very important to measure the deviant behaviour of employees over time and to examine the influence of their perception on interactional, distributive and procedural injustice on this deviant behaviour. Longitudinal change in deviant behaviour is investigated and variables that can explain the observed pattern of change are identified.
It is important to distinguish within-person from between-person variability to understand their change and stability over different and fluctuating periods (Mroczek & Spiro, 2003). Regardless the period of time between the different measures, it is important for researchers, as well as practitioners, to understand what it means for one person to differ from another and what it means for a person to vary from him- or herself. In this study, a within-person design, instead of a between within-persons design, allows us to demonstrate that it is not between person differences that are responsible for the observed effects. As a result, this will justify the idea of causal relationships between independent and dependent variables some more.
To analyse data in repeated measures design, Latent Growth Curve Modelling (LGM) has emerged as a versatile tool for studying longitudinal changes (McArdle & Epstein, 1987; Meredith and Tisak, 1990; Duncan, Duncan & Strycker, 2006). When data change over time is seen as an underlying latent process, the LGM identifies a pattern of change over time for each individual in the sample. The characteristics of each ‘pattern’, varying across individuals, are treated as latent variables. By using the LGM model, a large flexibility is provided to represent complex patterns of growth and correlates of change of employees’ deviant behaviours as a function of time (Duncan et al, 2006).
Within this longitudinal analysis, based on the theoretical framework above, the first ‘general’ hypothesis is:
22 H7: higher initial levels of (employees’ perception of) distributive, procedural and interactional injustice will result in higher initial levels of anger and sadness which will lead to a higher initial level of workplace deviance, or stated differently, that higher levels of distributive, procedural and interactional injustice will result in employees being more angry or sad which will lead to a higher level of workplace deviance.
The intercept states what the level is of where the average person starts. In this research this answers to the level of deviant behaviour in the workplace of employees. The slope states what the average rate of change is, so how deviant behaviour changes over the different weeks measured. This, in combination with the theoretical framework above, leads to the following hypotheses:
H8: Lower initial levels of (employees’ perception of) distributive, procedural and interactional injustice will lead to lower levels of employees’ anger or sadness, which will dampen the increases (i.e., reduce the slope) observed in employees’ deviant behaviour longitudinally throughout the research period.
H9: The effect of distributive, procedural and interactional injustice on the workplace deviance initiation (intercept) and increase (slope) will be mediated through employees’ anger and sadness.
23 2.6 Research Model
In short, this study will examine the influence of organizational injustice, perceived by the employees, on organizational and interpersonal deviance over time, mediated by anger and sadness, as one can see in figure 1.1 and figure 1.2.
24 Figure 1.2 Research model
25 Within this chapter the start of the empirical part of this study is represented. The instruments
used for the data collection in this study, and their reliability, will be discussed. Afterwards, a brief description of the statistical approach used to test the expected relationships, will be given. Finally, the most evident characteristics of the sample will be outlined. For the complete questionnaire, please refer to the appendix.
3.1 Measures of variables
The following instruments are used for the data collection in this study.
Workplace Deviance
Employees’ workplace deviance behaviours was measured by using the 19-item workplace deviance scale developed by Bennett and Robinson (1995). 12-items are indicative of organizational deviance (OD- deviant behaviours directly harmful to the organization) and 7-items are indicative of interpersonal deviance (ID- deviant behaviours directly harmful to other individuals within the organization). These scales have internal reliabilities of .81 and .78, respectively and are frequently used for measuring organizational deviance behaviours. O’Neill et al (2011) used this scale and found Alpha Reliabilities of .82 for the 12-item OD scale and .88 for the 7-item ID scale. A 7-point likert type response scale was used to measure the extent to which participants are engaged in workplace deviance behaviours. Items responses ranged from 1=never, 2=once a year, 3=twice a year, 4=several times a year, 5= monthly, 6=weekly, 7=daily.
Anger
Anger was assessed with two adjectives from PANAS-X (Watson and Clark, 1999), Anger and Hostility. Following the instructions of PANAS, participants were asked: “indicate to what extent you experience the following states in the last week” on a five-point scale ranging from 1=not at all to 5=extremely. This scale was used by Rodell and Judge (2009) who reported an Alpha Reliability of .85.
Sadness
Sadness was assessed with adjectives from PANAS-X (Watson and Clark, 1999), Sadness. Following the instructions of PANAS, participants were asked: “indicate to what extent you experienced the following states in the last week”, consisting of five states (sad, blue,
26 downhearted, alone, lonely), on a five-point scale ranging from 1=not at all to 5=extremely. According to research of Watson and Clark (1999), this scale has an Alpha Reliability of .87.
Organizational Justice
The Organizational Justice Scale developed by Niehoff and Moorman (1993) was used to measure organizational justice. This scale consists of three dimensions; Distributive, Procedural, and Interactional Justice. Reported reliabilities (cf. Niehoff and Moorman, 1993) for all the three dimensions are above .90. Distributive justice was measured by using a four item scale assessing the fairness of different work outcomes, work schedule, work load and job responsibilities. Procedural justice was measured using a six items scale and Interactional justice (four items) was used to measure the degree to which employees felt their needs are considered in, and adequate explanations are made for, job decisions. Employees had to indicate how fairly they felt they were treated in the last week. All items were administered using a seven-point response format. Items responses ranged from 1=strongly disagree, 2=disagree, 3=slightly disagree, 4=neither agree nor disagree, 5=slightly agree, 6=agree, 7=strongly agree.
3.2 Sample and Statistical Procedure
A five-wave longitudinal study of the influence of organizational injustice on organizational deviance, influences by employees’ anger and sadness, was initiated on the 14th
of April 2015, and respondents filled in a weekly survey for five times until the 28th of May 2015. As a research design, a (weekly) longitudinal diary study (Gleason et al., 2001; Bolger et al., 2003; Ohly et al., 2010; Yang et al., 2010) was used to test the hypotheses. In this study it is important to have the possibility to do research on an intra-individual level. The data were gathered within the HR-department of the Employee Insurances Implementing Agency UWV. In March they organized the ‘week of the trade union’, so this was probably an extra reason for them to co-operate.
Respondents received a survey every Thursday and had till Wednesday the next week to complete it. The data was collected using the online survey software Qualtrics. The surveys are translated into Dutch as otherwise employees could possibly experience difficulties with
27 the English expressions and difficult words. Initially the surveys were in English, to be able to get a response as large as possible, and possibly for future research to also be able to reach non-Dutch respondents. Before the data collection started a registration survey was sent to all employees of the HR department of the UWV, by an employee of the UWV. To keep the data collection completely anonymous, the registration survey participants were asked to come up with a unique personal code. This code had to exist of 1-the first letter of your mothers’ first name, 2-the last letter of your mothers’ last name, 3-the first letter of your fathers’ first name, 4-the last letter of your fathers’ last name. With this, the responses on the weekly surveys could be matched to one another, but the responses were not traceable to a person.
The data that has been gathered for this study and which this study is reporting on, was collected as part of a larger study on justice, which also includes different personality traits. At the first day of data collection, participants had to complete an online general personality and justice questionnaire, which took about 15 minutes to complete. This survey was about employees’ personality traits, their behaviour at work and how they experience certain events at work. After this, weekly online diary surveys have been sent to the participants via email on every Thursday for four consecutive weeks. As explained before, this was necessary to measure employees’ rapidly changing emotions. On average, these measures took 6.35 minutes to complete (M = 6.35, SD = 2.85)
The registration survey was completed by 215 employees, the general personality and justice questionnaire by 196 employees, the first weekly survey by 141 employees, the second weekly survey by 118 employees, the third weekly survey by 101 employees and the last weekly survey had 98 respondents. Only the data from respondents who took part in the whole process were used in the final analysis of this study. Some respondents did not fill in all questions of the last survey, so in the end 89 respondents could be used for the analysis. Of these respondents, 55% were male and the average age was 38 years old (SD= 22).
Diary designs are suitable for studying the role of emotions, in this case sadness and anger, in predicting behaviour as they bring the level of analysis closer to the causative conditions leading to the emotions (Bolger et al., 2003). Latent Growth Structural Equation Modelling using IBM SPSS AMOS version 22 was implemented to test the hypotheses as stated in the previous section. One of the greatest advantages of LGM probably lies in its ability to examine changes of inter-individual differences over time, as well as incorporate
28 time-varying and time-invariant covariates into the model. Curran and Willoughby (2003) summarize it as “an intersection between variable-centred and person-centred analysis”. Conceptually, the growth-curve method creates a regression-type line for each respondent’s deviant behaviour over time (see figure 2). Two latent factors were estimated, one representing the respondent’s baseline deviant behaviour (i.e., the intercept) and the other representing changes in deviant behaviour over time (i.e., the slope). To represent employees’ baseline levels of deviant behaviour, the intercept factor was created with a fixed loading of 1.0 deviant behaviour at each measurement wave. To represent employees’ changing deviant behaviours over time, the slope factor was estimated with a fixed 1 loading to deviant behaviour at measure 1, a fixed 2 loading to deviant behaviour a measure 2, etc.
Of distributive, procedural and interactional injustice, the results of the first measured moment were used. This also applies for the measures of anger and sadness, but controlling for measure moment 5. The choice to use the first measurement, controlling for the fifth measurement, was based on the fact that for behaviour to change, time is needed as this will probably be based on for example changed circumstances, which can take a while. For these variables it was impossible to use the slope and intercept of the different measures over time, as this would implicate the model too much, and with the program AMOS it would be impossible to create such a complex model which would work in combination with the data gathered for this study. On the far right-hand side of the model are the manifest deviant behaviour index scores corresponding to the five data collection waves.
4. Results
In this chapter, the conducted analyses and associated results that test the hypotheses of this study are described. The results pertaining to each of the hypotheses are showed.
4.1 Analytical strategy
It will now be made some clearer why for this study a latent growth modelling analysis is used. This study started with the idea of using the measures of deviant behaviour as well as anger, sadness and injustice over time, in the model. Unfortunately the model in combination with the data came out to be to complex. Therefore the choice has been made to only look at the slope and intercept of the outcome variable, deviant behaviour, over time. Still the latent growth curve modelling analysis is a right method to test all of these hypotheses.
29 Measures of the three forms of injustice from wave 1 are used as injustice perceptions of the employees over time did not fluctuate so much. To do justice to the fleetingness of emotions, the relations between the three forms of injustice and employees’ anger and sadness were also tested for wave 5. Ideally the five measures of anger and sadness over time would have been used in the model, but as explained earlier this was not feasible. Therefore the hypotheses are also tested for anger and sadness wave 5. As deviant behaviour is the dependent variable in this study, and seemed to be the most interesting variable to look at over time, of this variable the intercept and slope are used.
Before definitely choosing the model used in this study, different models were also looked at and tried out. A model with the slope and intercept of anger and sadness as well as the intercept and slope of deviance was too complex for the data to run. Also a model with justice measure 3 was used but this made the model fit worse with the data than when justice measure 1 was used. Models with other time measures of anger and sadness were also tried but as explained above, measure 1 and 5 of these variables were chosen.
4.2 Results linked to the hypotheses
First, a check of frequencies was done to check for errors in the data. Missing values were left out by ‘excluding cases listwise’ to make sure only cases that had no missing data in any variable were analysed. Twelve cases dropped out as a result of this. Items that are phrased so that an agreement with the item represents a low level of the construct being measured were recoded into different variables.
As one can see in table 2, the mean of Anger measure 1 is 2.51 (SD= .883), and for measure 5 the mean is 1.82 (SD= .706). For Sadness measure 1 the mean is 2.80 (SD= .851), and for measure 5 this is 2.06 (SD= .883). This means that on average, employees felt more angry and sad in week 1 than in week 5.
30 Table 2: Means, Standard Deviations, Correlations
31 mean is 359 (SD= .740). The mean for procedural injustice measure 1 is 3.61 (SD= .944), and
for measure 5 this is 3.51 (SD= .978). For interactional injustice the mean of measure 1 is 3.48 (SD= .843), and for measure 5 this is 3.40 (SD= .812). This means that on average employees experienced more distributive injustice in week 5 of the study than in the first week. Employees on average experienced less procedural and interactional injustice in week 5 than in week 1.
Means on the workplace deviance index are displayed in Figure 1 for each wave of data collected; the means show that workplace deviance sloped downwards over time. This also counts for the first and fifth measures of Anger and Sadness.
Table 1. Means of the five different measures of Workplace Deviance over time
Unfortunately, the model does not fit the data well (Chi-square= 7.317, p=.038) with major fit indices, normed (NFI), non-normed (TLI), comparative (CFI), and RMSEA of -.014, -.171, .000, and .242 respectively. As the cut-off criteria of NFI is ≥ .95 (Hu & Bentler, 1999), a score of -.014 is far from ideal. TLI should be ≥ .90, so a score of -.171 does not mean a good fit at all. The same counts for a CFI score of 0.000, which should be ≥ .95, and a RMSEA score of .242 which should be ≤ .08. The ‘modification indices’ were examined to find a potential way to improve the model empirically, but unfortunately there were no changes, which made sense theoretically, that could be made to improve the model fit.
1.494 1.321 1.299 1.237 1.184 1.1 1.2 1.3 1.4 1.5 1.6
Measure 1 Measure 2 Measure 3 Measure 4 Measure 5
Organizational Deviance Interpersonal Deviance Workplace Deviance
32 The predictors in the model were from Wave 1, controlling for anger and sadness Wave 5, so that the predictive relations would be prospective. In this prospective study we look at the development of employees’ their emotions, during the study period. As explained earlier it was not possible for all variables, within this model, to take into account the five measures over time and look at the intercept and slope. A prospective study involves taking a cohort and watching them over a long period. Therefore the two measures with the biggest time span between it, the first and fifth measure, are used in the model. Prospective studies in this, have fewer potential sources of bias than retrospective studies.
Figure 2 shows that distributive injustice is significantly and positively related to employees’ anger in Wave 1 (.260, p < .001). It also shows that distributive injustice is positively related to employees’ sadness in Wave 1 (.370, p < .001). With regard to the hypothesized relations in Hypothesis 1, higher levels of distributive injustice do lead to higher levels of anger and higher levels of sadness. However, as the hypothesis stated, the positive relationship between distributive injustice and employees’ anger is not stronger than the positive relationship between distributive injustice and sadness, but vice versa. Although, the first hypothesis is partially supported since evidence is established for significant relationships between distributive injustice and both anger and sadness.
33
34 1 (.173, p <.05). There is also a positive relation between procedural injustice and employees’
sadness in Wave 1 (.128, p < .05) as stated in hypothesis 2. But differently than the hypothesis states, the relationship between procedural injustice and employees’ sadness is not stronger than the relationship between procedural injustice and anger, but the other way around. So the second hypothesis is partially supported since evidence is established for significant relationships between procedural injustice and both anger and sadness.
Between interactional injustice and anger a positive relationship is also found (.161, p <.05) as well for interactional injustice and sadness (.163, p <.05). The positive relationship between interactional injustice and employees’ anger and sadness is almost equally strong. Therefore hypothesis 3, which stated that the positive relationship between interactional injustice and anger is stronger than the relationship between interactional injustice and sadness is not supported.
As explained the hypotheses are also tested for anger and sadness wave 5. When looking at these results, not all of these relations are significant anymore. The relationship between interactional injustice in wave 1 and anger in wave 5 is .073 (p < .001), and for sadness wave 5 this is .087 (p < .001). For procedural injustice wave 1 and sadness wave 5 this is .086 (p < .001) and for anger wave 5 this is .025 (p= .220). The relation between distributive injustice wave 1 and anger wave 5 is .056 (p < .05) and for sadness wave 5 this is .042 (p= .335). So the relationship between anger wave 5 and procedural injustice is not significant anymore as well as the relationship between distributive injustice in wave 1 and sadness in wave 5.
Controlling for anger and sadness wave 5, as explained above, hypothesis 1 would now be supported for the relationship between distributive injustice and anger and sadness wave 5. The positive relationship between distributive injustice and employees’ anger is stronger than the positive relationship between distributive injustice and sadness wave 5, but distributive injustice and sadness are not significantly related anymore.
For hypothesis 2 this is also the case. The positive relationship between procedural injustice and employees’ anger wave 5 is stronger than the positive relationship between procedural injustice and sadness wave 5, but unfortunately the relationship between anger and procedural injustice is not significant anymore.
35 For the relationship between interactional injustice and anger and sadness wave 5, hypothesis 3 is still not supported. Interactional injustice now has a stronger positive relationship with sadness than anger, instead of the other way around.
For there to be a true mediation, or explanation of the mechanism by which the effect occurs, there should be a significant path between organizational injustice and deviant workplace behaviour with the potential mediators, anger and sadness, not in the model. Therefore this additional analysis was conducted.
The coefficients of interactional (-.011, p<0.01), distributive (.028, p<0.001) and procedural (.008, p< 0.05) injustice are all significantly related to workplace deviance slope with anger and sadness removed from the model; thus, organizational injustice is related to the increase in organizational deviance (slope) apart from employees’ anger and sadness. Only the coefficients of distributive injustice (-.154, p<0.001) and interactional injustice (.040, p< .01) are significantly related to the workplace deviance intercept. The coefficient of procedural injustice to the deviance intercept was not significant (-.011, p= .476). However, because in this study the focus is on the fluctuation of workplace deviance it is not extremely important. As one would expect that interactional injustice leads to deviant behaviour it is striking to conclude that interactional injustice reduces the slope of deviant behaviour.
Now, that it has been concluded that the three forms of injustice are significantly related to the slope of deviant behaviour, the mediating effect of anger and sadness can be examined. First it is concluded that higher levels of anger as well as sadness in wave 1 and 5, significantly lead to deviant behaviours. Both mediating effects of anger (.009, p= .530) and sadness (.013, p=.444) of wave 1, between the relationship of the three forms of organizational injustice on the deviant workplace behaviour slope, are not significant. It is even more remarkable to see that in wave 5 sadness and anger, although not significant, have a negative mediating effect between organizational injustice and workplace deviance. This would mean that when employees are more angry or sad, these feelings reduce their utterance of deviant behaviour.
