Factors contributing to the perpetration of workplace incivility: the importance of organizational aspects and experiencing incivility from others.

In recent years a growing amount of research has been conducted in the area of workplace incivility. Whereas many studies have focused on the victims and the consequences of incivility, little attention has been paid to the perpetrators and antecedents of workplace incivility. This study aims to identify possible antecedents of workplace incivility, by investigating organizational aspects as well as the possibility that being the target of incivility from co-workers and supervisors could induce incivility. A total of 512 employees (378 women and 133 men) in the school sector in a Swedish municipality completed an online questionnaire. Overall, the results of structural equation modelling analyses showed that organizational variables were related to the perpetration of incivility. A direct relationship was found between being uncivil and organizational change, job insecurity, low social support from co-workers and high job demands. However, the strongest relationship was found between experienced incivility from co-workers and instigated incivility. This could be reflecting a climate or culture of incivility in the organization, and carry implications for future practice in interventions against workplace incivility. The results indicate the importance of focusing on the perspective of the instigator to gain knowledge about the process of workplace incivility.

Introduction

In recent years, many studies have been conducted on various types of workplace violence and its negative effects on both individuals and organizations (Hershcovis et al., 2007). These studies have concerned different dimensions of aggressive behaviours ranging from physical violence and harassment to milder forms such as psychological aggression, all of which can be seen as counterproductive work behaviour (Pearson, Andersson, & Porath, 2005).

One of these dimensions is workplace incivility, a subtle form of interpersonal negative behaviour characterized by rudeness and disrespect (Andersson & Pearson, 1999; Cortina, Magley, Williams, & Langhout, 2001). These behaviours constitute the basis of norms for consequent behaviour, which affects the entire organization through a generally rude climate. Incivility is a phenomenon that has attracted attention in international research on contemporary working life (Cortina, 2008; Lim & Lee, 2011). Workplace incivility has been defined as “ … low-intensity deviant behavior with ambiguous intent to harm the target, in violation of workplace norms for mutual respect. Uncivil behaviors are characteristically rude and discourteous, displaying a lack of regard for others” (Andersson & Pearson, 1999, p. 457). It is characteristic of uncivil behaviour that it has a low intensity and that the intention to harm is unclear (Pearson, Andersson, & Wegner, 2001). This phenomenon is closely related to and partially overlaps with other types of workplace mistreatment such as bullying, rude treatment, abuse, harassment, antisocial behaviour and social undermining, but it differs as it is lower in intensity (Hershcovis, 2011; Martin & Hine, 2005). Despite the lower intensity, incivility has been considered as a precursor to more serious aggression and negative consequences. In addition, workplace incivility can be seen as part of the climate or culture in the workplace rather than as an individual phenomenon (Leiter, 2013), and its subtle form makes it more difficult to detect (Lim & Lee, 2011). Pearson et al. (2005) give several examples of how workplace incivility can appear. It can be manifested by taking credit for others’ efforts, spreading rumours about colleagues, leaving the office untidy, not fixing errors, sending nasty emails to colleagues or not praising subordinates.

Discourteous behaviour is occasionally seen as unproblematic and harmless, but many researchers suggest that as the phenomenon of workplace incivility is common in today’s organizations the adverse effects and costs are large (Pearson & Porath, 2009). Evidence shows that employees who perceive incivility are at greater risk of mental health problems (Cortina et al., 2001; Lim & Lee, 2011; Torkelson, 2011). It has also been found that incivility is related to decreased job satisfaction, decreased willingness to remain in the organization, impaired performance, lower productivity (Cortina et al., 2001), lower job satisfaction and increased turnover intentions (Laschinger, Leiter, Day, & Gilin, 2009), lost commitment to work, lack of loyalty to the organization (Pearson et al., 2005), decreased satisfaction with managers and colleagues, a sense of injustice and conflict between work and family (Lim & Lee, 2011). In addition, uncivil behaviour may spread as a norm which affects the entire organization. In that way a culture of uncivil behaviour may be developed and the long-term consequences of such a culture may be extensive (Pearson et al., 2005).

Several authors (Estes & Wang, 2008; Pearson et al., 2005) have reported that workplace incivility is a phenomenon that has increased in working life in recent years, and research has demonstrated that incivility at work is highly prevalent (Cortina et al., 2001; Lim, Cortina, & Magley, 2008; Pearson, Andersson, & Porath, 2000). In an American study, 54% of participants reported that they had instigated incivility towards others in the organization in the past year, 12% had instigated it several times and 3% had done so daily (Reio & Ghosh, 2009). In another American study, Pearson et al. (2005) found that 25% witnessed workplace incivility daily and 50% had themselves been victims at least once a week in the past year.

Although many studies have been conducted focusing on victims and the consequences of incivility, several authors (Jex, Burnfield Geimer, Clark, Guidroz, & Yugo, 2010; Liu, Chi, Friedman, & Tsai, 2009; Meier & Semmer, 2013) argue that relatively little attention has been paid to study the instigators and the triggers of workplace incivility. Understanding workplace incivility from an instigator perspective may be the key to reducing and preventing incivility at work according to Jex et al. (2010). Along the same lines, Meier and Semmer (2013) highlight the importance of such studies in order to understand and gain knowledge about the antecedents of incivility as a basis for finding remedies for the phenomenon. The present study aims to address the antecedents of workplace incivility, by investigating organizational factors as well as factors related to being targeted by incivility that could contribute to instigating workplace incivility.

Incivility as a social process

Even if workplace incivility is low in intensity there is a risk that it will grow into more serious forms of aggressive behaviour. Andersson and Pearson (1999) describe workplace incivility through a social exchange perspective, as social interactions in the form of a spiral of reciprocal interpersonal conflicts. The interactions can easily escalate from initially low intensity to more distinct forms where the intention to harm is more clear. Incivility affects the entire organization (Pearson et al., 2005) and can cascade downwards through the organization as third parties to the uncivil interaction observe and subsequently adopt similar behaviour, either directly or by receiving second-hand information about an uncivilized encounter in the workplace (Pearson et al., 2000). In that way a supervisor that acts uncivil can become a role model for behaviour in the workplace (Estes & Wang, 2008). As incivility affects both those exposed and those who observe the process, it can influence the entire organization (Lim et al., 2008) and the negative process can result in a climate of incivility (Andersson & Pearson, 1999; Pearson et al., 2000).

Various studies have shown that incivility is in several ways related to the power position of the instigator. First, it is common for the instigator to be found higher up in the organizational hierarchy (Estes & Wang, 2008). Based on the social power theory (French & Raven, 1959), Cortina et al. (2001) argue that incivility can serve as a means of exercising power. In an organizational context, social power theory posits that employees of lower social status, such as those lower in the organizational hierarchy or those who are part of low-status groups in the workplace, may be more susceptible to incivility from higher status employees. In this way, the incivility process often starts from the top of the organization when high-status employees enact uncivil behaviours towards lower status employees. Pearson and Porath (2009) found that about 60% of all reported workplace incivility in their study was initiated by a person with a higher organizational rank and was directed to a person of lower rank. Second, incivility may take different forms that are related to the instigator’s power position in the organization. Pearson and Porath (2009) found that incivility that starts from the bottom of the organizational hierarchy and directs upwards is exerted in other ways than incivility exerted in the opposite direction, such as subtle forms of sabotage. A third way in which incivility also may be related to power is that the target’s perception of bad behaviour may be linked to the power position of the instigator. A study by Cortina and Magley (2009) revealed that employees experienced rude treatment in a more negative way if it was initiated by someone who had a higher position. In the light of the relationship between incivility and power position, it is relevant to investigate incivility from co-workers and incivility from supervisors separately.

Organizational antecedents

Pearson et al. (2005) suggest organizational pressures to be among the potential causes of workplace incivility. They mention organizational pressures such as changes at work, resulting in uncertainty about downsizing, restructuring and mergers. They also mention high demands, new technologies for communication, poor leadership, compressed time and deadlines, work and information overload as factors that have been potential causes of workplace incivility. The relationship between organizational change and workplace aggression has been partially supported, that is, organizational change has been related to a higher incidence of workplace aggression (Baron & Neuman, 1996).

In the research on bullying, relationships have been found between workplace changes and bullying. As Hershcovis (2011) showed, there is a certain degree of overlap between workplace bullying and workplace incivility; therefore, the research on bullying is relevant when considering antecedents of workplace incivility. Major organizational changes have been associated with exposure to bullying (Hoel & Cooper, 2000), and especially change of management has been proved to be particularly important. In a comprehensive longitudinal study by Oxenstierna, Elofsson, Gjerde, Magnusson Hanson, and Theorell (2012), various predictors of workplace bullying were explored. Organizational change was found to be a significant component in the onset of workplace bullying over a two-year period (Oxenstierna et al., 2012). In a study by Hauge, Skogstad, and Einarsen (2009), a stressful working environment was identified as an important contributor to the explanation of workplace bullying. By means of a longitudinal study design, Finne, Knardahl, and Lau (2011) found that bullying at work caused mental distress and that mental distress led to bullying.

De Cuyper, Baillien, and De Witte (2009) found that job insecurity was associated with both targets and perpetrators’ reports of bullying at work. A study by Baillien, De Cuyper, and De Witte (2011) revealed that high-strain jobs as defined by the Demand Control model proposed by Karasek (1979), that is, in jobs that lack autonomy in combination with high workload, the worker is especially sensitive to both being a target and a perpetrator of workplace bullying. The findings are supported by an additional study, by Notelaers, Baillien, De Witte, Einarsen, and Vermunt (2012), showing high job demands and low control to be associated with being targeted by workplace bullying. Karasek’s model also includes social support as a relevant factor for high-strain jobs (Karasek & Theorell, 1990), and associations have been found between reporting low social support and workplace bullying (Hansen et al., 2006). Moreover, Agervold (2009) found a relationship between bullying and fear of organizational change as well as other organizational factors such as work pressure, performance demands, role conflicts and poor social climate.

A longitudinal study by Blau and Andersson (2005) revealed that perceived exhaustion was positively related and that distributive justice and job satisfaction were negatively related to perpetrating workplace incivility. Perceived feelings of injustice were shown to potentially develop from uncivil interactions (Leiter, 2013), and this could contribute to incivility, as the victim is prone to retaliate for the unjust behaviour. Leiter (2013) argued that organizational change could increase the risk of a negative spiral occurring in an organization. Fevre, Lewis, Robinson, and Jones (2012) suggested that explanations for workplace violence are likely found in the workplace, rather than among the individuals. Additionally, they showed that experiencing less control and having a too intense work pace were associated with incivility.

Other suggested causes for workplace incivility have been lack of job autonomy (Jex et al., 2010), job insecurity through downsizing, greater stress, overworked employees and an increased diversity leading to interpersonal misunderstandings (Blau & Andersson, 2005). Based on the literature about workplace aggression and incivility and the connection made between workplace bullying and organizational factors, the following hypothesis is investigated in the present study:

Hypothesis 1: Workplace incivility will be more frequently reported when employees have recently been exposed to (a) organizational change, (b) perceived job insecurity, (c) low social support from co-workers, (d) low social support from supervisors or (e) job demands and (f) low control.

Being a victim as a possible antecedent to perpetrating incivility

In addition to organizational antecedents, the workplace aggression literature has suggested that being a victim of aggression is a possible antecedent to perpetrating workplace aggression. Hauge et al. (2009) found that being a target of bullying was the strongest predictor of bullying others. A meta-analysis by Hershcovis and colleagues (2012) showed that workplace deviance is a common response to workplace aggression. Specifically, victims have been shown to be more prone to retaliate against high power perpetrators if they have low task interdependence (Hershcovis, Reich, Parker, & Bozeman, 2012). Both being targeted by workplace aggression and a higher mean level of aggression in the work group proved to be significant predictors of engaging in interpersonal aggression in the workplace (Glomb & Liao, 2003). Bowling and Beehr (2006) proposed a reciprocity-attribution model of workplace harassment. They argue that individuals reciprocate actions towards the perpetrator once victimized and at the same time blame the organization for allowing the behaviour to exist. In relation to workplace incivility, Jex et al. (2010) illustrated how incivility can escalate in the form of a spiral, as victims subsequently act in an uncivil manner in order to retaliate for being treated uncivilly. Furthermore, Porath and Pearson (2012) found that being targeted by incivility was associated with anger and fear. Anger and fear were in turn related to increased indirect aggression towards the instigator. Additionally, Gallus, Bunk, Matthews, Barnes-Farrell, and Magley (2014) found that experienced incivility predicted perpetrated incivility. They also found that 70% of participants labelled themselves as both targets and perpetrators. It was argued that being targeted by incivility triggered a sense of retaliation in the victim, as targets of these behaviours proceed to be uncivil as a way to get back at their antagonist (Greenberg, 2010). With this theoretical background, being a target of workplace aggression also appears to be a possible antecedent of perpetrating acts of aggression at work. A further hypothesis is proposed:

Hypothesis 2: Experienced incivility from co-workers or supervisors will contribute uniquely to perpetrated incivility beyond the relationships in Hypothesis 1.

Sample

The teaching profession has been pointed out to be a high risk sector for workplace bullying from colleagues, possibly due to the high focus on interpersonal relationships within the sector (Hubert & van Veldhoven, 2001). In a study conducted in Turkey, 50% of primary school teachers had been subjected to bullying at work during the past six months (Cemaloglu, 2007). Public teachers in the United States report high levels of stressors and violent acts relating to the work environment (Fox & Stallworth, 2010). Workplace mistreatment, including behaviours such as abusive supervision, undermining and incivility, has also been related to teacher burnout in a Romanian study (Sulea, Filipescu, Horga, Orţan, & Fischmann, 2012). Of secondary school teachers in Sweden, 22% have been subjected to violence, threats or harassment, while 13% of college teachers have experienced the same behaviour. In addition to this, 45% of teachers in Sweden have reported that they are stressed most of the time, or all of the time (The Swedish National Agency for Education, 2009). Galand, Lecocq, and Philippot (2007) also note that despite the widespread media coverage of workplace violence as a risk for teachers, minor repetitive negative behaviours were much more frequent and had negative effects for teachers. In the light of this, it stands to reason that the educational sector could be particularly subject to milder forms of workplace aggression, such as incivility.

Method

Participants

An online survey was sent to 1733 individuals employed in the school sector in a Swedish municipality. Responses from 512 employees were received (378 women and 133 men), resulting in a response rate of 29.5%. The mean age was 48 years (SD = 10.4), they had been working at the current workplace on average 9.9 years (SD = 8.5), 83 (16.2%) of the participants had a managerial or executive position, 471 (92%) were born in Sweden and 40 (7.8%) were born in another country. The majority, 471 (92%), were permanently employed while 40 (7.8%) had a different form of employment. Of the 512 employees who answered the questionnaire, 25 (4.9%) were school principals, 194 (37.9%) secondary school teachers, 116 (22.7%) primary school teachers, 23 (4.5%) youth leaders, 24 (4.7%) preschool teachers, 26 (5.1%) administrators, 30 (5.9%) employees from the school health and 73 (14.2%) participants belonged to any other category of staff.

Measures

Demographic variables

Demographic information was collected about gender, age, managerial position (if any), whether participants had been born in Sweden or not, if they held temporary employment and length of employment. The demographic variables were not included in the analyses as this was beyond the scope of the study.

Organizational aspects

The Copenhagen Psychosocial Questionnaire (COPSOQ II) (Pejtersen, Kristensen, Borg, & Bjorner, 2010) in a Swedish version (National Research Centre for the Working Environment, 2014) was used for subscales measuring psychosocial factors of the working environment. The subscales were job insecurity, social support from co-workers, social support from supervisors, job demands and control. In addition to the subscales from COPSOQ, a single-item question regarding exposure to organizational change was included.

Job insecurity. Job insecurity was measured by a four-item subscale. A sample item is “Are you worried about becoming unemployed?” and “Are you worried about being transferred to another job against your will?” Response alternatives ranged from 0 (to a very small extent) to 4 (to a very large extent). Cronbach’s alpha was .78.

Social support from co-workers. To measure social support from co-workers, a three-item subscale was used. Sample items are “How often do you get help and support from your colleagues?” and “How often are your colleagues willing to listen to your problems at work?” Response alternatives ranged from 0 (never/hardly ever) to 4 (always). Cronbach’s alpha was .78.

Social support from supervisors. To measure social support from supervisors a three-item subscale was used. A sample item is “How often is your nearest superior willing to listen to your problems at work?” Response alternatives ranged from 0 (never/hardly ever) to 4 (always). Cronbach’s alpha was .86.

Job demands. For job demands, a four-item subscale was used. A sample item is “Is your workload unevenly distributed so it piles up?” Response alternatives ranged from 0 (never/hardly ever) to 4 (always). Cronbach’s alpha was .82.

Control. Control was measured through a four-item subscale. A sample item is “Do you have a large degree of influence concerning your work?”. Response alternatives ranged from 0 (never/hardly ever) to 4 (always). Cronbach’s alpha was .76.

Exposure to organizational change. A single-item question was used to measure whether the participants had been exposed to organizational change during the last year. This item was “Has any reorganizations occurred at your workplace during the last two years?” Response alternatives ranged from 0 (no) to 1 (yes).

Workplace incivility

Experienced workplace incivility from both supervisor and co-worker, as well as having been a perpetrator of incivility were measured.

Experienced workplace incivility (from co-workers and supervisors). Experienced incivility from co-workers and incivility from supervisors were measured by the seven-item Workplace Incivility Scale (WIS; Cortina et al., 2001), which was translated into Swedish (Schad, Torkelson, Bäckström, & Karlson, 2014). The scale included both direct and indirect forms of workplace aggression. The perceptions of co-workers’ and supervisors’ incivility were rated separately as advocated by Smith, Andrusyszyn, and Laschinger (2010). The scale assessed the frequency of experienced incivility in the last year, which is a shorter time frame than originally used by Cortina et al. (2001). The response alternatives ranged from 0 (never) to 4 (most of the time). A sample item is “During the past year while employed in the current organization have you been in a situation where any of your co-workers: Have made demeaning or derogatory remarks about you?”. When measuring experienced supervisor incivility co-workers was replaced by supervisors. Cronbach’s alpha for perceived incivility from co-workers was .91 and for supervisors .92.

Perpetrated workplace incivility. In accordance with Blau and Andersson (2005), a modification of the WIS was made to measure having been a perpetrator of workplace incivility. The scale’s seven items ranged from 0 (never) to 4 (most of the time), whereby participants were asked to rate their own behaviour rather than the behaviour of others. A sample item is “During the past year while employed in the current organization, how often have you: Made demeaning or derogatory remarks about others?” Cronbach’s alpha was .79.

Procedure and ethical considerations

The study was approved by the Regional Board of Ethical Review (EPN) (www.epn.se/sv/lund) in Lund, Sweden (dnr 2012/626). Participants were fully informed of the confidential nature of the study, and that any participation was voluntary. As the online survey was administered by the organization through e-mail distribution, no identifiable information beyond the demographic inquiries of the questionnaire was collected in the survey. Data were stored and analysed securely.

Strategy of analysis

The hypotheses were tested with structural equation models using latent variables. One of the chief advantages of using this method is that coefficients can be measured without error. It is easy to estimate how well the tested models fit the empirical data and get estimation of indirect relations in the model. The relationships between organizational aspects in the form of (a) organizational change, perceived (b) job insecurity, (c) social support from co-workers, (d) social support from supervisors, (e) job demands and (f) control and perpetrating incivility. Unique contributions beyond these relationships between experienced incivility from co-workers and supervisors and perpetrating incivility were tested by creating a number of structural models.

Before testing the models we tested the measurement models of the latent variables. The workplace incivility variables (having experienced co-worker’s and supervisor’s incivility and having perpetrated incivility) and the organizational aspects (job insecurity, social support from co-workers and supervisors, job demands and control) were measured by items from their respective scales as observed variables. One organizational variable, organizational change, was measured with only a single item; this variable was included in models as an observed variable.

The first structural model (Model 1) consisted only of the organizational variables as independent variables and perpetrated incivility as the dependent variable. After testing this model, as well as a second model (Model 2) where two variables (social support from supervisors and control) that did not contribute much were excluded, we created a third model adding the experienced incivility latent variables (Model 3). The last addition can be compared with a hierarchical or sequential regression with the difference that we made use of latent variables. All independent variables were defined to be correlated in the three models, as is usually the case in standard regression models. More exploratory, using the last model, we investigated indirect relations by estimating the paths from the independent to the dependent variables when these relations included the incivility variables as possible mediators.

MPLUS was used to estimate the models. All variables were defined as categorical and therefore the weighted least square estimator, with adjusted mean and variance adjusted χ 2 values (WLSMV), was used. All observed variables were ordinal; therefore, they were analysed as ordered categories and in that case estimations perform better using WLSMV. The fit indices used were comparative fit index (CFI > 0.95 representing excellent fit) and 
root mean square error of approximation (RMSEA < .05 representing excellent fit, and < .08 representing acceptable fit); in the measurement models, we chiefly relied on the CFI since the RMSEA was very unstable and CFI was very close to 1.0, representing an almost perfect model fit.

We tested the measurement models of all the latent variables and found that they all individually resulted in a close fit to the data (see Table 1). Next we tested the two latent variables measuring experienced incivility from co-workers and from superiors. The two models were not nested; therefore, it was not possible to make a formal test deciding if there were one or two distinct latent variables. However, the two-factor model was supported since a measurement model with one factor was considerably worse (χ 2 = 815.9, CFI = 0.93 compared to χ 2 = 243.0, CFI = 0.98). If all items had been included in the latent variables the number of degrees of freedom in our structural models would have been very large. Therefore, in the next step, the number of observed variables, estimating the incivility variables, was reduced from 7 to 3 by making random parcels of the variables consisting of two observed variables each. The item with the least variability was excluded (“Ignored or excluded you from professional camaraderie?”, item number 5) to make the reliability of the parcels more even. This reduced the number of covariances in the baseline model from 780 to 378. During our estimations of the models, we compared other possible combinations of items to parcels to guarantee that the results did not depend on item parcelling.

Table 1.

Standardized loadings and fit of the measurement models.

Measure	Item 1	Item 2	Item 3	Item 4	Item 5	Item 6	Item 7	CFI
JI	.890	.800	.786	.805				1.000
SSC	.925	.854	.578					1.000
SSS	.939	.842	.774					1.000
JD	.643	.817	.989	.622				0.999
C	.695	.629	.794	.714				0.991
EIC	.938	.818	.937	.899	.780	.730	.644	0.991
EIS	.946	.889	.984	.908	.885	.859	.851	0.997
PI	.815	.661	.941	.758	.717	.579	.746	0.967

Notes: JI = job insecurity, SSC = social support from co-workers, SSS = social support from supervisors, JD = job demands, C = control, EIC = experienced incivility from co-worker, EIS = experienced incivility from supervisor, PI = perpetrated incivility.

Results

Table 2 shows the correlations between the latent variables used in the model. The highest correlations with instigated incivility were found for experienced incivility from co-workers and supervisors, but all variables were significantly related to instigated incivility.

Table 2.

Correlations among the latent variables used in the model.

Variables	1	2	3	4	5	6	7	8
1. Organizational change
2. Job insecurity	.062
3. Soc. support co-workers	–.252	–.037
4. Soc. support supervisors	–.212	–.092	.537
5. Job demands	–.032	.082	–.147	–.344
6. Control	–.321	.019	.424	.641	–.338
7. Experienced co-worker incivility	.295	.184	–.469	–.348	.155	–.359
8. Experienced supervisor incivility	.265	.190	–.267	–.662	.276	–.480	.604
9. Perpetrated incivility	.191	.217	–.226	–.201	.161	–.186	.625	.514

Notes: Organizational change is an observed variable. Correlations larger than .110 were significant (p ≤ .001).

The first model (Model 1 in Table 3) tested Hypothesis 1 (a–f) that organizational factors will contribute to instigated incivility. In this model all independent variables were free to correlate. The model had acceptable fit but two of the organizational variables, control and support from supervisors, had very weak and insignificant relations with instigated incivility. A modified model (Model 2 in Table 3), where the paths to instigated incivility from these variables were fixed to zero, did not change the fit of the model significantly (the variables were not excluded from the model). In the modified model organizational change was found to have the strongest standardized coefficient; being exposed to organizational change was connected to instigated incivility. Another comparatively strong path was the one from low social support from co-workers to instigated incivility. Somewhat weaker, but still significant relations were found for job insecurity and job demands. Together these four variables explained 13.7% of the variance in the latent instigated incivility variable.

Table 3.

Standardized paths and fit indices from SEM models; having perpetrated incivility as dependent variable.

Variables	Model 1	Model 2	Model 3
Standardized paths
Organizational change	.196***	.192***	.096
Job insecurity	.135*	.152*	.000
Social support from co-workers	–.145*	–.175**	.131
Social support from supervisors	–.013
Job demands	.109	.133*	.032
Control	–.013
Experienced incivility from co-workers			.601***
Experienced incivility from supervisors			.125
R2	.130***	.137***	.444***
Fit indices
χ2	634.2	609.7	867.5
df	189	191	317
p	<.001	< .001	< .001
RMSEA	0.068	0.066	0.058
CFI	0.949	0.952	0.959

Notes: RMSEA = root mean square error of approximation; CFI = comparative fit index.

*p < .05; **p < .01; ***p < .001.

The next model (Model 3 in Table 3) tested the hypothesis that the two latent variables measuring experienced incivility, from co-workers and supervisors, respectively, contributed unique variance from instigated incivility. All independent variables were free to correlate in this model. The addition of the two incivility latent variables resulted in a model with good fit. The path from experienced incivility from co-workers was significantly related to instigated incivility. This path was very strong (see Model 3 in Table 3). Two variables were close to significant, the path from social support from co-workers (β = .131, p = .051) and the path from organizational change (β = .096, p = .055). As regards the path from social support from co-workers, the same variable was negatively related to instigated incivility in previous models, but when experienced incivility from co-workers was added, the coefficient changed sign, suggesting a suppressed relationship. The path from experienced incivility from supervisors to instigated incivility was not significant. The model explained 44.4% of the variance in the latent variable instigated incivility. To test whether the experienced incivility variables uniquely contributed variance, we tested a model where the paths from experienced incivility to instigated incivility were fixed to zero. The changed model had significantly worse fit, Δχ 2(2) = 119.2, p < .001, RMSEA = .068 and CFI = 0.943, according to the MPLUS DIFFTEST (the same models were tested using different versions of parcel aggregation; four were tested, but none of them was substantially different from the presented version).

The importance to instigated incivility of experienced incivility from co-workers was very obvious from the last model, but many of the independent variables in the model had moderately to strong correlations, suggesting that the variables were collinear. As an additional analysis, we tested whether the organizational variables had indirect relations to instigated incivility through experienced incivility from co-workers. The indirect relationships were estimated by multiplying the two coefficients of the paths from the independent to the dependent variable. The standardized coefficients were reported with MPlus significance test using Model 3. All the independent latent organizational variables and the observed variable organizational change were included as having possible indirect effects on instigated incivility. Since we used Model 3, all indirect effects were estimated based on the paths of this model. It was found that three of the organizational variables had a significant indirect relationship with instigated incivility; social support from co-workers (β = –.227, CI 95% [–.318, −.137], p < .001), job insecurity (β = .093, CI 95% [.028, .159], p < .005) and organizational change (β = –.093, CI 95% [.037, .150], p <  = .001).

Lastly, we found a strong indirect relationship from experienced incivility from supervisors to instigated incivility, through experienced incivility from co-workers (β = .338, p = .001, CI 95% [.219, .458]). This indirect effect was based on a model where incivility from supervisors was the sole predictor of incivility from co-workers. We did not test whether there were indirect relationships from organizational variables through experienced incivility from supervisors since this variable did not have a significant path to instigated incivility.

To summarize, the organizational latent variables organizational change, job insecurity, social support from co-workers and job demands were all related to instigated incivility, supporting the first hypotheses (a, b, c and e). However, the strongest relationship with instigated incivility was found for experienced incivility from co-workers.

Discussion

The aim of the present study was to examine possible antecedents of instigated workplace incivility, in the form of organizational aspects or having experienced incivility. There was some support for the first hypothesis that employees more frequently report having perpetrated workplace incivility when they have recently been exposed to (a) organizational changes, (b) have perceived job insecurity, (c) low social support from co-workers, (d) low social support from supervisors, (e) high job demands and (f) low control. This hypothesis was supported as significant correlations were found for all of the organizational variables and instigated incivility. A model where all the organizational variables were included resulted in four direct relations to instigated incivility. These variables were being recently exposed to organizational change and having perceived job insecurity, low social support from co-workers and job demands.

Partly in line with the second hypothesis, the results showed that being targeted by incivility from co-workers was a unique predictor of perpetrating incivility, whereas being targeted by incivility from supervisors was not.

The direct relationships between having experienced organizational change, having perceived high job insecurity, experienced low social support from co-workers, and high job demands and having instigated incivility partly support parts a, b, c and e of the first hypothesis. The relationships indicate that the nature of today’s modern working life and aspects of uncertainty and the changes associated with it can contribute to workplace incivility and are in line with previous studies of workplace bullying. Organizational change has been shown to predict the onset of workplace bullying (Oxenstierna et al., 2012), and job insecurity has similarly been demonstrated to be associated with reports of bullying at work (De Cuyper et al., 2009). Additionally, Baron and Neuman (1996) found organizational change to be tied to increased reports of workplace aggression. In the same way, experienced organizational change and job insecurity are related to the perpetration of workplace incivility in the present study. Furthermore, the relationship between low social support and instigated incivility in the present study is comparable with results from an earlier study concerning bullying (Hansen et al., 2006). Those authors found experiencing low levels of social support to be related to being targeted by workplace bullying.

High-strain jobs, defined through the Demand Control model (Karasek, 1979) by low autonomy and high job demands, have previously been tied to respondents being both the target and the perpetrator of workplace bullying (Baillien et al., 2011; Notelaers et al., 2012). However, in the present study only one of the components, job demands, of a high-strain job had a significant path to instigated incivility. This could imply that incivility is in some way differentiated from workplace bullying. It is thus interesting that in the present study the organizational aspects ‒ being exposed to organizational change, job insecurity, social support from co-workers, as well as job demands – contributed to perpetrating incivility. In addition, three of the organizational variables revealed indirect relationships through incivility from co-workers, suggesting that this kind of incivility is a mediator of problems in the organization that trigger uncivil behaviour. An organization having lower support from co-workers, lower levels of job insecurity and more organizational change seems to be related to a general higher level of incivility from co-workers. This underlines that there is a connection between the organizational aspects and uncivil behaviour. A potential explanation for this has been offered by Pearson et al. (2005), who suggest that organizational pressures are connected to eroding norms in the workplace and to incivility.

These results may carry theoretical implications for the field of workplace incivility, as they indicate that the phenomenon should be studied in the organizational context. Furthermore, the development of interventions against workplace incivility could be improved by including the importance of organizational aspects.

The results partially supported the second hypothesis that being targeted by incivility from co-workers contributes unique variance to instigating incivility. Interestingly, this is similar to findings in the literature on both bullying and workplace aggression. For both these constructs, it has been shown that being a victim is the strongest predictor of perpetration (Glomb & Liao, 2003; Hauge et al., 2009). Gallus et al. (2014) also found this relationship for workplace incivility, which is in line with this study’s findings. The fact that the direct relationship to perpetrated incivility only appeared from co-workers and not supervisors is an interesting addition in relation to power positions. The strong contribution of co-workers’ incivility stands in contrast to the literature which proposes that incivility often starts at the top of an organization and is exerted downwards (Estes & Wang, 2008; Pearson & Porath, 2009). That co-workers to a higher degree than supervisors affect the individual’s behaviour could possibly be explained by a larger degree of identification with colleagues than with supervisors. A study by Ferguson and Barry (2011) showed that individuals were more likely to act in a deviant manner if they had information about deviant action by colleagues, particularly if group cohesion was high. This implies that co-workers model the behaviour of their colleagues, increasing the tendency to act accordingly themselves. In line with these results, the importance of co-workers’ behaviours for instigating incivility could be highlighted in interventions against workplace incivility that over time can escalate to more aggressive forms of behaviour, as suggested by Pearson et al. (2005). A similar type of process could underlie the relationship found when testing this study’s second hypothesis. In support of the incivility spiral proposed by Andersson and Pearson (1999), several scholars have noted that incivility could occur as retaliation for being targeted by incivility from co-workers (Greenberg, 2010; Jex et al., 2010). Moreover, the overlap of individuals who have been found to rate themselves as both targets and perpetrators (Gallus et al., 2014) may indicate that incivilities are exchanged among colleagues. Similar social exchange perspectives and reciprocity models have been discussed in relation to workplace harassment (Bowling & Beehr, 2006) and workplace bullying (Parzefall & Salin, 2010). This mechanism of reciprocal exchanges and the contagious nature that the social process of workplace incivility contains could potentially explain the relationship between perpetrating incivility and experiencing co-workers’ incivility that was found in this study.

The findings are particularly relevant to employees in the educational sector (as in the present study), a sector that has been shown to be especially aggression-laden (Fox & Stallworth, 2010). Focusing on the role of co-workers’ interactions could be beneficial to the school sector to prevent incivility from escalating into more aggressive forms of mistreatment. This is in line with previous research (Wittmer, Sinclair, Martin, Tucker, & Lang, 2013) suggesting a prevention-focused approach to reduce workplace aggression in educational environments. Wittmer et al. (2013) found that workplace aggression has adverse consequences, not only for the affected individuals, but for the entire organization, showing the importance of approaching workplace incivility at an early stage in order to prevent a culture of uncivil behaviour from being developed.

Limitations

The cross-sectional nature of the present study limits the ability to draw any conclusions regarding actual antecedents of workplace incivility, as there is no way to establish causality. In the discussion, causality has been implied, but these causal paths have to be followed up in future studies, maybe with a more experimental or longitudinal research methodology.

The study would have benefited from a larger sample size in order to assert even larger statistical power in the estimations, as well as a higher response rate to the questionnaire. It was not possible to investigate factors contributing to the rather low response rate. In addition, as the sample was drawn from the school sector, the findings are not necessarily generalizable to other sectors.

One of the variables, organizational change, was measured with a single item, making it somewhat more sensitive to different interpretations by the participants. Such differences could have attenuated this variable’s relationships. This could potentially be remedied in future studies by using a number of objective measures of organizational change, so that results do not only reflect the individual’s perception.

Future research

In future studies, it would be important to expand the work on antecedents of workplace incivility and supplement the current literature with longitudinal data in order to establish causal relationships of triggers to workplace incivility. In addition to this, the relationships between perpetrated incivility and organizational aspects in the form of organizational change, perceived job insecurity and social support and job demands could be more vigorously approached by researchers. The relationship found between being targeted by incivility from co-workers and perpetrating incivility is an additional factor in need of further investigation, as it was shown to be the largest contributor to instigated workplace incivility in the present study. Focus should be applied to determine whether this finding can be replicated and if it is present in other sectors than the school sector. It could also prove beneficial to investigate the different outcomes of co-worker’s and supervisor’s incivility, as only incivility from co-workers had a significant direct relationship with instigated incivility. The results of the present study suggest that organizational variables contribute to workplace incivility, which stresses that incivility should be examined in its organizational context.

Conclusion

Overall the organizational variables were related to instigated incivility. A direct relationship was found between the perpetration of incivility and the organizational variables organizational change, job insecurity, low social support from co-workers and job demands. However, the strongest relationship was found between experienced incivility from co-workers and instigated incivility. The results of this study could be used as a base for developing intervention initiatives in organizations to prevent workplace incivility from escalating to more aggressive forms of behaviour. The findings highlight the importance of focusing on the perspective of the perpetrator as a basis for gaining knowledge about the phenomenon of workplace incivility in working life of today.

Disclosure statement

No potential conflict of interest was reported by the authors.