Introduction

Workplace bullying is a particularly insidious form of counterproductive workplace behavior with an increasing research stream developing in the last two decades (Branch et al. 2013; Hershcovis 2011; Neall and Tuckey 2014). The persistent abuse aimed at targets of workplace bullying results in a hostile work environment for all stakeholders of the organization. Estimates for exposure to bullying in American workplaces vary considerably, with reports of approximately 15% up to nearly 50% of workers experiencing bullying either directly or indirectly (Keashly and Jagatic 2011; Namie 2016) with annual costs estimated at over $40 billion including sick leave, absenteeism, lost productivity, and decreased employee morale (Indvik and Johnson 2012). Bullying is not a uniquely American phenomenon and researchers across the globe have found that between 10% and 18% of workers report exposure to this destructive behavior (Nielsen and Einarsen 2012; Zapf et al. 2011). The general conclusion can be drawn that workplace bullying is not an insignificant fact of work life and one that organizations are wise to address promptly and effectively.

Bullying is not defined as a one-time conflict or flare-up between individuals. Rather, it is a prolonged course of escalating torment that may be inflicted by an employee’s supervisor, subordinates, or peers (Einarsen 2000). Further, the target is seen to be in a less-powerful position (either formally or informally) than the bully within the work unit, rendering him or her unable to compel the bully to stop the abuse. Many targets respond to bullying by quitting their jobs in order to avoid further abuse (Brown n.d.).

Bullying can manifest in several forms. In general, researchers have found that bullying includes actions related to the work or task, actions targeting personal or psychosocial elements, and actions comprising physically threatening behaviors (Bartlett and Bartlett 2011; Dick 2010). Typical bullying behaviors include being given unrealistic performance targets or meaningless tasks, ignoring, socially isolating, or belittling the victim, and public humiliation. The typical course of a bullying campaign involves multiple forms of abuse over an extended period (typically six months or longer) (Dick 2010; Rayner and Keashly 2005; Zapf et al. 1996).

The etiology of workplace bullying is complicated, involving both individual (target and bully) and environmental (organizational culture, structure, and leadership) factors (Agervold 2009; Bowling and Beehr 2006; Einarsen 2000; Salin 2003). A majority of the research on individual factors has focused on targets of bullying with a limited study of perpetrators. An archetypal profile of bullied targets has not been discovered although numerous attributes of the target have been investigated. For example, negative affectivity (Duffy et al. 2002), submissiveness (Aquino 2000), organizational status (Hoel and Salin 2003), and personality (Glaso et al. 2012) have all been shown to influence the likelihood of becoming a target. Much of what is “known” about perpetrators of bullying has been gleaned from second-hand reports from targets or witnesses to bullying (Brotheridge 2013; Parkins et al. 2006).

The occurrence of workplace bullying represents a major challenge to ethical workplace conduct in that the organization values and norms define what any given organization considers right or wrong (Joyner and Payne 2002). The nature of workplace bullying is clearly a violation of ethics as targets are treated in a sub-human manner at the hands of the perpetrators and the abuse is condoned by the organization at large, whether explicitly or tacitly (LaVan and Martin 2008). Further, a lack of legal protections for targets of bullying renders them helpless if the organizational culture/environment and values reinforce this destructive behavior. Thus, the organizational environment and culture play a major role in the emergence of workplace bullying.

In this study, we focus on environmental factors as predictors of self-reported bullying in a public-sector organization. Specifically, the factors of interest are organizational culture, commitment to change, and leader-member exchange. We also investigate newcomer status and its relationship to reported bullying.

Environmental factors that have been studied in the past include stressful working conditions (Hauge et al. 2007), role demands and role conflict (Skogstad et al. 2011), organizational change (Baillien and De Witte 2009), and leadership style (Hoel et al. 2010; Skogstad et al. 2011). In their meta-analysis focused on the victim’s experience, Bowling and Beehr (2006) found that while individual differences accounted for little influence on the occurrence of workplace bullying, work environment factors were a much stronger predictor of bullying in the workplace. Salin (2003) posits that workplace bullying occurs as the result of a confluence of environmental factors, namely, enabling structures, incentives/motivating structures, and precipitating processes that provide the setting in which bullying can flourish.

Leadership both at the executive and work unit levels is a powerful force in shaping employee behavior. Stouten et al. (2010) found that ethical leadership is a powerful deterrent to workplace bullying through effective work environment design. The key relationship between individual employee and manager is best understood as a reciprocal exchange, often defined and studied as leader-member exchange (LMX) (Graen and Uhl-Bien 1995). LMX has been shown to influence subordinates’ work performance at all levels (objective task, citizenship, and counterproductive behaviors) (Martin et al. 2016). The quality of the supervisor-subordinate relationship is heavily influenced by leader behaviors and perceptions, with follower characteristics playing a lesser role in its assessment (Dulebohn et al. 2012). Of note is that despite the prevalence of research on abusive supervision (Tepper 2007), LMX has been understudied in the bullying literature.

Limited prior research has linked the likelihood of being targeted by bullies to worker tenure. Research has shown that novice nurses (those in their first year of service) experience high rates (over 70%) of bullying by those staff nurses assigned to train and mentor them (Berry et al. 2012). Additional research has shown that 57% of nurses with less than five years of service report being bullied, primarily by senior nurses, charge nurses, or nurse managers (Vessey et al. 2009). Preliminary interviews in our study hinted at particular abuse being leveled at newcomers akin to a ‘hazing’ ritual by coworkers to gain work unit acceptance. We therefore explore the phenomenon of newcomer hazing as an additional, albeit unknown, aspect of workplace bullying.

Importance and Rationale of the Study

Research shows that effects of bullying on workers and workplaces is serious and expensive, and detrimental outcomes to individuals and organizations have been identified in numerous studies (Aquino and Thau 2009; Bartlett and Bartlett 2011; Nielsen and Einarsen 2012). Targets of workplace bullying experience psychological and emotional problems such as depression, anxiety, and post-traumatic stress disorder (Hansen et al. 2011; Nielsen and Einarsen 2012; Zapf et al. 1996) coupled with reduced job satisfaction and organizational commitment and increased absenteeism and propensity to quit their jobs (Deery et al. 2011; Hoel et al. 2011).

Further, much of the research in workplace bullying in the United States has been conducted in private enterprises, with less attention in public entities (Boyne 2002; Glasø and Einarsen 2008; Strandmark and Hallberg 2007). However, workplace conditions in the public sector are often quite different from those in the private sector, with greater union representation and more traditional employee benefit programs leading to longer average tenure for workers (Boyne 2002; Glasø and Einarsen 2008; Strandmark and Hallberg 2007). A sense of “job lock” may permeate the workforce, as generous retirement benefits and job security are difficult to match in private employment. Therefore, public employees may remain in less-than-ideal work settings in the hopes of “making it” to retirement unlike their counterparts in the private sector who may find it easier to change jobs if conditions are unpleasant (Boyne 2002; Glasø and Einarsen 2008; Strandmark and Hallberg 2007). The present study offers a rare look at the organizational culture within a public works department of a municipal workplace and its impact on the occurrence of workplace bullying and newcomer hazing.

Workplace bullying is a complex phenomenon with numerous antecedents and consequences. As noted by Branch et al. (2013), the academy is beginning to recognize the complex nature of bullying and the interplay among the individual and environmental elements that can lead to bullying. In this paper, we investigate elements of the work environment (organizational culture, organizational change, leadership relationship quality, and newcomer status) rather than attitudes or other attributes of workers in an effort to focus on precursors to bullying that research has shown to exert greater influence on its occurrence (Bowling and Beehr 2006).

We integrate theories from several disciplines, notably, organizational behavior (culture and change), leadership (LMX), and psychology (occupational stress) to investigate the following issues:

  1. 1.

    We explore additional workers’ perceptions of work environment characteristics, primarily related to organizational culture and change, to determine their relationships with the occurrence of reported workplace bullying.

  2. 2.

    We investigate the role of leader-member exchange (LMX) in influencing the occurrence of reports of workplace bullying.

  3. 3.

    We examine the phenomenon of newcomer hazing as a specific manifestation of perceived workplace bullying.

Literature Review and Hypotheses

Work Environment

The work environment hypothesis (Hauge et al. 2007) postulates that specific circumstances in the work context, such as stressful and poorly organized work conditions, threats generating from organizational change, and deficient leadership may set the stage for bullying to occur. This perspective has become more widely accepted by researchers, particularly in light of findings that individual differences in organizational actors play a lesser role in the emergence of bullying (Bowling and Beehr 2006).

Seminal research by Brodsky (1976) introduced the notion that bullying can only occur in an environment that is accepting and favorable for its emergence. Salin (2003) builds on Brodsky’s approach along with previous research by Einarsen and Skogstad (1996), Vartia (1996), Zapf et al. (1996), and others to explicate three processes or structures that, when combined, can lead to workplace bullying: enabling structures, incentives/motivating structures, and precipitating processes. It is not necessary for a work environment to have elements from all three structures or processes; however, her major premise is that bad actors cannot operate in contrast to accepted work culture. The ‘host environment’ must be hospitable and agreeable to the dysfunctional behavior for it to persist.

The first of Salin’s (2003) three categories of antecedents are enabling structures, which are necessary but insufficient conditions for bullying to occur. They can be seen as the ‘fertile soil’ in which the seeds of bullying may be planted and flourish. Enabling structures include a perceived power imbalance (a classic element of bullying), low perceived costs (such that getting caught carries only a nominal penalty), and dissatisfaction and frustration (including lack of autonomy or clear goals). Additionally, the leadership style of those in charge can also be seen as an enabling process when that leadership is weak or inadequate (Einarsen et al. 1994; Leymann 1996). Further, bullying can more easily occur when the organizational culture allows or encourages such conduct. If bullying is considered the norm, new managers will behave accordingly. Bandura’s (1973) Social Learning Theory argues that people observe salient others and mimic conduct that appears to benefit them. Therefore, integrating oneself into an organization where bullying is ‘normal’ would likely lead one to also bully others.

Incentives or motivating structures establish the rewards for certain behavior in the organization. Motivating structures or processes include internal competition, organizational reward systems and expected benefits, and bureaucracy. Incentive processes therefore provide the payoff to bullies for their abusive behavior. Workers placed into competition with one another to obtain scarce resources or management’s favor may find that bullying delivers the promised rewards (Wheeler et al. 2010). When the organization’s reward system recognizes and awards bullies with promotions, bonuses, or other compensation for achieving results without regard to method, then organization members soon learn that harming their colleagues is the way to get ahead (O'Leary-Kelly et al. 1996; Treadway et al. 2013). Finally, bureaucratic rules that make it difficult to discharge employees may lead to bullying as a way to force out members of the work group as targets often seek job transfers or quit their jobs (Berthelsen et al. 2011; Soylu and Sheehy-Skeffington 2015). Thus, bullying is more likely when conditions are ripe for its emergence, whether or not incentives for bullying are present.

Precipitating processes serve as the catalyst for bullying events. These triggers often involve some change to the organizational status quo and include downsizing, organizational changes, or changes in the work group composition. Downsizing or reductions in force can be traumatic events for organization members and create a stress-filled environment in which workers are in a fight for survival. This survival mentality can lead to destructive behavior within the work group (Iverson and Zatzick 2011; Mishra et al. 2009). Organizational change can take many forms (such as restructuring, mergers and acquisitions, and IT system implementation), but all major changes lead to uncertainty and fear for workers – a breeding ground for bullying and dysfunctional behavior (Schumacher et al. 2016; Smollan 2016). These triggering events create a setting that is unpredictable and ambiguous, leaving employees feeling vulnerable and anxious. These conditions provide the fertilizer and water to the seeds of bullying, thus stimulating destructive behavior from those predisposed to bully (Baillien and De Witte 2009). This discussion suggests that the culture of the organization will be a critical determinant of whether or not bullying is likely to occur.

Organizational Culture

Organizational culture is an overarching concept that encompasses the ‘way of life’ within any organization. Extant work by Schein (1992) resulted in the commonly-used definition of culture as the “basic assumptions and beliefs that are shared by organizational members”. Beliefs, values, formal and informal processes, the “glue” that holds the organization together all represent aspects of culture in the scholarly and popular press (Watkins 2013). Organizational culture is a somewhat amorphous construct that has been studied extensively. Researchers in the organizational culture stream have yet to coalesce on a factor structure that explicates the construct. However, organizational culture can be considered as an enabling structure with respect to the emergence or repression of bullying as posited by Salin (2003). The quality of organizational culture thus serves as an important antecedent to worker behavior and bullying within the organization (Agervold 2009; Bowling and Beehr 2006; Pilch and Turska 2015; Samnani 2013).

Dysfunctional organizational cultures are described as those that limit individual and group capabilities and reward mediocrity (van Fleet and Griffin 2006). Workers in a dysfunctional organization therefore learn that harassing and destructive behavior is not only tolerated but encouraged and rewarded. Abusive cultures have been described by Wright and Smye (1998 as cited in Pilch and Turska 2015) as “win/lose,” “blaming,” and “sacrificing.” Additionally, Hodson et al. (2006) identified organizations high in chaos as being deficient in transparency, accountability, and capacity. Chaotic and abusive organizational cultures abet bullying via implicit and explicit support for destructive worker conduct.

We adopt O'Reilly et al.’s (1991) assessment of organizational culture as refined by Sarros and colleagues (Sarros et al. 2008; Sarros et al. 2002; Sarros et al. 2005) to identify seven factors of culture: competitiveness, social responsibility, supportiveness, innovation, emphasis on rewards, performance orientation, and stability. We propose that not all seven factors of culture will influence the emergence of workplace bullying. We attempted to map these factors on to the cultural framework proposed by Power et al. (2013) in their comparative study on workplace bullying which distinguished three cultural orientations: humane orientation, performance orientation, and future orientation. Humane orientation is high when culture is fair, caring, and kind and low when culture is more transactional and relationships are formal and standardized. Performance orientation is high when culture focuses more on results than people, values accomplishment, and exhibits a sense of urgency; it is low when relationships and loyalty are valued and cooperation is important. Future orientation is high when culture emphasizes delayed gratification, diversity, and long-term relationships and low when immediate results are preferred, and policies and practices are inflexible. High humane orientation and high future orientation were found to be negatively related to bullying while high performance orientation was positively related to bullying (Power et al. 2013).

As indicators of humane cultural orientations, we identify social responsibility and supportiveness. Thus, we expect perceptions of these cultural orientations to be associated with less reported bullying. We conceptualize a future orientation as perceptions that the organization’s culture emphasizes stability, and thus also expect to find a negative relationship between it and reported bullying. Finally, we see a performance orientation in perceived organizational competitiveness and an emphasis on rewards, and so we expect to find positive relationships between these variables and reported bullying. We therefore hypothesize that the organizational culture in the current study will reflect these relationships as follows:

  • H1a: Perceptions that the organization’s culture is socially responsible will be negatively related to reported workplace bullying.

  • H1b: Perceptions of supportiveness will be negatively related to reported workplace bullying.

  • H1c: Perceptions of stability will be negatively related to reported workplace bullying.

  • H1d: Perceptions of competitiveness will be positively related to reported workplace bullying.

  • H1e: Perceptions of an emphasis on rewards will be positively related to reported workplace bullying.

Commitment to Organizational Change

The uncertainty created by organization change represents Salin’s (2003) precipitating process or trigger for bullying. In the present study, the city’s mayor recently announced a major initiative to change the organization’s culture to one focusing on improved customer service and public perception of the city’s service orientation. Bullying has been linked to a stressful work environment wherein some individuals respond to the stress by acting out in a harmful manner and committing counterproductive acts (Hauge et al. 2009; Spector and Fox 2005). Indeed, prior research has shown that workers faced with substantial changes in the organization may exhibit psychosomatic complaints in response to the heightened stress of uncertainty regarding the future as well as experiencing insecurity regarding their job status (Cheng and Chan 2008; Schumacher et al. 2016).

The stressor-emotion model (Spector and Fox 2005) posits that an organization event (such as a major culture change) is perceived by each individual who then determines if the event represents a controllable or uncontrollable stressor. Individuals who perceive a high level of stress may experience negative emotions (i.e. anger and anxiety) (Spector and Goh 2001). These negative emotions may then drive counterproductive behavior if the individual feels powerless to control the stressor (i.e. organizational change). In one of the few studies on organizational change and bullying, Baillien and De Witte (2009) found that role conflict, job insecurity, and social support were among the most potent factors related to bullying. Vakola and Nikolaou (2005) found that individuals who are most acutely impacted by the stress of change are less likely to commit to the change. Additionally, they found that negative work relationships in particular contributed to lowered commitment to organizational change initiatives. Therefore, we predict that lower levels of commitment to change will be associated with higher levels of reported bullying.

  • H2: Perceived commitment to change will be negatively related to reported workplace bullying.

Leader-Member Exchange (LMX)

Finding its roots in social exchange theory (SET; Blau 1964), leader-member exchange (LMX: Graen 1976 as cited in Graen and Uhl-Bien 1995) focuses on the specific interpersonal relationship between a supervisor/manager and his or her direct reports. The underlying theory rests on the premise that relationships between individuals evolve based on expectations and fair exchanges of valued resources (such as employee effort and manager support). Relationship quality is typically assessed on a continuum from low or economic exchange to high or social exchange (Wayne et al. 1997). Low quality relationships are transactional in nature with limited regard for the partners’ feelings or emotions. In contrast, high quality relationships embody a reciprocal exchange between the partners that results in mutual loyalty, commitment, and support (Cropanzano and Mitchell 2005).

Although SET has been suggested as an explanatory perspective for workplace bullying (Bowling and Beehr 2006; Branch et al. 2013), limited prior research could be located that directly assessed LMX in relation to workplace bullying. The closely related concept of reciprocity has been shown to influence retaliatory (Skarlicki and Folger 1997) and revenge (Bies and Tripp 2005) behavior in the workplace. Exchange relationships may also be seen as a reflection of the justice climate within an organization, leading workers to behave destructively when they perceive an injustice (Parzefall and Salin 2010).

The quality of the relationship between leader and members has been linked to employee job satisfaction (Volmer et al. 2011), turnover intentions (Flickinger et al. 2016), work engagement and job performance (Breevaart et al. 2015), and workplace bullying (Mathisen et al. 2011). Since a positive relationship between leaders and members (high LMX) implies a productive, supportive, and trusting relationship, we predict that:

  • H3: Perceived LMX will be negatively related to reported workplace bullying.

Newcomer Hazing

We located only two studies that reported on the treatment of newcomers and whether they are singled out for especially abusive behavior as part of an initiation or rite of passage into the organization or work unit. Turnover, especially early in an employee’s tenure, is costly to organizations (Ramsay-Smith 2004) as short-tenure employees have little time to contribute while costing the organization valuable training and onboarding resources. Incongruence or being ‘out of step’ with coworkers has been shown to result in higher levels of voluntary turnover (Dong et al. 2012) while perceived organizational support leads to reduced new employee turnover (Allen and Shanock 2013). We can conclude that early interactions with coworkers and one’s supervisor can exert a strong influence on the new employee’s intention to stay or to quit. Josefowitz and Gadon (1989) found that mistreatment early in an employee’s tenure led to costly voluntary turnover with 10 % of their study respondents quitting due to hazing experiences.

While the overarching construct of workplace bullying has a rich literature, there has been little investigation of the newcomer’s experience to determine if they are subjected to higher levels of abuse. Hazing is defined as “initiation rituals by which newcomers to an organization are harassed and humiliated as a test or preparation for acceptance into the group” (Bersani et al. 1980 from Ostvik and Rudmin 2001). It differs from bullying in several ways, primarily that its main goal is to bring newcomers into the group whereas the main goal of bullying is to keep the target outside of the group (Ostvik and Rudmin 2001).

Hazing is a common practice in sports teams, college fraternities and sororities, and the military as a way to inculcate new members into the group’s norms, ensure that the group’s culture is maintained, and solidify the new member’s commitment to the group and its beliefs (Cimino 2011). Novice nurses reported high levels of abuse (Berry et al. 2012) with subsequent decreases in productivity due to target’s inability to handle cognitive demands and manage their workload. Norwegian soldiers reported higher levels of hazing than of bullying, supporting our contention that newcomers may be targeted more frequently as a rite of passage to indoctrinate them to the work group (Nielsen et al. 2010). We thus propose our final hypothesis:

  • H4: Newcomer employees (with 1 year of service or less) will report higher levels of workplace bullying than those with longer service.

Method

Sample and Procedures

Data for this study were collected from employees within the Department of Public Utilities for a large Midwestern city. Within this organization, three subdivisions were targeted for study: Water, Sewer, and Public Power. For each of these subdivisions the participants self-described as field employees, customer service representatives, or management. The study was requested by the city’s mayor in his desire to understand the organizational culture as part of a larger change initiative.

Prior to survey development, 25 semi-structured interviews were conducted with volunteers from the three subdivisions. The interviews were exploratory and completed first in order to guide the questionnaire development and to determine the constructs to assess in the main study.

Based on the results of the interviews, the following study variables were selected: bullying behavior, current organizational culture, commitment to change, and LMX. Surveys were provided to a sample of 374 employees and were completed by 357 respondents (for a 95% response rate). All surveys were completed in a pen/paper format in the presence of the first author and those surveys that were incomplete were not included in the final sample.

Measures

Although some past research (Bartlett and Bartlett 2011; Dick 2010) has identified three types of bullying (work-related bullying, person-related bullying, and physically intimidating bullying), we used a unidimensional measure for three reasons. First, limited prior research studied the antecedents in our study and therefore we had no evidence to support predictions of differential effects between these three types of bullying and our predictor variables (organizational culture, commitment to change, and LMX; Bartlett and Bartlett 2011; Nielsen et al. 2010). Second, past research frequently utilizes a single bullying measure (Nielsen et al. 2010). Third, ease of analysis and presentation makes one bullying measure more appropriate. As can be seen, we use post-hoc analyses below to explore bullying as a three-dimensional construct. Thus, a unidimensional measure of perceived bullying was calculated by taking the mean of all 22 items.

Bullying

The Negative Acts Questionnaire, (NAQ-R) created by Einarsen et al. (2009), was used to measure bullying. The NAQ-R consists of 22 items (a = .95). Responses were collected using a five-point Likert-type scale with responses ranging from 1 (never) to 5 (daily). Sample items include: “being ignored or excluded”; “hints or signals from Culture, LMX, and Reported Bullying 16 others that you should quit your job”; and “having insulting or offensive remarks made about your person, attitudes, or your private life.”

Organizational Culture

O'Reilly et al.’s (1991) 54 item Organizational Culture Profile (OCP) scale was used to measure the independent variables in this study. Although items in this profile were originally developed to be analyzed with a Q-Sort methodology, we incorporated them in our survey format because we believe the items describe the most accepted and universal organizational cultural dimensions. Consistent with Sarros et al. (2005), we utilized a five-point Likert-type scale (1 for “strongly agree” to 5 for “strongly disagree”) for ease of completion in a survey format. Respondents were asked to assess the importance of the 54 items to their organization. Sample items from the subscales include: “flexibility”; “being calm”; and “being demanding”. We created five subscales by averaging their items: competitiveness (a = .68), social responsibility (a = .80), supportiveness (a = .72), emphasis on rewards (a = .86), and stability (a = .74). It should be noted that the performance orientation subscale used by Sarros et al. (2005) was not included in this study due to its low reliability (a = .63).

Commitment to Change

Herscovitch and Meyer’s (2002) nine item Commitment to Change scale (a = .77) was selected. Employees were asked to “Consider the current culture change announced by the Mayor’s office when answering the following questions. Evaluate your commitment to this change effort by indicating the extent to which you agree with the following survey items.” Responses were collected using a five point Likert-type scale, ranging from 1 (“strongly agree”) to 5 (“strongly disagree”). Sample items from the subscales include: “I believe in the value of this change”; “I have no choice but to go along with this change”; “I feel a sense of duty to work toward this change.” The scale was created by averaging the nine items.

Leader-Member Exchange

Liden and Maslyn’s (1998) 12 item LMX scale (a = .92) was used to measure the quality of supervisor-subordinate relationships by asking respondents “how do you feel about the relationship between you and your current manager/supervisor?” Responses were collected using a five-point Likert-type scale, ranging from 1 (strongly agree) to 5 (strongly disagree). Sample items from the subscales include: “I like my supervisor very much as a person”; “I do not mind working my hardest for my supervisor”; “My supervisor would come to my defense if I were attacked by others.” The LMX variable was created by averaging the 12 items.

Descriptive Variables

Several descriptive variables were collected for this study. These included the division in which each respondent worked: “Water,” “Sewer,” and “Public Power,” and the position in which each respondent worked, including “Field,” “Customer Service,” and “Management.” Gender and age were also measured, as were number of years and months of service.

Methodological Issues

Common Method Variance Analysis

Since this study is based on data from a single survey, we wanted to ensure common method variance was not influencing outcomes. Accordingly, we followed the recommendations of Conway and Lance (2010), who believe studies using single surveys should 1) provide a rationale that the method is appropriate for the topic at hand; 2) show the measures have construct validity; 3) show that items do not overlap in content; 4) explain how authors minimized potential common method issues.

First, the purpose of this convenience sample was to assess workers’ perceptions. Because of the nature of the intervention, it was not possible to evaluate workers’ experiences in other ways, such as deploying neutral observers of bullying behaviors. However, we believe that the interviews conducted prior to the construction of the actual survey instrument provided proper direction and framework that clearly delineated the constructs that were relevant for this study. We do believe that more “objective” measures of these constructs would add value to future investigations, and this will be expanded upon in the discussion section below.

Second, all measures were from frequently used, validated scales. Although the organizational culture measure was used differently from its original Q-sort formulation, other researchers have used the items as we have. Thus, we have built on this technique and believe the construct validity of all the measures is strong.

Third, all scales measure distinctly different content and thus items are not redundant across the scales. To ensure this, we checked for inter-item redundancy as described below. Fourth, the order of the items was strategized so that bullying items came last in the survey, which avoided priming respondents with the bullying theme in order that their answers to the predictor variables (organizational culture, commitment to change, and LMX) would less likely to be affected.

Further, a symptom of common method variance is high multicollinearity. As can be seen, where appropriate we assessed this in our analyses. Finally, the Harman single-factor test (Podsakoff et al. 2003) found that when forced to load on to one factor, that factor accounted for only 24% of the variance, well below the 50% cut-off to determine the presence of common-method variance.

Investigation of Scale Item Independence

An anonymous reviewer suggested that inter-item correlations across the scales might be affecting outcomes. To ensure item independence, we fit the model using structural equation modeling (insufficient sample size prohibited using SEM for hypothesis testing). Modification indices, which represent the potential change in parameters should the path between two variables be freed, showed only five out of 528 possible pair comparisons with modification indices of over .15. Further, an exploratory factor analysis showed only three items (out of 54) loaded on more than one factor (based on factor loadings of .4 or higher; maximum likelihood extraction, varimax rotation). Thus, we determined that scale items were sufficiently independent of other scale items.

Analysis Strategy

Because the variables of interest in the hypotheses were continuous scales, multiple regression was the optimal data analysis strategy. In the post-hoc analyses, we were first seeking to see simple group differences (positions and divisions), thus analysis of variance (ANOVA) was most appropriate given we compared categorical variables. Further post-hoc analyses tested the hypotheses within each group using analysis of covariance (ANCOVA). This method applies the general linear model to multiple groups at a time, and shows how the dependent variables differ from group to group. ANCOVA is more appropriate than individual multiple regression equations for ascertaining how groups differ in dependent variables because it allows all groups to be simultaneously tested and thus accounts for shared variance (Cohen et al. 2003).

Results

The sample participants included more males (191, or 65%) than females (102, or 35%). The average age of participants was 28.5 years (sd = 10.5) with an average length of service of nearly five years (mean = 4.8; sd = 6.2). Forty-three percent (124) of the participants were field employees, 33% (94) were customer service, and 25% (71) identified as management. The total public works department workforce demographics were 71% male; 29% female with an average age of 34 and 9 years average length of service. The correlation matrix with means, standard deviations, and reliabilities can be found in Table 1. Thus, our sample was similar to the full departmental workforce but slightly less male, younger, and had fewer years of service.

Table 1 Correlations and descriptive statistics
Full size table

Multiple regression was conducted with reported bullying as the outcome (dependent) variable and all predictor (independent) variables entered along with the control variables of gender and age. These results can be found in Table 2. We tested for multicollinearity, finding that tolerance factors and VIFs (variance inflation factors) were within normal boundaries (over .10 for tolerance factors and under 10 for VIF).

Table 2 Hypothesis testing controlling for other predictor variables
Full size table

Hypothesis 1

The first hypothesis tests the relationships between organizational culture variables and bullying. Perceptions of social responsibility, supportiveness, and stability are proposed to negatively predict reported bullying (H1a, H1b, and H1c) and perceptions of competitiveness and an emphasis on rewards are proposed to be positively related to reported bullying (H1d and H1e). As can be seen, of the five organizational culture variables, stability was negatively associated with bullying and emphasis on rewards positively predicted bullying, supporting hypotheses 1c and 1e.

Hypothesis 2

This hypothesis predicts that perceptions of commitment to change will be negatively related to perceptions of workplace bullying. As Table 2 indicates, while commitment to change significantly predicted reported bullying, the relationship was the opposite of what was predicted: higher commitment to change was associated with more reported bullying. Thus, hypothesis 2 was not confirmed.

Hypothesis 3

This hypothesis predicts LMX will be negatively related to bullying, and Table 2 shows that it was supported: those perceiving higher LMX report less bullying in their workplaces.

Hypothesis 4

Hypothesis 4 investigates the possibility of “hazing,” and proposes that newcomer employees (with 1 year of service or less) will report higher levels of workplace bullying than those with longer service. Analysis of variance was used to test this hypothesis. Gender and age were controlled for by entering them as covariates. As Table 3 shows, the hypothesis was supported: newcomers (mean = 3.38, sd = .67, N = 71) reported significantly higher bullying perceptions than those with longer service (mean = 3.11; sd = .83, N = 215). The multiple regression model (Table 2) also shows a negative relationship between newcomer status and bullying (newcomer status coded as 1, non-newcomer status coded as 2).

Table 3 Hypothesis 4: reported bullying on newcomer status
Full size table

Summary

The first hypothesis, that perceptions of the organization’s culture will predict bullying was supported for stability and emphasis on rewards. Although results for the second hypothesis were statistically significant, they were contrary to our expectations and thus H2 was not supported: higher levels of commitment were associated with less bullying. Hypotheses 3 and 4 were supported: respondents who perceived higher LMX in their workplace tended to experience less bullying, and newcomer employees reported higher levels of bullying than did longer-serviced workers.

Post Hoc Analyses

Positions and Divisions

To understand these findings more fully, we conducted some post-hoc analyses. First, we investigated the three positions represented (field (N = 124), customer service (N = 94), and managers (N = 71) as well as the three divisions; Sewer (N = 98), Water (N = 102), and Public Power (N = 90). We were first interested in finding mean differences in reported bullying among positions and divisions.

ANOVA was conducted to test if the three positions differed in levels of reported bullying (F = 3.17, df = 2, ρ < .04). While bullying differed significantly between the field and customer service workers (mean = 3.05, sd = .86 and 3.32, sd = .63, respectively), managers did not differ from either (mean = 3.20, sd = .85). Bullying was also not significantly different between divisions. However, perceptions of organizational culture differed based on divisional membership, with the Water Division perceiving lower levels of social responsibility, emphasis on rewards, and stability than one or both other divisions. Thus, we were curious as to whether our hypotheses would show differential results, so we split the data according to position and divisions and tested the hypotheses on them separately, using ANCOVA.

Hypothesis 1: Organizational Culture and Bullying

We found differences in how organizational culture affected field workers and customer service representatives. Supporting hypothesis 1 in field workers, perceptions of stability were negatively associated with reported bullying (β = −.54, ρ = .000). However, opposite predictions, supportiveness was positively associated (β = .29, ρ = .009) and competitiveness negatively associated with bullying (β = −.26, ρ = .033) in field workers.

Also supporting hypothesis 1 in customer service workers, perceptions of stability and social responsibility negatively predicted reported bullying (β = −.44, ρ = .003 and β = −.36, ρ = .008, respectively). Perceptions of competitiveness and an emphasis on rewards were positively associated with reported bullying (β = .34, ρ = .005, β = .29, ρ = .052; note the relationship of emphasis on rewards to bullying was only marginally significant). No organizational culture variables were associated with reported bullying for managers.

Separate analyses based on division showed that as hypothesized, stability negatively predicted bullying for the Water Division (β = −.33, ρ = .017) and the Public Power Division (β = −.44, ρ = .012). No other organizational culture variables predicted bullying in any division (although competitiveness was marginally significant in the hypothesized direction in the Pubic Power Division (β = .29, ρ = .069).

Hypothesis 2: Commitment to Change and Bullying

Separate regression analyses showed positive relationships between commitment to change and bullying for field and customer service workers (β = .21, ρ = .027) and (β = .40, ρ = .000), as well as across the three divisions (β = .29, ρ = .015) for Water; (β = .26, ρ = .032) for Sewer; (β = .28, ρ = .012) for Power, results mirroring those found in the full sample.

Hypothesis 3: LMX and Bullying

When positions were individually investigated, regression analyses showed a negative relationship between perceptions of LMX and reported bullying for field employees (β = −.40, ρ = .000) as well as for customer service (β = −.57, ρ = .000), but not for managers. When the data was split based on divisions, all showed significant negative relationships (β = −.31, ρ = .012) for Water; (β = −.43, ρ = .001) for Sewer;(β = −.31, ρ = .012) for Power.

Hypothesis 4: Hazing and Bullying

When positions were analyzed separately, only field employees who were newcomers reported higher bullying. Customer service and management workers did not differ in their perceptions of bullying based on newcomer status. When data was split based on divisions, in no division did newcomers report higher bullying. This finding did not surprise us, as the preliminary interviews suggested that new field employees experienced particularly heinous abuse. One interviewee reported being locked in a sewer hole while his coworkers drove away, leaving him for 20 min unable to escape. Perhaps the high percentage of males in field service positions (84% were men) breeds a more brutal initiation for newcomers.

Bullying as a Three-Dimensional Construct

Because bullying has been construed as a three-dimensional construct, we also tested the hypotheses looking at work-related bullying, person-related bullying, and physically intimidating bullying (Bartlett and Bartlett 2011). For all three dimensions, our findings were identical to those using bullying as a single dimension. The organizational culture variables of stability (negatively) and emphasis on rewards (positively) predicted reported bullying (hypothesis 1); commitment to change did not predict reported bullying (hypothesis 2); perceived LMX was negatively predictive of reported bullying (hypothesis 3). The only varying finding was that while newcomers perceived more work-related and personal-related bullying than longer-serviced workers, reported physically intimidating bullying did not differ based on length of service categories.

Discussion

This study is notable for its investigation of an important dysfunctional workplace behavior – bullying – and a unique set of variables that have not been extensively studied as predictors of bullying. It set out to investigate four initial hypotheses regarding perceived organizational culture and LMX as antecedents to workplace bullying as well as exploring newcomer hazing within a municipal service operation. The results show support for hypothesized effects of perceptions of organizational culture as stable and emphasizing rewards and reported workplace bullying. A central tenet of a public context is the benefit of stable employment, often marked by union contracts. When employees feel secure in a stable work environment, there is less stress and subsequently less dysfunctional behavior (Spector and Fox 2005). Stability may produce a less chaotic and more harmonious work environment where employees do not feel the need to “act out” and bullying should occur less frequently. Further, we predicted – and found – that an emphasis on rewards is characteristic of a highly competitive, performance-based culture that would foster aggressive behavior that could lead to perceived bullying. Within this study, rewards such as raises and promotions could potentially be areas of contention and employees might bully others in an effort to win. In addition, the stable nature of the organizational culture might give those with a propensity to bully the perceived freedom to use bully tactics in an effort to win rewards.

Surprisingly, our findings show that commitment to change was associated with higher reports of bullying. We had predicted that those who saw change as more critical would experience higher levels of stress and thus be less prone to perceive bullying. However, our unexpected findings may suggest the alternative relationship: those perceiving a more hostile work environment may crave change to improve the culture and workplace relationships. Further, pre-study interviews yielded a consensus that the mayor’s proposed “change” was not to be taken seriously and that there was little belief that any real change would ever take place. Typical comments from interviewees were:

“Nothing ever changes . . . nothing and we know it never will. So I just show up, do my job, keep my head down, and count the days toward retirement.”

“Nothing ain’t ever going to change here; I’ve heard this stuff for years. Nobody believes anything is going to change. It’s a joke; I ignore those promises.”

This study also showed support for a relationship between perceived level of LMX and reported workplace bullying. This is very much in line with the LMX literature, which demonstrates that higher levels of LMX between employees and their supervisors creates a relationship where the supervisor treats the employee more favorably (Graen and Uhl-Bien 1995). Such favorable treatment might mean the supervisor will confide in the employee, offer them better job tasks, better working conditions, and potentially better opportunities for promotion. Therefore, it is plausible when the supervisor and the employee have a less than ideal relationship and lower levels of LMX, the employees might feel or perceive they are being bullied as they do not get the choice job tasks, opportunities for advancement, or enjoy a level of confidence with their supervisor. Our post-hoc analyses showed that this finding was robust across all three types of positions and two of the three divisions.

Although we found an overall hypothesized relationship between newcomer status and bullying, post hoc analyses found this applied only to field employees. This finding is of particular interest, as employees in field positions often work away from administrators and other groups that might tamp down abusive conduct. Further, citizens can directly observe public service workers and form perceptions of “city workers” based on what they see happening within a work crew. Compared to other jobs, the working conditions for field employees can be quite challenging. They work outdoors regardless of the weather and must deal with many unpleasant tasks such as cleaning sewer lines. Newer employees who are not mentally or physically equipped for such conditions might feel they are being bullied until they become accustomed to the job. For example, a new employee (10 months of service) in the sewer department stated during his interview:

“I take a lot of ribbing, they play a lot of pranks on me, but, that is part of being the newest. It would help if they took the time to train me but, they are busy and don’t have time…they aren’t paid to train me either…. I get all the worst jobs too. If we don’t have tools they make [me] use my hands to clean out the lines that get clogged.”

This might be construed as a rite of passage into the field worker context or perhaps the newer employee perceived this treatment as bullying behavior.

Non-supported Hypotheses and Post-hoc Analyses

Also worthy of discussion are the hypotheses that were not supported. First, of the five organizational culture variables we studied, only stability and emphasis on rewards supported our hypotheses. Our post-hoc analyses showed that these relationships were dissimilar across the types of workers and divisions. For example, three of the five organizational culture variables were associated with reported bullying in the predicted directions for field and customer service workers, but not for managers. Of further interest is that while supportiveness was unrelated to reported bullying in the full sample, it was positively related to reported bullying for field workers. It may be that these workers, who work off-site and often are unsupervised, interpreted the items measuring supportiveness (team-orientation, information sharing, people orientation, and collaborative work) as more representative of their work group rather than an immediate supervisor.

Thus, it seems that the perceived organizational context differs across positions and affects perceptions of bullying differently. It may be that the requirement to work with the public, prominent for field and customer service workers but less so for managers, changes the dynamics for how bullying is perceived. Further, as our study was the first that we know of to test attributes of organizational culture, it is possible that our measure or interpretation of culture did not accurately assess crucial aspects of culture with respect to the manifestation of bullying.

Commitment to change was positively predictive of reported bullying for the full sample, but not in the post-hoc analyses. This is inconsistent with our expectation that in highly changing environments job insecurity, role ambiguity, and other work stressors will enhance bullying behavior.

Practical Implications

Analysis of the full sample showed two organizational culture dimensions were related to reported bullying and our post-hoc analyses revealed that dimensions of culture seem to be associated with reported bullying for some, but not all, positions. Thus, culture may be a critical element in promoting or discouraging bullying, but its influence may be targeted and specific. Organizational leaders spend billions of dollars investing in training, development, selection, and performance management (Gavino et al. 2012). A primary purpose of this investment is to create an organizational culture or context that will act as a catalyst for improvement and success. This study demonstrates that it is important for organizational leaders to maximize their return on this investment by gaining insight into the presence or perceived presence of bullying.

Little research has focused on workplace bullying and the LMX dyad. Our results show that good leader-member relationships may be associated with lowered perceptions of bullying. This might point to the importance of additional leadership training for supervisors.

Finally, there are also practical implications for organizations in the treatment of newcomers, particularly if there is an embedded culture of treating them more harshly than other employees. Leaders within such organizations might assess if this is the case, how such treatment might affect the newer employees’ long term, and how the larger organization might be ultimately affected. Proactive onboarding programs involving longer-term employees in the socialization and welcoming of newcomers may help to alleviate perceptions of bullying of new entrants to the organization provided the longer-term employees do not bully their mentees.

Limitations and Future Research

The contributions of this study should be considered in light of its limitations. The data were collected at the same time using a single instrument. Consequently, we cannot infer causal relationships among the variables. It is also possible that common method variance played some role in the relationship among the variables (Lindebaum and Cartwright 2010); however, given the precautions we took, as well as the low shared variance found in the Harman’s test as noted, we believe this is unlikely.

As noted above, we study reported bullying and perceptions of the environmental context. Future research should investigate actual behavior and shared contextual beliefs in order to enhance our knowledge of bullying dynamics. Nevertheless, we believe that the perception of bullying is a legitimate and important way to measure bullying – after all, if workers do not perceive they are being mistreated, there is less likelihood it will negatively affect them, even if observers believe they see mistreatment. Thus while we believe investigation into more objective measures would expand knowledge of bullying, we also affirm the usefulness of the present research.

An additional concern is the representativeness of this sample. While probably not unlike many other municipal public works divisions, it may not reflect how organizational culture, commitment to change, LMX, and newcomer status are associated with bullying in other work groups, in the private sector, or even in other national cultures.

Future research might consider looking at the study relationships from a dyadic perspective and gaining insight into the perceptions within specific work groups. It might also be beneficial to explore the differential implications of specific organizational practices likely to impact perceptions of workplace bullying. Subsequent research should also focus on newcomer employees and how their perceptions of bullying or actual bullying might manifest and influence work attitudes, performance, and turnover.