Job-Related Burnout Among Civilian and Sworn Police Personnel

Employee burnout can affect workers’ health, motivation, and job performance, and speed staff turnover. In law enforcement, burnout has been attributed to a variety of job-related, organizational, and personal factors, ranging from the danger inherent to the job to the liabilities of late shift work, tension with supervisors, and gender relations in the organization. Overlooked in almost all of these studies has been the place of civilians in police work, and how their burnout experiences differ from—or resemble—those of their sworn counterparts. This study is based on surveys of both sworn and civilian employees of 12 police agencies from across the United States. In the survey they described their extent of emotional exhaustion, and reported on features of their lives and work that have been hypothesized to magnify or minimize this stress reaction. The study found that the burnout process is a universal one, driven by virtually the same factors among both civilians and sworn officers. Difficulties balancing work and life responsibilities, the support they receive from coworkers and supervisors, the fairness of personnel policies, and several personal factors contributed to burnout levels. The implications of these findings for both research and practice are also explored.

. Given its pernicious consequences, employee burnout has long been a concern of organizational scientists. This is reflected in the many studies focusing on the prevalence and correlates of staff burnout in a wide variety of organizational settings and professions, ranging from dentistry (Ahola & Hakanen, 2007) to the military (Vinokur, Pierce, & Lewandowski-Romps, 2009).
Police have been included in this research as well. Focusing on the dangers inherent in the profession, the litany of rules that constrain employees' working lives, and their frequent and often fraught interaction with the general public, researchers have examined burnout in law enforcement work (e.g., Schaible & Gecas, 2010). Virtually all of these studies have focused just on sworn personnel, overlooking the experiences of their civilian counterparts. This is despite the fact that the effective delivery of policing services is dependent on the performance of both sworn and civilian staff members (Wright, 2010).
This study examines burnout issues in samples of nearly 500 civilian employees and more than 2,000 sworn officers representing 12 U.S. law enforcement agencies of varying sizes. The study has two goals. First, we compare levels of burnout among civilian and sworn law enforcement employees, and test if the two differ significantly. Second, regression analyses are presented that examine whether the predictors of burnout differ between civilian and sworn employees. We then conclude with a discussion of the implications of the findings for both research and practice.

The Civilianization of Law Enforcement Personnel
Since the 1950s, civilians have commanded a much larger presence in law enforcement. In 1975, Schwartz, Vaughn, Waller, andWholey (1975) enumerated some of the factors leading to increased reliance on civilians in policing. One was the growth of American cities following World War II, and corresponding increases in crime and disorder. This led to new demand for police services. Recognizing these needs, but facing difficult budgetary constraints, agencies responded by hiring civilian employees to handle administrative and clerical tasks, freeing officers to devote more attention to field duties. Second, experts in the realm of policing supported the idea of hiring more civilians. Wilson and McLaren (1972), for example, advocated hiring civilians to perform specialized upper management duties, including finance, personnel, and data processing. They argued that assigning sworn police officers to clerical duties and other record-keeping functions undermined efficiency and economy. This sentiment was echoed in the report issued by the President's Commission on Law Enforcement and Administration of Justice (1967), which held that many tasks in a law enforcement agency could be performed better by civilians with specialized training than by sworn officers. Finally, this trend toward civilianization was encouraged and legitimized by various federal and state agencies and programs. Schwartz et al. (1975), for example, referenced the creation of the Office of Law Enforcement Assistance, which later became the Law Enforcement Assistance Administration (LEAA), and its efforts to provide grants for salary subsides, training and technical assistance for civilian employees.
Echoing many of Schwartz et al.'s (1975) observations, Forst (2000) provided a more contemporary look at the civilianization of law enforcement. He traced the roots of this trend to the Professional Era of policing, a period roughly spanning 1930 to 1980. The Professional Era was marked by numerous accomplishments, including improved training of officers, decreased reliance on the spoils system, and a heightened focus on justice as the primary mission of policing. Despite these advances, police-citizen relations suffered during the Professional Era, as the desire for efficiency and reliance on new technology created a divide between officers and the public. These developments necessitated a new approach to policing, with greater involvement of the general public in helping the police prevent and solve crimes. This approach helped spur the process of civilianization and would eventually develop into the Community Era of policing.
Many police agencies initially resisted this trend because the presence of civilian employees was perceived as a threat to the job security of sworn officers, and it undermines the then-favored militaristic image of the police organizational culture (Crank, 1989). Despite these concerns, civilian employment in policing has grown steadily. From 1987 to 2007, the number of full-time civilian employees in local police departments increased by about 47%, from 94,000 to 138,000 (Reaves, 2010). In comparison, the number of full-time sworn officers in local police departments increased by only about 30% during that same time period. Civilian employees have proven to be cost-effective because they are paid less, cost less to train and equip, and usually enjoy fewer benefits while working more days each year (Forst, 2000). Civilians have been hired to take on clerical tasks, thus freeing sworn officers to perform law enforcement duties (Heininger & Urbanek, 1983). They may also assist those officers directly, aiding in a variety of field assignments. Civilians have also been hired for jobs that sworn officers may not have the skills to perform, such as photography, computing, and other technical support services (Ellison, 2004).
The civilianization of law enforcement represents a trend that is likely to continue. As Forst (2000) argued, civilianization is an aspect of organizational structure that helps departments serve their communities in a flexible and responsive manner, which is consistent with the tenets of community policing. Furthermore, the economic malaise hitting the United States following 2007, and its corresponding strain on municipal budgets, could increase reliance on less expensive civilian employees.
Although their numbers and importance are increasing, little is known about the work-related views of civilian employees or how they compare to sworn personnel. A small number of studies have explored the working conditions of civilian employees, primarily those who work in dispatch. Doener (1987), for example, found that 31 civilian dispatchers in Tallahassee, Florida, expressed elevated levels of stress when compared to 200 sworn officers, although they were not a pathologically stressed group of employees. Shernock (1988) also explored the salience of sworn status on various perceptions of the work experience among employees from 11 municipal police departments in small-to medium-sized cities in three New England states and New York State. Although their perceptions were similar on a variety of issues, the 42 civilian employees expressed greater job satisfaction, enjoyment of their work, and job security than did the 36 sworn officers included in the study. These results contrasted with Harring (1981), who found low morale among civilians working in the New York City Police Department (see also Burke, 1995;Sewell & Crew, 1984). In light of this sparse and somewhat dated research, there is a place for a contemporary and more thorough understanding of personnel issues among civilian as well as sworn police employees. This includes learning more about the general prevalence of burnout, the predictors of burnout, and how burnout differs among sworn and civilian personnel. It is these employees together who determine the functioning of the agencies in which they work.

Burnout: A Closer Look
First labeled by Bradley (1969), burnout is a form of psychological strain that afflicts those working in the human service professions, including health care, social work, and law enforcement. People working in these environments were hypothesized to evidence a high incidence of burnout due to the seemingly never-ending demands of their job and an apparent lack of reciprocity in their relationship with patients, clients, or community members (Buunk & Schaufeli, 1993). Empirical research has supported this hypothesis, finding that a perceived lack of equity in the relationship with the population served, coupled with inadequate coping mechanisms, results in higher rates of burnout (Schaufeli & Janczur, 1994). Maslach (1982) and colleagues (Maslach & Jackson, 1981) clarified the concept by noting that burnout is a multifaceted phenomenon that encompasses three core components: emotional exhaustion, depersonalization, and personal accomplishment. Emotional exhaustion, which represents a depletion of emotional energy, is often cited as the key element of burnout and the initial outcome of chronic job demands and feelings of inadequate equity with the population of individuals served (Cooper, Dewe, & O'Driscoll, 2001;Maslach, 1982). Emotional exhaustion is the condition of interest in this article.
Existing research has identified multiple correlates of emotional exhaustion, in particular, and burnout, in general. These correlates are often divided into three categories, those associated with job-related, organizational, and personal factors (Cooper et al., 2001). The first set of factors includes the rigors of the job. Shift work, for instance, has the potential to affect relationships on the job as well as with family members and friends, which can affect burnout levels (Wittmer & Martin, 2010). In law enforcement, the dangers inherent in police work can also exacerbate feelings of burnout (Schaible & Gecas, 2010). Organizational factors have also been found to correlate significantly with burnout. Perceived support from managers and coworkers, for example, is cited as having a strong mitigating influence on feelings of burnout (Schaufeli & Buunk, 1996). Features of organizational climate or culture have also been examined extensively in the burnout literature (Cooper et al., 2001). Finally, personal factors are hypothesized to influence burnout. Despite inconsistent empirical findings, gender is a commonly studied demographic correlate of burnout, with researchers hypothesizing that female employees manifest higher levels of burnout than their male colleagues (Pines, Aronson, & Kafry, 1981). As we shall see, a large majority of civilian police employees are female, so this finding is of particular potential significance. The implications of job tenure have also been examined, with researchers finding that younger employees may be more prone to burnout than older employees (Schaufeli & Buunk, 1996).

Burnout Among Law Enforcement Employees
A number of studies have explored the correlates of burnout among sworn officers (e.g., Hawkins, 2001;Kop, Euwema, & Schaufeli, 1999;Martinussen, Richardsen, & Burke, 2007;McCarty, Zhao, & Garland, 2007;Schaible & Gecas, 2010). Highlighting the demands, scrutiny, and potential danger of the job, researchers have hypothesized that police officers should manifest higher levels of burnout than do employees in other professions. The implications of this are serious, as burnout has been found to affect the quality of service provided to citizens (Hawkins, 2001;Jackson & Maslach, 1982).
How those results apply to civilian law enforcement employees is unclear. On one hand, all jobs in law enforcement are demanding, requiring employees, both sworn and civilian, to make emotional connections with citizens or suppress emotions when being exposed to information about crime. This type of emotional labor inflicts a toll on employees, which can result in high levels of burnout (Guy, Newman, & Mastracci, 2008). On the other hand, aspects of the work environments occupied by sworn and civilian employees differ. For example, civilians are often described as occupying a lower stratum in the police hierarchy. This manifests in lower levels of pay, fewer benefits, and less training, when compared to sworn officers. Inadequate training, in particular, has been found to be related to high levels of burnout as those individuals feel less prepared for handling the wide array of ordeals that can arise in emotional work, like law enforcement (Newman, Mastracci, & Guy, 2005).
Occupying a lower stratum in law enforcement organizations may also result in lower levels of respect from their sworn counterparts and supervisors. Civilians may be looked upon unfavorably by their sworn colleagues, who feel that a civilian employee is depriving a member of the force a desirable detail or assignment (Chess, 1960;Swanson & Territo, 1983). Sworn officers may also harbor animosity toward the general public, which may be perceived as overly critical or incapable of understanding the true nature of police work. These feelings may then be transferred to civilian law enforcement employees, further widening any divide between the two groups (Chess, 1960). Furthermore, upper management may neglect civilian employees because they are preoccupied with street policing and less concerned about administrative functions (Chess, 1960). As a result, as Guyot (1979) noted, sworn officers may be treated like members of the organization while civilians may be treated like employees of the organization. This perception was corroborated by Burke (1995), who found that civilian employees felt as if they were treated as second class citizens in the organization. Their lower position in the organizational hierarchy, coupled with low pay, inadequate training, and potential tensions with sworn officers and supervisors, suggests that civilians may enjoy a less burnout-moderating job environment than that faced by officers, even in the same organization.
But the job experiences of civilian and sworn personnel also differ. Sworn officers typically have more personal interaction with the general public than do civilian employees. They may feel a need for hypervigilance, as these encounters have the potential to become volatile or dangerous, and can call for discretionary decisions regarding arrest or the use of force (Schaible & Gecas, 2010). Heightened scrutiny over the handling of those situations by supervisors, the media, prosecutors and the courts is a constant feature of their working lives. While civilian employees certainly serve and interact with the general public, they may not face those circumstances to the degree that they are faced by sworn officers (Sewell & Crew, 1984). These differences in the work environment and job experiences between civilian and sworn personnel may have implications for both the levels and predictors of burnout for the two groups of employees. The methods used to test this proposition are outlined below.

Method Participants
The participants in this study included 486 civilian law enforcement employees and 2,078 sworn personnel representing 12 law enforcement agencies across the United States. Descriptive statistics for the civilian and sworn personnel are presented in Table 1. Among civilian employees, close to 70% were female; about 40% had attained a Bachelor's degree or higher; approximately 30% were African American, and 16% Hispanic. The average civilian employee had been on the job approximately 13 years. Among sworn personnel, 16% were female, and over half had attained at least a Bachelor's degree. Approximately 11% were African American and 17% reported being Hispanic. The average sworn officer had been on the job for over 17 years. Table 1 also presents tests of the statistical significance of sworn-civilian differences for the continuous measures used in this study, including the dependent variable. These provide a guide for identifying differences between the samples, but they should not be confused with the substantive significance of the differences, because of the large samples involved.

Data and Measures
All data and measures for this study were drawn from a 55-item burnout, health, and job satisfaction survey conducted as part of the National Police Research Platform, and funded by the National Institute of Justice. In total, 28 agencies were selected to participate in a series of Platform surveys, pertaining to a variety of topics. The 28 agencies were selected to represent six broad size bands and the nation's geographic regions. Most were municipal police departments, but county police, county sheriffs and one tribal agency participated as well. The burnout, health, and job satisfaction survey was fielded in 12 of those 28 agencies. This subset included eleven municipal agencies and 1 full-service county police agency. The 12 agencies are quite diverse in their location and size. Three of the agencies are located in the Northeast; four are in the Midwest; two each are located in the Southeast and West; and one is found in the Southwest. Four of the agencies serve populations of 250,000 or above, while the other eight serve populations of fewer than 250,000. Collectively the 12 do not in any way constitute a "sample" of agencies, but they do reflect some of the diversity of American policing. In line with our agreement with the organizations, their identities are not revealed in this report.
Procedures. The surveys were conducted online using software hosted by Qualtrics, Inc© between April, 2010, and January, 2011. Civilian and sworn personnel completed essentially the same survey, with an appropriate skip pattern for questions about their job status. Participation was voluntary, and completing the survey took an average of 8.3 minutes. Procedures for administering the online survey varied slightly with agency size. In the three largest agencies, this survey was presented to a random subsample of employees; other subsets of employees participated in different studies. In nine smaller agencies the same survey link was offered to all members of the organization.
Administration of the surveys was coordinated with agency representatives, and the e-mail invitations to participate that were sent to the employees came from the departments' senior leaders. This mailing, and subsequent reminder messages, included a link to the survey site. The home page for each survey described the project's potential risks and benefits; respondents who chose to participate were next presented the survey questions.
The average agency response rate for the survey was 57% for sworn employees and 39% for civilian employees. The gender breakdown of the civilian sample, which was 68% female, mirrors exactly the 2009 national average for all civilian employees of municipal agencies in the United States (U.S. Department of Justice, 2010).
Dependent variable. Burnout was measured using a four-item composite scale (α = .91). The items were modified, with permission, from Maslach's Burnout Inventory (MBI), which is the most widely used measure of burnout (Cooper et al., 2001). A subset of the 22-item MBI was chosen because of the severe time constraints imposed by the Internet survey format; this issue is revisited below, in the "Discussion" section. The four items that were chosen reflect the emotional exhaustion component of burnout, which has been hypothesized to be most prevalent among social services professions such as law enforcement (Maslach, 1978). They asked respondents about the frequency with which they felt "used up" at the end of the day, burned out from their work, frustrated by their job, and emotionally drained from work. Responses were scored on 7-point Likert-type scales using these response categories: never, less than once a month, once a month, 2 to 3 times a month, once a week, 2 to 3 times a week, and daily. A high score on the composite measure indicates frequent manifestations of emotional exhaustion. Like all measures in this study, a full list of the survey questions can be found in the Appendix.
Independent variables. Based on the literature, measures of a number of independent variables were included in the survey. They assessed job-related stress factors, organizational stress factors, and demographic factors linked to reports of burnout.
Three job-related measures were included in the multivariate models. The first, perceptions of danger, was measured by a two-item scale (α = .75). The scale reflects the unique pressures of working in law enforcement, including threats to personal safety. For example, one item solicited reactions to the statement "I am concerned for my safety due to the danger of the job." While feelings of danger are presumed to be felt more so by sworn officers than civilians, the increasing use of civilians in field and other related assignments may make these prominent for them as well, so they are included for all respondents in this study. The two items were measured on a 5-point Likert-type scale, with response categories ranging from strongly agree to strongly disagree with a neutral (neither agree nor disagree) middle category. A high score on this index indicates high levels of perceived danger.
Second, work-life conflict was measured by a two-item scale (α = .76). The scale reflects the implications that the job has for spending time and maintaining relationships with family and friends. One item queried responses to the statement "My family and friends don't see me enough because of this job." Difficulties with balancing work and life responsibilities often fall disproportionately on women (Davidson & Cooper, 1994). This measure thus may be especially important when studying civilian law enforcement employees, given the large percentage of women working in these positions. Both items were measured on a 5-point Likert-type scale, with response categories ranging from strongly agree to strongly disagree with a neutral (neither agree nor disagree) middle category. A high score on this index indicates high levels of worklife conflict.
Third, shift assignment work was measured dichotomously, identifying individuals working midnights in contrast to working mornings or afternoons. Late and overnight work can also disrupt family life and away-from-the-job friendship patterns, and to isolate workers from many of the benefits of community life.
Three variables in the models assess different aspects of the organization and its culture. A four-item scale (α = .78) was used to measure the extent of social support from coworkers and supervisors felt by the respondent. One item in this set stated that "I feel my supervisor supports me and has my back." Responses were again measured using 5-point, Likert-format agree-disagree response categories, with a high score indicating high levels of peer and supervisor support. A three-item scale (α =.75) measures perceived unfairness of the organization in the same fashion. In an organization that is perceived as unfair, respondents replied negatively to the statement "I feel landing a good assignment is based on merit." A high score on this index indicates high levels of perceived unfairness. The analyses also controlled for the size of the organization, which was measured dichotomously. Respondents working in agencies serving populations of less than 250,000 are compared with respondents working in agencies serving populations of 250,000 or more. Brooks and Piquero (1998) argued that work in larger police agencies is more frustrating and stressful than work in smaller agencies. The tall rank structures, rigid bureaucracy, and greater distance between supervisors and subordinates in large agencies can create an environment conducive to high levels of burnout (see also Crank & Caldero, 1991).
Finally, four individual demographic factors were included in the analysis: gender, race, education, and length of service. Recognizing the historical dominance of White males in the policing profession (Haarr, 1997;Martin, 1992) and the corresponding effects on female and minority officers, studies of burnout among law enforcement agencies often examine differences by gender and race. Race was measured using three dummy variables. These captured respondents self-identifying as African American or Hispanic in origin, plus those who placed themselves in Asian, Native American, or multiracial categories. Self-identifying Whites served as the reference category. Research has also indicated that manifestations of stress and ensuing coping mechanisms may be dependent on educational attainment (Zhao, He, & Lovrich, 2002). Therefore, a dichotomous measure identifying respondents with a college degree or above was included in the analysis. Finally, we also examined the effect of the number of years respondents had been working in their current agency. Research has found that employees may experience different levels of burnout depending on whether they are early or late in their careers (Schaufeli & Buunk, 1996).

Results
The first goal of this study was to determine if there are significant differences in the level of burnout described by civilian employees and their sworn counterparts. Table 1 reports that the average burnout score was 3.79 for civilians and 3.76 for officers, a difference that is not statistically significant (t = .29, p > .05). Translating these scores into the response categories presented in the survey, both groups reported episodes of emotional burnout somewhere between once a month and 2 to 3 times a month.
A detailed examination of the burnout composite scale shows that the item probing feelings of being "used up" at the end of the day pointed to the most commonly occurring issue among both civilian employees (4.36) and sworn personnel (4.31). Translating their average scores into the response categories for the survey item, both groups felt used up somewhere between 2 to 3 times a month and once a week. Conversely, reporting being emotionally drained from work was the least problematic issue for civilian employees (3.51) and sworn personnel (3.44). Translating these scores into the response categories for the survey item, both groups felt emotionally drained from work between once a month and 2 to 3 times a month. Responses were most divergent on the item measuring how often respondents felt burned out from their work. Approximately 28% of civilian employees felt burned out from their work 2 to 3 times per week or daily. In comparison, 20% of sworn personnel felt burned out from their work 2 to 3 times per week or daily. Sworn respondents did report slightly more frequent feelings of frustration with their jobs than civilian employees as approximately 27% of sworn personnel and 26% of civilians reported those feelings 2 to 3 times per week or daily.
While average levels of burnout did not differ significantly between civilian employees and sworn personnel, the two groups did differ significantly on the stressrelated aspects of their job discussed above. Civilians expressed significantly lower (t = −23.60, p < .05) perceptions of danger (2.10) than did the sworn officers (3.23). Civilians also expressed significantly lower (t = −13.11, p < .05) feelings of work-life conflict (2.27) than did the sworn officers (2.97). Their perceptions of unfairness by the organization (3.20) were also significantly lower (t = −4.57, p < .05) than that expressed by the officers (3.42). Civilian employee perceptions of social support (3.29) were also significantly lower (t = −5.37, p <.05) than that reported by sworn (3.56) respondents.
The second goal of the study was to determine whether the predictors of burnout differed significantly between civilian and sworn respondents. Table 2 contains the results of ordinary least squares regression analyses for civilian employees and sworn personnel. The unstandardized b-coefficients are reported, along with the standard errors, the difference in the b-coefficients, and the corresponding z-values. There are many similarities in the two models. For both civilian and sworn employees, work-life conflict had a positive and statistically significant effect on levels of burnout. The direction of the coefficients indicates that, as work-life conflicts increased, so too did reports of burnout. Measures of social support and perceptions of unfairness were also statistically significant predictors of burnout for both civilian employees and sworn personnel. The negative coefficients for social support indicate that, as perceptions of camaraderie increased, feelings of burnout decreased. The positive coefficients for perceptions of unfairness indicate that as those perceptions increased, feelings of burnout also increased. Being an African American was the only demographic factor with a statistically significant effect on burnout, among both civilian and sworn personnel. African Americans recalled significantly lower levels of burnout, when compared with Whites. Differences between the two models are also apparent. Perceptions of danger had a positive and statistically significant effect on burnout, but only for sworn officers. The direction of the coefficient indicates that as perceptions of danger increased, so too did reported levels of emotional burnout among sworn respondents. Agency size also had a positive and statistically significant effect on burnout for sworn officers, but not for civilian employees. The direction of the coefficient indicates that sworn respondents in agencies serving populations of less than 250,000 had higher levels of burnout than those sworn respondents in agencies serving populations 250,000 or greater. This runs counter to the conclusions of the few studies that have examined this matter. Gender was also a positive and statistically significant predictor of burnout among sworn personnel, but not for civilian employees. Female officers reported significantly higher levels of burnout than did male officers. Finally, sworn Hispanics reported lower levels of emotional burnout when compared to Whites, but this was not the case for civilian employees. The adjusted R-square measures for civilian employees (.37) and sworn personnel (.32) were similar, and moderate.
To test whether the estimated b-coefficients for civilian employees and sworn personnel were equivalent, or significantly different, the following formula outlined by Paternoster, Brame, Mazerolle, and Piquero (1998) 1 is the unstandardized coefficient for civilian employees and b 2 is the unstandardized coefficient for sworn personnel. SEb 1 is the standard error of the b-coefficient for civilian employees. SEb 2 is the standard error of the b-coefficient for sworn personnel. The two far-right columns of Table 2 examine the difference in the unstandardized coefficients between civilian and sworn employees and whether the corresponding z-value is statistically significant. The results indicate that the effects of virtually all of the explanatory variables on burnout, with one exception, are invariant across civilian and sworn personnel. The one exception is agency size. The statistical significance of the z-value associated with the difference in unstandardized b-coefficients indicates that the effect of agency size on burnout is significantly more pronounced for sworn personnel than civilian employees.
To further test for differences between sworn and civilian employees, a one-way analysis of covariance (ANCOVA) was conducted that combined all of the civilian and sworn respondents, with burnout as the dependent variable and all of the independent variables outlined above included in the model. Then, a dichotomous variable capturing whether the respondent was a civilian or sworn respondent was entered into the analysis. The ANCOVA was statistically significant (F = 21.09, p < .05). However, only approximately 1% of the variance in burnout was accounted for by civilian or sworn status, once controlling for all of the other independent variables. The full results of the ANCOVA are available upon request. A discussion of the implications of these findings for both research and practice is outlined below.

Discussion
Civilians represent a large and growing segment of the law enforcement labor market, but there has been little research on how they fit into the world of policing. This study focused on employee burnout, which can affect workers' health, motivation, job performance and commitment to the profession. Past burnout studies focused exclusively on sworn personnel, but there was reason to suspect that civilian members of police organizations also face stress-inducing on-the-job challenges.
Our study had two goals. First, we contrasted levels of burnout between 486 civilian law enforcement employees and 2,078 sworn officers. They represented 12 law enforcement agencies spread across the United States. There was no statistically significant difference between the two groups. Both reported that being "used up" at the end of the day was their biggest burnout issue, one that on average affected them between 2 to 3 times a month and once a week. Furthermore, an ANCOVA found that when all of the other independent variables in the study were controlled for, the two groups only differed slightly in terms of burnout level.
The second goal of the study was to determine if the predictors of burnout differed between civilian and sworn employees. Separate regression models, and subsequent tests of the differences in unstandardized b-coefficients, indicated that the burnout process is fairly universal. The effects of virtually all of the independent variables relating to job-related, organizational, and demographic factors, except for agency size, were invariant across civilian employees and sworn officers. This is not to say that the views of the two groups were identical. While the significance tests presented in Table 1 should be interpreted only as a guide because of the large samples involved in this study sworn participants were significantly more likely than civilians to report that the dangers of the job affected them, their families, and their friends. This was also reflected in the significantly higher reports of work-life conflict expressed by sworn respondents. Relations on the job also differed, with sworn respondents significantly more likely to think that their coworkers and supervisors were supportive, and "have their back." This could reflect the traditional solidarity of the sworn ranks, while civilians by contrast may see themselves outside of the "blue curtain" that signals the boundaries of this cohesiveness. On the other hand, officers were also more likely to think that their agency's personnel policies involving job assignments and discipline were unfair, another correlate of burnout. The effect of perceived unfairness-which is reflected in the b-coefficients in Table 2-was very similar among sworn and civilian employees, but more officers agreed that assignments were unfairly distributed and discipline unevenly enforced in their agency, driving up their average burnout score.
Several of the findings reported here run counter to expectations created by past research. One was the mixed consequences of agency size, which had a significantly different effect on burnout for sworn and civilian employees. Officers (only) serving in larger agencies reported lower levels of burnout. Our expectation was that bureaucratic processes and disconnects between line workers and top leaders that are characteristic of larger organizations would encourage burnout, but for officers it seems instead to relieve them from some of the performance tensions that-we find-permeate smaller organizations. We also found that African American employees-sworn and civilian alike-and Hispanic sworn officers reported less pervasive symptoms of burnout than did their White counterparts. Whether this finding is indicative of minorities being a relaxed and content group of employees or Whites being an anxious and uneasy group of employees is not clear given the absence of an appropriate benchmark to which the results can be compared. However, we attribute some of these differences in burnout between races and ethnicities to divergent perceptions of work-life conflict and unfairness. Grouping together civilian and sworn employees, African Americans (2.51) and Hispanics (2.69) expressed lower average perceptions of work-life conflict than did White employees (2.93). Furthermore, African Americans (3.28) and Hispanics (3.22) expressed lower average perceptions of unfairness than White employees (3.41). The underlying reasons why minority law enforcement employees, both civilian and sworn, perceived less work-life conflict and unfairness from the organization are beyond the scope of the current study, but provide an avenue for further research, both quantitative and qualitative, into how burnout and other perceptions of work differ across races and types of employees. Finally, consistent with expectations from prior research, female officers (16% of the sworn total) reported significantly higher levels of burnout than male officers, but this difference was not apparent among the civilians, where females made up almost 70% of the group.
The findings also have multiple practical implications. First, initiatives to help law enforcement employees understand and manage stress must include civilian employees, given their comparable levels of burnout to sworn personnel. Access to stress management training, for example, has been found to be a useful initiative for some forms of stressors (Dewe, 1994). Ensuring that civilians have access to such training should be a priority for law enforcement agencies. Other initiatives that target civilians, like mentoring programs, could also be implemented. These programs have been successful in helping minority and female officers increase camaraderie and mitigate stress in law enforcement organizations (Gibb, 1999). Ultimately, programs of this nature can help the health and job satisfaction of civilian employees while also minimizing the risk of staff turnover. Second, organizations must take steps to ensure more equitable systems of reward and punishment distributions. Perceptions of unfairness were high, particularly for sworn personnel. Chiefs should be cognizant of these concerns as they are found to significantly increase levels of burnout among all employees. Finally, a greater emphasis on the health and well-being of law enforcement employees in small agencies is needed. This emphasis can take the form of both stress prevention and treatment. Prevention can involve the restructuring of organizational units of redesigning the physical work environment (Elkin & Rosch, 1990). Treatment can entail counseling services or employee assistance programs.
There are a number of limitations of this study. First, the respondents were drawn from only 12 police agencies, albeit organizations of widely varying sizes and located around the United States. Second, our use of the internet constrained the length of the survey, because it is easy for respondents to break off if other tasks intrude or if the study becomes too onerous. We aimed for a 10-min completion period based on focus group discussions. This placed limits on the number of concepts that could be measured, and the number of items that were available for the scales that were developed. As a result, we were only able to replicate a small portion of the MBI, the full version of which includes 22 items. Additional items geared specifically toward civilian employees, capturing the amount and nature of their contact with fellow civilians, sworn officers, and the general public also had to be omitted. The inclusion of items of this nature may indicate more variance in the predictors and prevalence of burnout than what is suggested by the current study. Finally, the response rate was 57% for sworn officers but only 39% for civilians. However, these rates are based on complete lists of employee e-mail addresses, and the opportunity to participate in the surveys reached every corner of the organizations. Nonresponse included employees who were in court or on special assignment, on vacation or sick leave, or enjoying compensatory time off, setting a high standard. Furthermore, the disparity in the rates between sworn and civilian respondents could be indicative of multiple things. For example, civilian law enforcement employees may not be familiar with being surveyed since most studies are focused on sworn officers. Or, civilian employees may not have responded because of their high levels of stress and burnout. While the answer is not certain, it does suggest the need to conduct more research into the work-related perceptions of this understudied group.
An obvious next step in this line research is to expand the number of participating agencies. Twelve agencies were involved in this study, but to formally model the influence of organization-level factors on employee attitudes and behavior would require an agency sample that is about three times as large. This would allow incorporation into the analysis of agency characteristics that could play an important role in exacerbating or ameliorating employee burnout. These could include systematic measures of workload, stressors in the environment, bureaucratization, and even employee protections built into the labor contracts in the agencies. Regardless of whether these specific directions for future research are undertaken, it is hoped that researchers will continue to explore the experiences and perceptions of civilian employees, who occupy an essential and expanding role in the modern law enforcement agency.