Decoding the Code of Silence

Abstract

The code of silence—the informal prohibition of reporting misconduct by fellow police officers—has long been viewed as a serious obstacle in control of police misconduct and achievement of police accountability. The purpose of this article is to study the key correlates of police officers’ reluctance to report. Relying upon a theory of police integrity and the accompanying methodology to study the code, a police integrity survey was administered in 2013 and 2014 to measure the contours of police integrity among 604 police officers from 11 police agencies located in the Midwest and the East Coast of the United States. The questionnaire contains descriptions of 11 scenarios describing various forms of police misconduct, each followed by seven questions measuring officer views of scenario seriousness, the appropriate and expected discipline, and willingness to report misconduct. Multivariate analyses reveal that the key factor related to the police officers’ reluctance to report is the perception that the other officers would not report. The code is also negatively related to familiarity with the official rules, evaluation of misconduct as serious, and the expectation of harsher discipline. The methodology can be used either by the police agencies themselves or by the civilian oversights to assess the nature and extent of the code in the police agency.

Keywords

police integrity misconduct code of silence discipline fairness supervisors

The Code of Silence

The code of silence, the blue curtain, or the blue wall of silence is a part of the police subculture. The code refers to the informal norms in the police subculture that prohibit reporting misconduct of other police officers. In his testimony before the Mollen Commission, Bernard Cawley, a former New York police Officer, explained that the code of silence simply means that, “Cops don’t tell on cops” (New York City Commission to Investigate Allegations of Police Corruption and the Anti-Corruption Procedures of the Police Department [Mollen Commission], 1994, p. 53). Similarly, the Independent Commission on the Los Angeles Police Department [Christopher Commission] (1991), which investigated the use of excessive force and racism in the Los Angeles Police Department (p. 168), reported that an officer described the code of silence as “a non-written rule that you do not roll over, tell on your partner.” In a nationwide study of police officers, one of six police officers adhered to the view that “the code of silence is an essential part” of good policing (Weisburd & Greenspan, 2000, p. 3).

The New York City Commission to Investigate Allegations of Police Corruption and the Anti-Corruption Procedures of the Police Department [Mollen Commission] (1994, p. 51) reported that the code of silence is pervasive across the New York Police Department and that it “influences the vast majority of honest and corrupt officers alike.” About two thirds of the police officers in the Weisburd and Greenspan’s (2000) study reported that police officers do not always report even serious misconduct. Clearly, pressure exists in the police subculture to adhere to the code; commissions have documented that whistleblowers are ostracized and harassed, face physical threats, and are left alone on the streets (e.g., Commission to Investigate Allegations of Police Corruption and the City’s Anti-Corruption Procedures [Knapp Commission], 1972; New York City Commission to Investigate Allegations of Police Corruption and the Anti-Corruption Procedures of the Police Department [Mollen Commission], 1994). At best, a police officer who violated the code and reported misconduct would face a “cold shoulder” by fellow officers (Weisburd & Greenspan, 2000, p. 3).

More than 20 years after the Mollen Commission’s findings, the New York Police Department yet again finds itself amidst a developing corruption scandal (Neumeister & Armstrong, 2016; Vasilogambros, 2016). In the past few months, a number of high-ranking police commanders were charged with various forms of misconduct and reassigned to other duties. In the most extreme case, one ranking officer took his own life (Schramm, Cohen, & Fredericks, 2016). The most recent disclosures underscore the importance of the research on the code of silence. Given the nature of the current accusations (bribes from high profile businessmen) and the rank of the officers charged, it is plausible that these incidents were not isolated and a number of other officers, yet to be identified, had, at minimum, some knowledge of the alleged misconduct yet chose, once again, to remain silent. Hence, the importance of the code of silence cannot be overestimated.

The consequences of the strong and widespread code of silence are serious. For one, the code allows the misconduct to continue. Not only are police officers reluctant or unwilling to report even serious misconduct by fellow officers, but they are even willing to take the blame for misconduct by other officers (e.g., New York City Commission to Investigate Allegations of Police Corruption and the Anti-Corruption Procedures of the Police Department [Mollen Commission], 1994, p. 55). Furthermore, the code of silence separates the police from the community (e.g., New York City Commission to Investigate Allegations of Police Corruption and the Anti-Corruption Procedures of the Police Department [Mollen Commission], 1994, p. 58) and diminishes police accountability. Based on their investigation into the Los Angeles Police Department, the Independent Commission on the Los Angeles Police Department [Christopher Commission] (1991, p. 168) concluded that “perhaps the greatest single barrier to the effective investigation and adjudication of complaints is the officer’s unwritten ‘code of silence.’” Similarly, the New York City Commission to Investigate Allegations of Police Corruption and the Anti-Corruption Procedures of the Police Department [Mollen Commission] (1994, Exhibit 6, p. 6) reported that the code of silence is “the greatest barrier to effective corruption control” in the New York Police Department.

Studying the code of silence by directly asking police officers to report on fellow officers’ misconduct is riddled with pitfalls. To avoid these pitfalls, the approach we pursue in this article inquires about police officers’ views and opinions. It relies on the police integrity theory and the related methodology to explore the critical determinants of the code of silence. The article goes beyond assessing what the contours of the code are; it provides an in-depth systematic analysis of the factors that could be used as predictors of a strong code of silence.

The Code of Silence and the Theory of Police Integrity

Klockars and Kutnjak Ivković (see, for example, Klockars & Kutnjak Ivković, 2004; Klockars, Kutnjak Ivković, & Haberfeld, 2004; Klockars, Kutnjak Ivković, Harver, & Haberfeld, 1997) argued that the challenges associated with the direct measurement of police misconduct could be alleviated by addressing its opposite—police integrity. They proposed the theory of police integrity and developed the related empirical methodology (see, for example, Klockars & Kutnjak Ivković, 2004; Klockars et al., 2004; Klockars et al., 1997). The code of silence and its control are one of the four dimensions of the theory. Consequently, almost all studies using this approach study not only police integrity in general, but also the code of silence in particular. In this article, we will focus on the studies primarily exploring the code of silence.

Klockars, Kutnjak Ivković, Harver, and Haberfeld (2000) argued that the parameters of the code of silence, inherent in the police subculture, vary across police agencies. Different police agencies could have the codes that differ not only in terms of what misconduct is covered by the code, but also in to whom the benefits of this coverage extend (Klockars et al., 2004, p. 8). In their initial study of more than 3,000 police officers from 30 police agencies, Klockars and colleagues (2004) found substantial variation in the contours of the code across the surveyed police agencies. For illustrative purposes, Klockars and colleagues used two agencies—Agency 2 and Agency 23—on the opposite sides on the police integrity scale—to depict the magnitude of the differences in the contours of the code across the agencies (Klockars et al., 2000):

In both agencies, few officers said that they or their police colleagues would report any of the least serious types of corrupt behavior (Cases 1, 2, 4, and 8). Officers from Agency 2 reported that they and their colleagues would report the behavior described in the seven other cases. In Agency 23, however, there was no case that the majority of officers indicated they would report. In sum, while The Code is under control in Agency 2, it remains a powerful influence in Agency 23, providing an environment in which corrupt behavior can flourish. (p. 8)

Subsequent U.S. applications of the police integrity theory and methodology confirmed the initial finding that the code of silence differs across police agencies (e.g., Hickman, Piquero, Powell, & Greene, 2016; Hickman, Powell, Piquero, & Greene, 2016; Marche, 2009; Micucci & Gomme, 2005; Rothwell and Baldwin, 2007a, 2007b; Schafer & Martinelli, 2008; Wolfe & Piquero, 2011). Heterogeneity in the contours of the code of silence was measurable even across the agencies that could be classified as having similar levels of police integrity. Specifically, Klockars, Kutnjak Ivković, and Haberfeld (2006) studied three municipal police agencies of high integrity and noted that the code of silence in one police agency seems to be much stronger than the codes of silence in the other two agencies.

Moreover, in large municipal police agencies such as the Los Angeles, New York, or Philadelphia Police Departments, variations in the code of silence may be found not only across police agencies, but also within police agencies. Hickman, Piquero, et al. (2016) find strong district-level variations in the willingness to report among police officers from Philadelphia (see also Greene et al., 2004). Characteristics of police districts, such as its “maleness,” may strengthen the code of silence (Hickman, Piquero, et al., 2016).

Lim and Sloan (2016) report that agency types and sizes matter for the code of silence. While constable supervisors were less likely to say that they would report misconduct than supervisors from municipal and specialized agencies (Lim & Sloan, 2016), the authors argued that these findings reflect the fact that constables typically work in small agencies, in which ties among the employees are stronger. By contrast, McDevitt and colleagues (2011, p. 3) found that, compared with the officers in medium and large police agencies, officers in small agencies were most likely to perceive that other officers would report misconduct. Yet, Rothwell and Baldwin (2007b) found the effects of the agency size on the willingness to report only for the less serious forms of misconduct, but not for the most serious forms of police misconduct.

Extant research demonstrated that one of the strongest predictors of the police officers’ willingness to report is their perceived seriousness of misconduct. The more serious the evaluation of misconduct, the less likely police officers were to adhere to the code of silence. In an early exploration on the topic, Klockars and colleagues (Klockars et al., 1997, 2000, p. 6) wrote that the majority of police officers in their sample of 30 agencies would not report the scenarios evaluated as the least serious. On the other hand, the majority of the officers stated in the same study that they would report behavior evaluated as the most serious. A bivariate correlation between the code of silence and perceptions of seriousness seems to be very high (e.g., Klockars et al., 2004, 2006, 1997; Kutnjak Ivković, Haberfeld, & Peacock, 2013; Kutnjak Ivković, Peacock, & Haberfeld, 2016; Kutnjak Ivković & Sauerman, 2013). Similarly, empirical studies demonstrated that seriousness of misconduct is one of the strongest predictors, if not the strongest predictor, of police officers’ willingness to report (e.g., Hickman, Piquero, et al., 2016; Hickman, Powell, et al., 2016; Long, Cross, Shelley, & Kutnjak Ivković, 2013; Lim & Sloan, 2016; Porter & Prenzler, 2016; Westmarland, 2006).

Lim and Sloan (2016) find that the “the strength of discipline” (i.e., perceptions of disciplinary fairness) is related to the supervisors’ willingness to report misconduct. Previous research has also analyzed the bivariate relation between perceptions of disciplinary fairness and the code of silence. Kutnjak Ivković and Klockars (1998) developed three theoretical models that can be used to hypothesize about the relation between the code and the disciplinary fairness: the discipline indifference model (no relation between the code and disciplinary fairness), simple deterrence model (a negative relation between the code and discipline severity), and the simple justice model (a positive relation between the code and discipline severity). In addition to the two tests in the United States (Kutnjak Ivković et al., 2016; Long et al., 2013), these theoretical models have been tested in Bosnia and Herzegovina (Kutnjak Ivković & Shelley, 2005), Croatia (Kutnjak Ivković & Klockars, 1998), Czech Republic (Kutnjak Ivković & Shelley, 2010), and South Africa (Kutnjak Ivković & Sauerman, 2013).

A comparison between the respondents who evaluated the expected discipline as too harsh and the respondents who evaluated the expected discipline as fair yields mixed conclusions. The simple justice model is predominant in Croatia (Kutnjak Ivković & Klockars, 1998), the Czech Republic (Kutnjak Ivković & Shelley, 2010), and South Africa (Kutnjak Ivković & Sauerman, 2013). While Long and colleagues (2013) found support for the simple justice model in about one-half of the scenarios, Kutnjak Ivković et al. (2016) reported support for the simple justice model in over three quarters of the scenarios. On the other hand, the results of prior research exploring the results for the respondents who evaluated the expected discipline as fair with the results for the respondents who evaluated the expected discipline as too lenient are more uniform: in most of the scenarios, extant research reported no relation between perceived discipline fairness and the code of silence (Kutnjak Ivković & Klockars, 1998; Kutnjak Ivković et al., 2016; Kutnjak Ivković & Sauerman, 2013; Kutnjak Ivković & Shelley, 2005, 2010).

Previous research on the code of silence studied the influence of demographic variables on the code of silence as well. Rank was frequently explored, but the results are mixed, either proposing no relation or suggesting that supervisors are less likely to adhere to the code (Hickman, Powell, et al., 2016; Huon, Hesketh, Frank, McConkey, & McGrath, 1995; Klockars et al., 2006; Kremer, 2000; Kutnjak Ivković & Klockars, 1998, 2000; Kutnjak Ivković et al., 2016; Kutnjak Ivković and Shelley, 2008; Lobnikar, Prislan, Cuvan, & Mesko, 2016; Long et al., 2013; Pagon & Lobnikar, 2000; Porter & Prenzler, 2016; Weisburd and Greenspan, 2000). Gender and race are used less frequently, but the outcome is similar—no uniform findings (e.g., Hickman, Piquero, et al., 2016; Hickman, Powell, et al., 2016; Lobnikar et al., 2016; Long et al., 2013; Porter & Prenzler, 2016)

Method

The Sample

In 2013-2014, we surveyed 11 U.S. police agencies, primarily located in the Midwest and on the East Coast of the United States. These agencies range from very large municipal agencies with more than 500 sworn officers to very small police agencies with fewer than 25 sworn officers. They constitute a convenience sample, seeking to obtain the views of a diverse range of police departments. About three quarters of the respondents are coming from either medium-size police agencies with about 76 to 200 sworn police officers (46.5%) or large police agencies with 201 to 500 sworn police officers (26.3%). The sample also contains respondents from very large police agencies with more than 500 sworn police officers (17.4%), small police agencies with 25 to 75 sworn police officers (7.3%), and very small police agencies with fewer than 25 sworn police officers (2.5%).

A comparison of the characteristics of the police agencies in the sample with the characteristics of local police agencies in the Bureau of Justice Statistics (BJS; 2011) survey of local police agencies indicates that the distribution of officers in 11 surveyed agencies is similar to the distribution of police officers working in large, medium, and small cities in the United States. (Kutnjak Ivković, Haberfeld, & Peacock, 2015; Kutnjak Ivković et al., 2016). For example, 2.5% of the respondents in our study were from small police agencies with fewer than 10 officers, whereas nationally more than 5% of officers work in the smallest agencies (BJS, 2011).

The chiefs of these police agencies agreed to the use of an online survey. We posted the questionnaire on Survey Monkey and password protected it. The response rate for the whole sample was 34%, consistent with what is viewed the appropriate response rate for one-time web surveys (e.g., Shih & Fan, 2008). However, the response rate was not identical across agency sizes; it was the lowest in the largest police agency in our sample (15%) and the highest in a large police agency (69%).

Because of the nature of the sampling (an online, one-time survey, voluntary participation), our sample contains only the answers provided by the respondents who voluntarily chose to participate. One of the issues is a potential for an embedded sampling bias (i.e., respondents who choose to participate may be more conscientious or have higher levels of integrity). However, our analyses go beyond simply assessing the extent of the code of silence (more likely to be affected by the potential sampling bias); they provide an in-depth systematic analysis of the factors that could be used as predictors of a strong code of silence (less likely to be affected by such bias). Moreover, even in the course of our measurement of the code of silence, we primarily compare responses across scenarios (e.g., how perceptions of seriousness compare across scenarios), rather than draw absolute conclusions about the extent of the code.

To maximize the likelihood that officers would participate and provide truthful answers, the survey was anonymous and the number of demographic questions was very limited. In addition, we purposely did not ask about their actual conduct or conduct of their fellow officers, but rather focused our questions on their views and opinions. Last, we explicitly asked them whether they lied while filling out the questionnaire and disregarded answers provided by the respondents who either openly said that they lied or did not answer this question.

The overwhelming majority (83%) of the respondents assessed that other police officers in their agencies would answer honestly while filling out the questionnaire. This percentage is virtually identical to the percentage of police officers (84%) in the survey of three large U.S. metropolitan police agencies (Klockars et al., 2006, p. 20). When asked whether the respondents themselves have answered honestly, an even larger number responded positively (91%; Table 1). About one in 10 respondents (9%; Table 1) either stated that the respondent himself or herself lied or simply skipped the question. We use only the answers provided by the respondents who reported truthfully answering the question (N = 604).

Table 1.

Respondents Demographic Characteristics.

	Number of respondents	% of respondents
Reported telling the truth
Yes	604	91.0
No	10	1.5
Missing	50	7.5
Supervisory role
Non-supervisors	430	72.2
Supervisors	166	27.8
Length of service
Up to 5 years	74	12.3
6-15 years	255	42.4
Over 15 years	273	45.3

To entice respondents to answer truthfully and to protect their anonymity, the number of demographic questions in the questionnaire was very minimal. Close to three quarters of our respondents are line officers (72.2%; Table 1). The overwhelming majority (87.7%) are also experienced officers with more than 5 years of service (Table 1). In fact, about one out of two respondents had 15 or more years of experience (45.3%; Table 1).

Questionnaire

The police integrity questionnaire relies on hypothetical scenarios describing 11 examples of police misconduct (Klockars et al., 2006; Kutnjak Ivković et al., 2013, 2015; Kutnjak Ivković et al., 2016).¹ The behaviors described in the scenarios range in terms of their seriousness and type of misconduct. Of the 11 scenarios, five depict police corruption (Scenario 1: Free Meals, Gifts From Merchants; Scenario 3: Theft of Knife From Crime Scene; Scenario 5: Supervisor Offers Holiday Off for Errands; Scenario 8: Cover-Up of Police DUI and Accident; Scenario 9: Auto-Body Shop 5% Kickback), four depict use of excessive force (Scenario 4: Unjustifiable Use of Deadly Force; Scenario 6: Officer Strikes Arrestee; Scenario 7: Verbal Abuse of Motorist; Scenario 11: Sgt. Fails to Halt Beating of a Child Abuser), one depicts the falsification of the official document (Scenario 10: False Report of Drug Possession), and one depicts the failure to execute a search warrant (Scenario 2: Failure to Arrest Friend on Felony Warrant). The respondents were told that the officer in question has no prior disciplinary record.

Upon reading each scenario, the respondents were asked the same seven questions based on the theory of police integrity (Klockars & Kutnjak Ivković, 2004). These questions focus on the respondents’ personal opinions and measure their views about scenario seriousness, appropriate and expected discipline, and willingness to report misconduct (Klockars et al., 2006; Kutnjak Ivković et al., 2015).

Dependent Variables

The dependent variables in the logistic regression models capture the respondents’ own willingness to adhere to the code of silence. The question measuring own reporting was worded as follows: “Do you think you would report a fellow police officer who engaged in this behavior?” The possible answers ranged on a 5-point scale from “definitely not” (1) to “definitely yes” (5). Answers “1” and “2” were merged into “would not report” and answers “3,” “4,” and “5” were merged into “would report.” In four scenarios (Scenario 3, Scenario 4, Scenario 9, and Scenario 10), the distribution of answers was very skewed toward “5,” so we had to merge answers “1” through “4” into “would not report” versus “5” being “definitely would report.”

Independent Variables

The respondents were asked to assess how serious the described behavior was (“How serious do YOU consider this behavior to be?”). The answers to own seriousness ranged on a 5-point scale from “1” being “not at all serious” to “5” being “very serious.” In the logistic regression analysis, the variables have been dichotomized in the following way: “1” and “2” into “not serious” and “3,” “4,” and “5” into “serious.” ²

The next question in the questionnaire inquired about the respondents’ familiarity with the legal rules (“Would this behavior be regarded as a violation of official policy in your agency?”). The possible answers to knowledge of rules ranged from “1” being “definitely not” and “5” being “definitely yes.” The distribution of answers was skewed toward “5,” so we dichotomized answers into “no” (“1” through “4”) and “yes” (“5”).³

One of the questions in the questionnaire asked the respondents to assess what discipline they thought their agency would mete out (“If an officer in your agency engaged in this behavior and was discovered doing so, what if any, discipline do you think WOULD follow?”). The possible answers to expected discipline had six answers: none; verbal reprimand; written reprimand; suspension; demotion in rank; and dismissal. They were recoded into “no discipline” (containing answers “no discipline” and “verbal reprimand”), “some discipline” (containing answers “written reprimand,” “suspension,” “demotion in rank”), and “dismissal” (containing answers “dismissal”).⁴

Discipline fairness⁵ was constructed by comparing the respondents’ individual answers to the questions about expected discipline with their answers to the corresponding question about appropriate discipline. When their answer to the question about expected discipline is subtracted from their answer to the question about appropriate discipline, the result could be a negative number (suggesting that the discipline is evaluated as too harsh), a positive number (suggesting that the discipline is evaluated as too lenient), or zero (suggesting that the discipline is evaluated as fair). Thus, discipline fairness has three categories of how expected discipline is evaluated: too harsh, too lenient, or fair.

Others reporting asked the respondents to assess whether other police officers in their agency would report the behavior (“Do you think MOST POLICE OFFICERS IN YOUR AGENCY would report a fellow police officer who engaged in this behavior?”). The possible answers ranged on a 5-point scale from “definitely not” (1) to “definitely yes” (5). Answers “1” and “2” were merged into “would not report” and answers “3,” “4,” and “5” were merged into “would report.⁶

The last two variables are the respondents’ demographic characteristics. Supervisory status measures whether the respondents were supervisors or line officers; length of service measures the respondents’ experience in policing. Length of service originally had seven possible answers which were merged into three categories (“up to 5 years,” “6-15 years,” and “over 15 years”).

Results

Measuring the Code of Silence

Table 2 illustrates the contours of the code of silence. It shows the percentage of respondents who responded that they would not report a fellow police officer who engaged in the behavior described in the scenarios and demonstrates that the code of silence is not a flat prohibition of reporting. In fact, the contours of the code substantially change from one scenario to the next. The difference in the percentages of police officers not willing to report between the scenario with the largest percentage and, thus, the strongest code (Scenario 1: Free Meals/Gifts From Merchants) and the scenario with the smallest percentage and, thus, the weakest code (Scenario 4: Unjustifiable Use of Deadly Force) is close to 60% (Table 2); about 60% of the respondents stated that they would not report a fellow police officer who accepted free meals and gifts from merchants (Scenario 1), while only 1.4% of the respondents said that the same for a scenario in which a police officer unjustifiably used deadly force (Scenario 4).

Table 2.

Extent of the Code of Silence (Percent Not Reporting).^a

Scenario number and description	% not reporting	Rank
Scenario 1—Free Meals/Gifts From Merchants	59.7	11
Scenario 2—Failure to Arrest Friend With Felony Warrant	24.5	8
Scenario 3—Theft of Knife From Crime Scene	3.5	2
Scenario 4—Unjustifiable Use of Deadly Force	1.4	1
Scenario 5—Supervisor Offers Holiday for Errands	22.6	7
Scenario 6—Officer Strikes Arrested	21.8	6
Scenario 7—Verbal Abuse of Motorist	45.7	10
Scenario 8—Cover-Up of Police DUI and Accident	41.1	9
Scenario 9—Auto Body Shop Five Percent Kickback	12.2	4
Scenario 10—False Report of Drug Possession	4.7	3
Scenario 11—Sgt. Fails to Halt Beating of Child Abuser	14.2	5

Note. DUI = driving under the influence.

The variable is coded as follows: “1” and “2” are coded as “not reporting,” whereas “3,” “4,” and “5” were coded as “reporting.”

The scenarios can be classified into three groups on the basis of the strength of the code of silence. First, about one-half or close to one-half of the respondents stated that they would not report a fellow police officer in three scenarios (Scenario 1: Free Meals/Gifts From Merchants; Scenario 7: Verbal Abuse of Motorist; Scenario 8: Cover-Up of Police DUI and Accident). These scenarios represent the least serious examples of police corruption and use of excessive force⁷ in the questionnaire. At the same time, the respondents are the least likely to recognize them as rule-violating behavior (Kutnjak Ivković et al., 2015).

On the other end of the spectrum is a group of three scenarios (Scenario 3: Theft of Knife From Crime Scene; Scenario 4: Unjustifiable Use of Deadly Force; Scenario 10: False Report on Drug Possession) in which the code seems to be the weakest (only 5% or fewer of the respondents said that they would not report such behavior; Table 2). These scenarios represent the most serious examples of police misconduct in the questionnaire (including both corruption and the use of excessive force) and were among the behaviors most likely to be recognized as rule-violating (Kutnjak Ivković et al., 2015).

We also explored the contours of the code across different forms of police misconduct. Five scenarios in the questionnaire describe examples of police corruption (Scenario 1: Free Meals/Gifts From Merchants; Scenario 3: Theft of Knife From the Crime Scene; Scenario 5: Supervisor Offers Holiday Off for Errands; Scenario 8: Cover-Up of Police DUI and Accident; Scenario 9: Auto-Body Shop 5% Kickback). Our respondents had no problems differentiating across these forms of police corruption; they were most likely to protect acceptance of gratuities (Scenario 1: Free Meals/Gifts From Merchants) and internal corruption (Scenario 8: Cover-Up of Police DUI and Accident) and least likely to tolerate opportunistic theft (Scenario 3: Theft of Knife From the Crime Scene) and kickback (Scenario 9: Auto-Body Shop 5% Kickback). The rank order of these scenarios, based on the likelihood of not reporting, is strongly and negatively correlated with the evaluations of their seriousness; the more serious the respondents evaluated police corruption, the more likely they were to say that they would not tolerate it without reporting.

The respondents also expressed substantial diversity of views across the four scenarios describing the variations in the use of excessive force (Scenario 4: Unjustifiable Use of Deadly Force; Scenario 6: Officer Strikes Arrestee; Scenario 7: Verbal Abuse of Motorist; Scenario 11: Sgt. Fails to Halt Beating of Child Abuser). In particular, the respondents were most tolerant of verbal abuse (45.7% not reporting; Scenario 7: Verbal Abuse of Motorist) and least tolerant of abuse of deadly force (1.4% not reporting; Scenario 4: Unjustifiable Use of Deadly Force). Hitting a person (Scenario 6: Officer Strikes Arrestee) or failing to stop the beating by a subordinate officer (Scenario 11: Sgt. Fails to Halt Beating of Child Abuser) were not to be protected to the same degree (21.8% not reporting and 14.2% not reporting, respectively) as verbal abuse was (45.7% not reporting). The rank order of these scenarios based on the likelihood of not reporting is strongly and negatively correlated with the evaluations of their seriousness; the more serious the evaluation of the abuse of excessive force, the more likely the respondents were to say that they would not tolerate it without reporting.

Multivariate Models of the Code of Silence

We utilize logistic regression to assess the relation between the respondents’ expressed willingness to report (own reporting) and organizational variables (own seriousness, knowledge of rules, expected discipline, discipline fairness, and others reporting) and individual variables (supervisory status and length of service). The results are shown in Table 3.

Table 3.

Logistic Regression of Own Reporting.^a

	Scenario 1 (gifts) Free Meals, Gifts From Merchants			Scenario 2 (warrant)			Scenario 5 (supervisor)			Scenario 7 (verbal)			Scenario 6 (hurts) Free Meals, Gifts From Merchants			Scenario 8 (DUI)			Scenario 11 (watch)			Scenario 3 (theft) Free Meals, Gifts From Merchants			Scenario 4 (deadly F.)			Scenario 9 (kickback)			Scenario 10 (report)
	B	SE	Odds	B	SE	Odds	B	SE	Odds	B	SE	Odds	B	SE	Odds	B	SE	Odds	B	SE	Odds	B	SE	Odds	B	SE	Odds	B	SE	Odds	B	SE	Odds
Supervisory status^b	.937**	.327	2.553	1.403**	.508	4.068	1.255*	.578	3.509	1.859***	.424	6.416	2.178***	.548	8.826	1.115	.612	3.049	.917	.743	2.501	1.192**	.429	3.294	.830	.632	2.293	1.369***	.365	3.933	1.447***	.384	4.249
Length of service^c
6-15 years	.438	.471	1.549	−.119	.559	.887	.336	.576	1.400	−.530	.588	.589	−.204	.659	.815	1.439*	.705	4.216	−.705	.812	.494	−.050	.520	.951	−.736	.863	.479	.243	.497	1.275	.560	.544	1.751
Over 15 years	.735	.487	2.085	.673	.590	1.961	.134	.587	1.143	−.130	.604	.878	−.579	.673	.560	1.926*	.808	6.861	.337	.857	1.401	.586	.533	1.798	−.143	.860	.866	1.011*	.511	2.748	.459	.553	1.582
Own seriousness^d	1.213***	.296	3.362	1.576**	.493	4.838	1.399**	.459	4.050	1.987***	.371	7.297	1.668**	.490	5.303	2.329***	.471	10.268	1.989**	.660	7.310				3.882**	1.406	48.502
Knowledge of rules^e	.399	.297	1.490	−.080	.445	.923	−.426	.432	.653	.709	.420	2.032	.662	.588	1.939	−.480	.493	.619	.836	.715	2.308				.887	.804	2.427	1.273*	.588	3.570		.
Expected discipline^f
Some discipline	.967**	.383	2.631	1.318**	.475	3.735	2.008***	.514	7.446	.539	.411	1.715	.153	.741	1.166	2.114**	.645	8.278	−.109	.841	.896	1.822	1.346	6.184	−1.456	1.147	.233	2.338**	.883	10.363	1.795*	.821	6.021
Dismissal													.748	1.090	2.112	1.557	1.086	4.746	1.447	1.192	4.251	3.789**	1.378	44.192	.548	1.278	1.729	4.562***	.980	95.775	3.463***	.869	31.921
Discipline fairness^g
Too harsh	−1.541**	.446	.214	−1.443*	.580	.236	−1.000	.847	.368	−1.442**	.470	.236	−1.984***	.495	.138	−1.269*	.511	.281	−1.497*	.618	.224	−1.408*	.632	.245	−.587	.809	.556	−1.403**	.503	.246	−1.722**	.583	.179
Too lenient	.856*	.380	2.353	1.238*	.547	3.448	2.006***	.556	7.435	−.522	1.117	.593	.932	.798	2.539	2.968**	.915	19.458	.885	.827	2.423	1.734***	.454	5.664	2.317*	.903	10.144	1.446**	.431	4.246	1.806**	.536	6.083
Others’ reporting^h	3.837***	.363	46.381	4.162***	.403	64.223	4.791***	.470	120.395	4.717***	.435	111.810	4.566***	.466	96.162	5.335***	.528	207.501	4.916***	.571	136.502	4.799***	.747	121.434	6.732***	1.141	839.224	5.264***	.790	193.214	4.723***	.634	112.521
Constant	−3.350	.500	.035	−3.746	.706	.024	−3.896	.689	.020	−3.422	.597	.033	−2.869	.859	.057	−6.483	1.041	.002	−3.358	1.024	.035	−3.182	1.450	.042	−4.586	1.559	.010	−5.343	1.048	.005	−3.382	.952	.034
Model Nagelkerke R²	.678			.735			.770			.807			.772			.868			.769			.588			.790			.719			.628

Note. DUI = driving under the influence.

p < .05. **p < .01. ***p < .001.

Reference category = “not reporting.”

Reference category = “non-supervisor.”

Reference category = “0 to 5 years.”

Reference category = “not serious.”

Reference category = “not.”

Reference category = “no discipline/verbal reprimand.”

Reference category = “fair.”

Reference category = “would not report.”

The strongest predictor of the respondents’ unwillingness to report is their assessment that other officers in the agency would not be willing to report (others reporting) (Table 3); in all 11 scenarios, the respondents who said that other police officers would report had higher odds of saying that they would report too than the respondents who said that other police officers would not report (Table 3). In other words, the respondents’ assessment of what their peers would do is the most powerful influence on their perception regarding whether they should stick to the code of silence.

Evaluations of seriousness were also strongly and negatively related to the code of silence; in all eight scenarios in which own seriousness was a part of the logistic model, the respondents who evaluated the behavior as serious had much higher odds of saying that they would report the behavior than the respondents who did not evaluate the behavior as serious (Table 3).

Knowledge of rules was a statistically significant predictor in only one scenario (Scenario 9: Auto-Body Shop 5% Kickback; Table 3). Because of its moderate to strong correlation with own seriousness (Spearman’s correlation coefficient ranged from .341 to .589, depending on the scenario), we also run a separate logistic regression with knowledge of rules included, but own seriousness excluded. When own seriousness was excluded from the model, knowledge of rules was significant in about one-half of the scenarios (five out of nine scenarios in which it could be included in the model).⁸ Compared with the respondents who did not evaluate the behavior as a violation of official rules, the respondents who evaluated the behavior as a violation of official rules had much higher odds of saying that they would report.

Expected discipline was a statistically significant predictor in about one-half of the scenarios. In particular, in six out of 11 scenarios, compared with the respondents who expected either no discipline or a verbal reprimand, the respondents who expected some harsher discipline but not dismissal had much higher odds of saying that they would report (Table 3). In addition, compared with the respondents who expected no discipline or a verbal reprimand, the respondents who expected dismissal had higher odds of saying that they would report in three out of seven scenarios (see Table 3).

As it turns out, not only was expected discipline relevant for the code of silence, but the evaluation of the expected discipline fairness (discipline fairness) turned out to be even more important (Table 3). Compared with the respondents who evaluated the expected discipline as fair, the respondents who thought that the expected discipline was too harsh had much lower odds of saying that they would report in the majority of scenarios (nine out of 11). On the other hand, compared with the respondents who evaluated the expected discipline as fair, the respondents who thought that the expected discipline was too lenient had higher odds of saying that they would report in the majority of the scenarios (eight out of 11; Table 3).

Length of service was not a significant predictor of the respondents’ expressed adherence to the code of silence (Table 3). The respondents’ length of service was relevant in only two scenarios (Scenario 8: Cover-Up of Police DUI and Accident; Scenario 9: Auto-Body Shop 5% Kicback); in both scenarios, the more experienced respondents had higher odds of saying that they would report than the respondents who were less experienced (Table 3). By contrast, supervisory position (supervisory status) was significant in the majority of the scenario (eight out of 11; Table 3). In particular, compared with non-supervisors, supervisors have a much higher odds of saying that they would report misconduct.

Conclusion

The control of the code of silence is one of the critical aspects of developing and maintaining a police agency of high integrity. In accordance with extant research (e.g., Hickman, Piquero, et al., 2016; Hickman, Powell, et al., 2016; Klockars et al., 2004, 2006, 1997; Kutnjak Ivković et al., 2013; Kutnjak Ivković et al., 2016; Kutnjak Ivković & Sauerman, 2013; Lim & Sloan, 2016; Long et al., 2013; Porter & Prenzler, 2016; Westmarland, 2006), our results show that the code is far from a simple prohibition of reporting; police officers do not adhere to the code of silence uniformly. In particular, we were able to detect and describe empirically large differences in the contours of the code of silence.

The study demonstrates that the police officers’ expected response varies with how serious they evaluate misconduct. When they were faced with less serious misconduct, such as verbal abuse, the acceptance of small gifts, or police officers’ DUI, most police officers in our sample seemed to be willing to adhere to the code of silence and tolerate those behaviors in silence. On the other hand, when confronted with very serious examples of police misconduct, such as stealing from a crime scene, abusing deadly force, or falsifying an official report, only a small percentage of our respondents indicated their willingness to stick with the code of silence.

Our in-depth exploration of the respondents’ willingness to report emphasizes the critical role of the perceptions of the code of silence in shaping their own responses. When the respondents, both supervisors and line officers alike, perceived that the majority of police officers in the agency would abide by the code of silence, they were more likely to indicate their willingness do so themselves. Lim and Sloan (2016) found that perceptions of deviant climate (containing perceptions of how likely others would be to report and perceptions of others’ evaluations of seriousness) directly influenced the supervisors’ likelihood of reporting misconduct. Long and colleagues (2013) relied on the legitimacy theory and the related endorsement process to assess the link between the likelihood of own reporting and perceptions of others’ reporting. Indeed, they found that, the greater the endorsement of reporting misconduct among others, the more likely the respondents were to report misconduct themselves.

The models not only provide support to the relevance of the perceived severity of expected discipline, but also provide even stronger support for the relevance of expected discipline fairness. While the severity of expected discipline had an independent effect in about one-half of the scenarios, perceptions of fairness were directly related to the respondents’ expressed willingness to report in almost all scenarios. Regardless of all other factors, the respondents who evaluated expected discipline as too harsh were more likely to adhere to the code of silence than the respondents who evaluated discipline as fair. These results fit well the simple justice model (Kutnjak Ivković & Klockars, 1998), suggesting that the perceptions of expected discipline are critical for the code of silence and providing additional evidence in support of this model (e.g., Kutnjak Ivković & Klockars, 1998; Kutnjak Ivković & Sauerman, 2013; Kutnjak Ivković & Shelley, 2010; Long et al., 2013).

Our complex models showed that experience in policing per se did not have an independent effect on the respondents’ expressed willingness to report, whereas the supervisory position did. As expected, in the majority of the scenarios, line officers were more likely to follow the code of silence than supervisors. While prior literature (Hickman, Powell, et al., 2016; Huon et al., 1995; Klockars et al., 2006; Kremer, 2000; Kutnjak Ivković & Klockars, 1998; Kutnjak Ivković et al., 2016; Kutnjak Ivković and Shelley, 2008; Lobnikar et al., 2016; Long et al., 2013; Pagon & Lobnikar, 2000; Porter & Prenzler, 2016; Weisburd & Greenspan, 2000) did not yield a uniform picture of the influence of supervisory position, our results show that supervisory status mattered for their willingness to report. By their position, supervisors have the responsibility to report and process misconduct of their subordinates. Consequently, they are almost mechanically more likely to say that they would report.

Although the characteristics of our sample resemble the distribution of municipal police agencies across the country, it is not a representative sample, thus limiting the extent to which findings may be generalized. Nonetheless, our description of the theory and the ensuing empirical methodology can be used by police administrators, oversight agencies, and scholars to assess the code of silence in police agencies.

First, the code could be measured empirically to assess its contours and the strength of its support. Studies can pinpoint where the code is the weakest and the easiest to break. Our results demonstrate that the code is the least likely to provide protection for the most extreme cases of corruption and use of excessive force.

Second, our analyses point that peer endorsement is the most critical factor influencing the police officers’ decision to report. Our respondents, just like other police officers in other agencies, do not really know what the contours of the code are, but are grounding their decisions on their perceptions of the code of silence. Such a state of affairs is good news for the police administrators and oversight agencies because it leaves a lot of room for the police administrators and oversight agencies to intervene and, while trying to curtail the code of silence, communicate to the officers how the code looks like and how it should look. If the officers knew with a high degree of confidence that other officers in their agency would report a police officer who engaged in the specific behavior, they would be more likely to report it as well.

Third, a uniform finding across prior studies and this study is that officers’ expressed willingness to report is strongly related to their evaluations of seriousness. Thus, one of the strategies of curtailing the code is to influence the gravity with which serious misconduct is viewed by the rank and file. This could be achieved by developing more detailed official rules, teaching about the importance of official rules, and enforcing them consistently.

Fourth, our results demonstrate that, for the purposes of addressing the code of silence, even more important than the harshness of discipline is the perceived fairness of the discipline. Officers are more likely to stick to the code of silence if they evaluate the expected discipline as too harsh than if they evaluate it as fair. Simply put, our results do not provide justification for the introduction of unnecessarily harsh discipline as a way of controlling the code. Instead, we believe that learning whether the police officers have accurate expectations about the discipline that the police agency is to mete out for such misconduct is particularly beneficial. If they do not have accurate expectations, the first step would be to correct their views and teach them about the actual outcomes. Once they do have an accurate picture in mind, the next role of the police administrators and oversight agencies would be to figure out why the officers perceive the discipline as too harsh and, subsequently, to rectify the issue directly, as appropriate.

The proposed theory of police integrity and the related methodology enable police administrators and oversight agencies to assess empirically the contours of the code and gain insight knowledge about the factors related to it. Once police administrators and oversight agencies have this specific knowledge, addressing the code of silence should be an easier and more manageable task. The ultimate reality, however, is that the code will never cease to exist, but it is quite possible to curtail it using the proposed methodology.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Sanja Kutnjak Ivković is a professor at the School of Criminal Justice, Michigan State University. She holds a doctorate in criminology (PhD, University of Delaware) and a doctorate in law (SJD, Harvard University). Her research focuses on police accountability, procedural justice, and legimacy, as well as international and comparative criminology and criminal justice.

Maria (Maki) Haberfeld is a former Lieutenant with the Israel National Police and a professor of Police Science at the John Jay College of Criminal Justice. She has been involved in research projects on police ethics, training and counter-terrorism. Her publications include books, book chapters and peer reviewed articles on police leadership, training, multicultural awareness and counter-terrorism response.

Robert Peacock is a doctoral candidate in the School of Criminal Justice, Michigan State University, and currently manages the US Department of Justice ICITAP Regional Program Office in Kyiv, Ukraine. He holds an MBA in Finance from Ohio University and a BA in Public Policy from the University of Michigan. Mr. Peacock’s research focuses on law enforcement personnel performance and accountability.

References

Commission to Investigate Allegations of Police Corruption and the City’s Anti-Corruption Procedures [Knapp Commission]. (1972). Report on police corruption. New York, NY: G. Braziller.

Greene

J.R.

Piquero

A.R.

Hickman

M.J.

Lawton

B.A.

(2004). Police Integrity and Accountability in Philadelphia: Predicting and Assessing Police Misconduct. National Institute of Justice. Retrieved from http://www.ncjrs.gov/pdffiles1/nij/grants/207823.pdf

Hickman

M. J.

Piquero

A. R.

Powell

Z. A.

Greene

(2016). Expanding the measurement of police integrity. Policing: An International Journal of Police Strategies & Management, 39, 246-267.

Hickman

M. J.

Powell

Z. A.

Piquero

A. R.

Greene

(2016). Exploring the viability of attitudes toward ethical behavior scale in understanding police integrity outcomes. Policing: An International Journal of Police Strategies & Management, 39, 319-337.

Huon

G. F.

Hesketh

B. L.

Frank

M. G.

McConkey

K. M.

McGrath

G. M.

(1995). Perceptions of ethical dilemmas. Payneham, Australia: National Police Research Unit.

Independent Commission on the Los Angeles Police Department [Christopher Commission]. (1991). Report of the independent commission on the Los Angeles Police Department. Los Angeles, CA: Author.

Klockars

C. B.

Kutnjak Ivković

(2004). Measuring police integrity. In Hickman

M. J.

Piquero

A. R.

Greene

J. R.

(Eds), Police integrity and ethics (pp. 1.1-1.20). Belmont, CA: Wadsworth Publishing.

Klockars

C. B.

Kutnjak Ivković

Haberfeld

M. R.

(2004). Police integrity in the United States of America. In Klockars

C. B.

Kutnjak Ivković

Haberfeld

M. R.

(Eds.), The contours of police integrity (pp. 265-282). Thousand Oaks, CA: Sage.

Klockars

C. B.

Kutnjak Ivković

Haberfeld

M. R.

(2006). Enhancing police integrity. New York, NY: Springer.

10.

Klockars

C. B.

Kutnjak Ivković

Harver

W. E.

Haberfeld

M. R.

(1997). The Measurement of police integrity. Final Report Submitted to the U.S. Department of Justice, Office of Justice Programs, National Institute of Justice, Washington, DC.

11.

Klockars

C. B.

Kutnjak Ivković

Harver

W. E.

Haberfeld

M. R.

(2000). The Measurement of police integrity (Research in Brief, U.S. Department of Justice, Office of Justice Programs, National Institute of Justice). Washington, DC: Government Printing Office.

12.

Kremer

(2000). Comparing supervisor and line officer opinions about the code of silence: The case of Hungary. In Pagon

(Ed.), Policing in Central and Eastern Europe: Ethics, integrity, and human rights (pp. 211-219). Ljubljana, Slovenia: College of Police and Security Studies.

13.

Kutnjak Ivković

Haberfeld

M. R.

Peacock

(2013). Rainless West: The integrity survey’s role in agency accountability. Police Quarterly, 16, 148-176.

14.

Kutnjak Ivković

Haberfeld

M. R.

Peacock

(2015). Police integrity in the United States. In Kutnjak Ivković

Haberfeld

M. R.

(Eds.), Police integrity across the world (pp. 295-327). New York, NY: Springer.

15.

Kutnjak Ivković

Klockars

C. B.

(1998). The code of silence and the Croatian police. In Pagon

(Ed.), Policing in Central and Eastern Europe: Organizational, managerial, and human resource aspects (pp. 329-347). Ljubljana, Slovenia: College of Police and Security Studies.

16.

Kutnjak Ivković

Klockars

C. B.

(2000). Comparing police supervisor and line officer opinions about the code of silence: The case of Croatia. In Pagon

(Ed.), Policing in Central and Eastern Europe: Ethics, integrity, and human rights (pp. 183-195). Ljubljana, Slovenia: College of Police and Security Studies.

17.

Kutnjak Ivković

Peacock

Haberfeld

M. R.

(2016). Does discipline fairness matter for the code of silence? Answers from the U.S. supervisors and line officers. Policing: An International Journal of Police Strategies & Management, 39, 354-369.

18.

Kutnjak Ivković

Sauerman

(2013). Curtailing the code of silence among the South African police. Policing: An International Journal of Police Strategies & Management, 36, 175-198.

19.

Kutnjak Ivković

Shelley

(2005). The Bosnian police and police integrity: A continuing story. European Journal of Criminology, 2, 428-454.

20.

Kutnjak Ivković

Shelley

(2008). The police code of silence and different paths towards democratic policing. Policing & Society, 18, 445-473.

21.

Kutnjak Ivković

Shelley

(2010). The code of silence and disciplinary fairness: A comparison of Czech police supervisor and line officer views. Policing: An International Journal of Police Strategies & Management, 33, 548-574.

22.

Lim

Sloan

J. J.

(2016). Police officer integrity: A partial replication and extension. Policing: An International Journal of Police Strategies & Management, 39, 284-301.

23.

Lobnikar

Prislan

Cuvan

Mesko

(2016). The code of silence and female police officers in Slovenia: Gender differences in willingness to report police misconduct. Policing: An International Journal of Police Strategies & Management, 39, 387-400.

24.

Long

Cross

J. E.

Shelley

T. O.

Kutnjak Ivković

(2013). The normative order of reporting police misconduct: Examining the roles of offense seriousness, legitimacy, and fairness. Social Psychology Quarterly, 76, 242-267.

25.

Marche

G. E.

(2009). Integrity, culture, and scale: An empirical test of the big bad police agency. Crime, Law, and Social Change, 51, 463-486.

26.

McDevitt

Posick

Zschoche

Rosenbaum

D. P.

Buslik

Fridell

(2011). Police integrity, responsibility, and discipline. National Institute of Justice. Retrieved from http://www.nationalpoliceresearch.org/storage/updatedpapers/Police%20Integrity%20Responsibility%20and%20Discipline.pdf http://uicclj.squarespace.com/police-integrity-responsibili/

27.

Micucci

A. J.

Gomme

I. M.

(2005). American police and subcultural support for the use of excessive force. Journal of Criminal Justice, 33, 487-500.

28.

Neumeister

Armstrong

(2016, June 20). NYC police officials charged in bribery corruption scandal. New York Times. Retrieved from http://www.nytimes.com/aponline/2016/06/20/us/ap-us-nypd-corruption-probe.html

29.

New York City Commission to Investigate Allegations of Police Corruption and the Anti-Corruption Procedures of the Police Department [Mollen Commission]. (1994). Commission report. New York, NY: Author.

30.

Pagon

Lobnikar

(2000). Comparing supervisor and line officer opinions about the code of silence: The case of Slovenia. In Pagon

(Ed.), Policing in Central and Eastern Europe: Ethics, integrity, and human rights (pp. 197-209). Ljubljana, Slovenia: College of Police and Security Studies.

31.

Porter

L. E.

Prenzler

(2016). The code of silence and ethical perceptions: Exploring police officer unwillingness to report misconduct. Policing: An International Journal of Police Strategies & Management, 39, 370-386.

32.

Reaves

B.A.

(2010). Local Police Departments. Washington, D.C.: Bureau of Justice Statistics. Retrieved from http://bjs.ojp.usdoj.gov/content/pub/pdf/lpd07.pdf.

33.

Rothwell

G. R.

Baldwin

J. N.

(2007a). Ethical climate theory, whistle-blowing, and the code of silence in police agencies in the state of Georgia. Journal of Business Ethics, 70, 341-361.

34.

Rothwell

G. R.

Baldwin

J. N.

(2007b). Whistle-blowing and the code of silence in police agencies policy and structural predictors. Crime & Delinquency, 53, 605-632.

35.

Shih

Fan

(2008). Comparing response rates from web and mail surveys: A meta-analysis. Field Methods, 20, 249-271.

36.

Schafer

J. A.

Martinelli

T. J.

(2008). First-line supervisor’s perceptions of police integrity. Policing: An international journal of police strategies & management, 31, 306-323.

37.

Schramm

Cohen

Fredericks

(2016, May 13). Cop questions in NYPD probe kills himself. New York Post. Retrieved from http://nypost.com/2016/05/13/cop-questioned-in-nypd-corruption-probe-kills-himself/

38.

Vasilogambros

(2016, June 21). The NYPD’s corruption scandal. The Atlantic. Retrieved from http://www.theatlantic.com/news/archive/2016/06/nypd-corruption-scandal/488054/

39.

Weisburd

Greenspan

(2000). Police attitudes toward abuse of authority: Findings from a national study. Washington, DC: Office of Justice Programs.

40.

Westmarland

(2006). Police ethics and integrity: Breaking the blue code of silence. Policing & Society, 15, 145-165.

41.

Wolfe

S. E.

Piquero

A. R.

(2011). Organizational justice and police misconduct. Criminal justice & behavior, 38, 332-353.