Sergeant Coffee Needs You: Evaluation of a Police Officer Bystander Intervention Program

Abstract

Police encounters with suspects may produce unfortunate outcomes, particularly when defiance of authority leads to usage of force. Application of excessive force may be interrupted, or obviated by the actions of bystander officers who become intervening officers. This study evaluated officer and supervisor perceptions of a training program which taught intervention strategies and tactics to help officers intervene with other officers to prevent harm. Data were collected very early in program implementation and approximately one year later. Survey data were collected with officers and qualitative focus groups were conducted with officers and supervisors. Findings indicate a marked change in perceptions by officers. Both quantitative and qualitative findings collected in early phase indicate a high degree of officer resistance. Late phase data indicate significantly higher endorsement of officer intervention strategies and participants reported higher frequencies of intervening with other officers. Findings suggest officer-based de-escalation and intervention training warrants implementation consideration from agency decision makers.

Keywords

police use of force policing de-escalation intervention active bystander culture change

Introduction

When Minneapolis Police Officer Derek Chauvin killed George Floyd, three other officers stood nearby and watched or provided crowd control. One officer twice asked Chauvin whether Floyd should be turned on his side but none of the officers present intervened in a manner that altered the excessive force administered by Chauvin. George Floyd died, Chauvin was tried and convicted in a federal court, and the three bystander officers were convicted of federal charges (for depriving Floyd of his rights). Two of the officers were also convicted of failing to intervene during the killing.

When Tyre Nichols was killed by Memphis Police officers, five officers were involved in beating him and were indicted with a host of charges, including second degree murder and kidnapping. Two bystander officers were suspended and may be charged.

When Rodney King was beaten by Los Angeles Police Department officers, approximately two dozen law enforcement personnel are visible on the video, including one supervisor. Several LAPD officers, Los Angeles County Sheriff’s Deputies, and California Highway Patrol personnel were present. All were bystanders while Rodney King was kicked and beaten. The bystander sergeant on scene (Stacey Koon) was charged concurrently with the officers who beat King.

These examples represent some of the worst moments for policing over the past few decades. Each incident followed apprehension of a suspect who committed, or was suspected, of minor offenses. Rodney King was speeding, then evaded police in a high-speed pursuit. Tyre Nichols was pulled over for a traffic offense, then ran from police. George Floyd was accused (by a cashier) of using a counterfeit twenty-dollar bill and then resisted arrest.

An important commonality is the presence of additional officers who failed to intervene as one or several officers utilized force on a suspect. These bystander officers failed in their duties, obviating their responsibility to protect life, in deference to another officer who was administering excessive force against an unarmed suspect. These egregious failures, and other similar incidents, have damaged public perception and trust of law enforcement.

The New Orleans Police Department developed the Ethical Policing is Courageous (EPIC) program in response to several incidents, including some regarding excessive force. While several police departments had developed programs that encouraged officers to intervene in situations which were unethical, EPIC was notable in that NOPD officers were trained to intervene in problematic situations. These situations could include excessive force, corruption, coercive behavior, or any other inappropriate circumstance. The EPIC program addressed three pillars: reducing mistakes, preventing misconduct, and promoting officer health and wellness.

The implementation of EPIC included a training extension, the Active Bystander Law Enforcement (ABLE) program, developed by a team of law enforcement officers, civil rights lawyers, psychologists, academics, and curricula design specialists (Aronie & Lopez, 2017). ABLE is a police training model which has been implemented at more than 350 law enforcement agencies. One component of the training teaches officers to intervene and de-escalate when other officers are engaging in scenarios where force is the likely result, or force is in progress but could lead to injury or death. This distinction (between training and encouraging) is an important one as officers must be provided a skillset to implement change. There have been few evaluations of training models for law enforcement bystander intervention. Findings are generally supportive of this effort, as described in the literature review.

This study utilized a mixed methods approach, incorporating both survey and qualitative data, to secure officer and administrator perspectives, at multiple points in time, on a police bystander intervention program. Findings may inform police agency training and policy.

Literature Review

A host of factors may influence an officer’s interaction with citizens. Police culture and organizational factors play a role, the urgency of the situation, exigent circumstances, officer personality, suspect behavior, civilian presence, and any number of other variables can influence the officer/suspect dynamic (Farrell, 2022; Hine et al., 2018; James et al., 2018; Klinger, 1994). Officer tenor, particularly related to aggressive tone, predicts escalated force with suspects (Terrill et al., 2023). Critical among these influencing factors is officer training, particularly relative to de-escalation (Engel et al., 2020, 2022; Todak & James, 2018). The following literature review addresses the influence of policing culture, factors that influence police use of force, the evolution of an officer training program, and the nature of bystander intervention. The extant literature leads to an evident knowledge gap and research questions intended to rectify this deficit.

Use of Force and Police Culture

Many variables influence police use of force. Fyfe’s seminal analysis of deadly force in New York framed an understanding of the variables that can precipitate officer use of force (1978). Fyfe found that environment and context were important, as were officer specific variables such as officer race and assignment. Organizational leadership plays a critical role. Skolnick and Fyfe (1993) provide case studies of Philadelphia (during the 1970s) and Los Angeles (during the 1990s) and cite the endorsement by the chiefs of those agencies towards officer use of force. Through a typology of police, Worden (1995) defines the individual perspective of officers inclined to use force, noting characteristics such as their narrow view of the police role, conceptualization of force as a tool to neutralize, and regard of the citizenry as unappreciative and hostile. There are few external controls on police use of force. Section 1983 civil lawsuits require substantial evidence to hold officers and agencies accountable. Department of Justice investigations and the subsequent consent decrees can lead to agency change but those changes take many years and may not be lasting. Consent decrees are political—during the Obama administration there were fourteen new consent decrees while the Trump administration ended several and backtracked on others. Internal influences of use of force are largely voluntary. An agency that wants to pursue accreditation must have lethal and less than lethal force policies commensurate with best practice. At the other end of the policy continuum, however, are those agencies with no use of force policy and are instead reliant upon basic state and federal laws (Alpert & MacDonald, 2001).

As with many hierarchical organizations, the police bystander problem is rooted in organizational culture. Police need the support of their peers—hesitation or reluctance in providing officer assistance can place an officer in danger. There are several key elements of police organizational culture that are related to the passive officer bystander problem.

Research on the culture of policing is replete with examples of authors connecting suspect defiance to negative outcomes. Paul Chevigny (1969) observed that defiance of police authority increases the likelihood of police resorting to force. Donald Black (1980) concurred, stating that disrespectful behavior by a suspect frequently leads to police use of force. Jerome Skolnick (1994) wrote eloquently about the two policing imperatives—danger and authority—and how they define the modalities in which police interact with suspects. When a suspect defies officer authority, it can be viewed as a dangerous act and therefore necessitates a quick and decisive response. Van Maanen (1978) echoes this theme, describing the three types of civilians: suspicious persons, know-nothings, and the asshole, where an asshole is anyone who defies police authority.

The connection between suspect defiance and police use of force is critical, particularly when coupled with an important internal police cultural code. Police are expected to not go against other police. Following his immersive research experience in policing, Rubenstein (1973) observed that police culture mandates unquestioning internal support of fellow officers. Elizabeth Reuss-Ianni (1978) concurred, stating that police who inform against officers (even corrupt or abusive police) are ostracized by their peers and run the risk of not having back-up arrive when needed. Alpert et al. (2021) expanded these concepts detailing the many factors that necessitate officer solidarity, supporting the code of silence despite observed officer malfeasance.

EPIC Program

Many aspects of the policing profession are hard. Learning the craft of policing as a rookie officer is particularly challenging. Most agencies assign rookies to a field training officer (FTO) who imparts the benefits of their experience with the new officer. In 2005, New Orleans Police Officer Dean Moore, a rookie, watched his FTO, veteran NOPD officer Melvin Williams, employ excessive force. The victim, Raymond Robair, was beaten and kicked by Officer Williams, likely as punishment or a show of power, leading to hospitalization and eventual death. Williams was charged and convicted. Interestingly, rookie Officer Moore also remained quiet at the hospital—had he provided information as to the nature of the beating, it is possible that Robair could have been saved.

The Robair incident was one of several problems that contributed to a call for an external investigation by the Department of Justice. That investigation led to a federal Consent Decree which determined that all officers should receive ethics training and be trained in peer intervention (NOPD Consent Decree, 2012).

After a leadership change, the New Orleans Police Department created the Ethical Policing is Courageous (EPIC) program which teaches officers the strategies and tactics they need to intervene in another officer’s conduct safely and effectively. While these situations might involve force, other actions could lead to officer intervention, including corruption, coercive behavior, transactional relationships, or any violation of police ethics. Encouraging and supporting officer intervention may reduce the insular role of police culture.

The Superintendent of the New Orleans Police Department then constructed a working group of patrol officers, supervisors, commanders, and community members (including a civil rights attorney, a community leader, a mental health worker, and similar) and tasked them with building a holistic program to support NOPD officers. The working group addressed a series of preconceived notions about the program (noted by Aronie and Lopez as Myths), addressing whether EPIC was necessary, duplicative, or designed to coerce officers to inform on other officers. The working group eventually recommended full implementation. Aronie and Lopez (2017) present a series of measures of effectiveness which could determine whether EPIC generated meaningful change with the NOPD. These include: reduced use of excessive force, fewer citizen complaints, and increased reporting of intervention. The authors note these are highly elusive measures and may not be determinable given the multiple confounds that influence each of these outcome categories.

While ethics training is usually included in Basic Law Enforcement Training (BLET), minimal hours are allocated in the training schedule, with only officer mental welfare receiving fewer hours (Raines & Merenda, 2022). Ethics training generally directs academy recruits towards what is right and wrong—EPIC training is different in that problem-based learning is a core component (Raines & Merenda, 2022) and it focuses on skills and tactics. Recruits are asked to reflect and talk through decision making, leading to a common understanding of the rationale behind ethical decision making. Raines and Merenda (2022) compared standard academy ethics training with EPIC training in a population of police academy cadets. They included an orientation towards bystandership and officer intervention as part of their review (Raines & Merenda, 2022). The authors collected survey data from multiple cohorts of academy students—some undergoing traditional BLET with standard ethics training and some that completed training with EPIC principles. When questioned about the likelihood of intervening with officers who were using excessive force, those officers that completed EPIC training reported a willingness to intervene at a statistically higher rate (Raines & Merenda, 2022). This finding is consistent with the core premises of EPIC training and indicates the potential of this approach.

Other studies have included questions on officer bystander consideration, usually as an adjunct to ethical decision-making by police. Fischer et al. (2006) determined that officers who felt that a situation was low danger would not be likely to intervene in officer interaction with a suspect. Hickman et al. (2016) found that the likelihood of officer intervention was proportional to the degree of perceived misconduct. While there is some understanding of officer perceptions of police intervention programs, direct evaluation of these programs is lacking.

Bystander Intervention

One of the founders of the EPIC program, Dr. Ervin Staub (known for research on the psychology of peace and violence), worked with the Georgetown Law Center to advance a training regimen for police. The Active Bystander Law Enforcement (ABLE) program provides trainers and a curriculum to law enforcement agencies. The core premise of the training is to foster officer intervention and de-escalation tactics. Facilitating officer intervention prior to excessive force could have obviated the death of George Floyd and other losses where bystander officers failed to prevent a tragedy.

Bystander intervention approaches have been employed in several fields and some of those concepts may be applicable to policing. The medical field encourages intervention if team members believe that the individual treating a patient is making the wrong decision or following the wrong intervention path (Stavert & Lott, 2013). The risk of failing to intervene in a crisis treatment situation is extremely high relative to the potential negative effects (Boynton, 2012). Bullying research has also suggested that teaching bystanders to intervene, particularly in cyberbullying scenarios, can defuse situations before they escalate into violence (Hinduja & Patchin, 2009; Pelfrey & Weber, 2015). Schools have been encouraged to advise students that deflecting a bullying scenario can reduce the impact on a victim, reduce the risk of violence, and limit the likelihood of retaliation tactics (Pelfrey & Weber, 2015).

The President’s Task Force on 21^st Century Policing noted the importance of peer intervention and encouraged wide advocacy as a means to prevent excessive force (President’s Task Force on 21st Century Policing, 2015). Many agencies already have duty-to-intervene policies as part of their use of force policies (Taniguchi et al., 2022). Unfortunately, these policies may not be consistently taught or supported through scenario-based training (Taniguchi et al., 2022). Aronie and Lopez (2017) note that officers who fail to intervene and prevent an officer from violating a person’s constitutional rights may be held liable under 42 U.S.C. § 1983. While this liability has existed for many years, it also appears insufficient to foster an officer bystander intervention imperative among police agencies.

Research Questions

When police decide to use force, it should be proportional and the minimum amount of force required to fulfill a defined, legitimate objective (Skolnick & Fyfe, 1993). Unfortunately, high profile excessive force incidents, with other officers in a bystander role, have adversely impacted policing relative to community trust. The EPIC and ABLE programs seek to encourage officer intervention in escalating situations. Although there is limited scholarly research on these initiatives, the extant literature suggests these approaches have merit. A knowledge deficit exists regarding officer level perceptions of these efforts. This gap suggests a need for research addressing police perceptions of officer intervention programs, particularly relative to usage and expected utility. Research questions which extend from the literature review include:

1) Do police officers perceive a need for an officer intervention training program?

2) How will officer views of officer intervention program evolve over time?

The following research methods were used to collect data which directly addresses those questions.

Methodology

To address research questions regarding the utility of an officer intervention training program in an urban law enforcement agency, a multi-methods research strategy was developed and implemented. The following sections detail the research setting, study design, quantitative and qualitative data collection instrumentation and approaches, and key methodological steps.

Research Setting

This research was conducted with an urban law enforcement agency in the mid-Atlantic region. The agency has approximately 750 sworn personnel but is allocated an additional 80—at the time of the study they were short approximately 10% of funded sworn personnel. The residential population of the city served by this agency is approximately 230,000 in a metropolitan statistical area of approximately 1.2 million. The city is approximately 40% African-American, 40% white, 10% Latino/Hispanic, and 10% other/multi-racial. While this city has had serious violent crime issues in the past, and a number of recent protests, the population has stayed relatively constant. The police agency is fully accredited with a major national law enforcement accrediting agency, indicating that policies and training modalities are consistent with national best practices. The police agency is divided into four approximately equivalent police precincts. This project worked with officers assigned to two of those four precincts. Each precinct houses approximately one-hundred sworn officers.

Design

A multi-methods approach was utilized to provide the most complete participant perspectives. The intersection of quantitative and qualitative methods allows researchers to triangulate findings (Yin, 2017). A pre-test/post-test model was employed to secure perspectives very early in program implementation and post initiative implementation. All sworn patrol officers in the two precincts were invited to participate in a pre-implementation survey and a post-implementation survey conducted approximately one year later. Focus groups with officers from each precinct were conducted early in implementation and approximately one year later. Officers were selected for participation in the focus groups via a non-random, convenience approach. Several officers representing multiple shifts were asked to stay after a shift or come in slightly early to participate. Additionally, interviews with senior command personnel were conducted pre and post implementation.

Initiative

Following the death of George Floyd, the participating city and police agency experienced several protests and citizen/police conflicts. Police leadership elected to implement the ABLE program and directed all sworn officers to complete officer intervention training. The capacity of the police academy and a finite number of ABLE instructors represented a limiting factor relative to implementation rollout. Within a few months, all sworn officers completed training during a phased approach. All incoming officers complete the officer intervention training program. At the time of this study, approximately 99% of sworn officers had completed bystander intervention training and it has become a regular piece of new officer training at the academy.

A core tactic of officer de-escalation is to tell an officer involved in a potentially negative suspect encounter that a fictional administrator wants to talk to that officer. The statement “Sergeant Coffee wants to see you” is a code phrase. This statement should inform one officer that their colleague would like to take over the encounter and the principal officer should step away and cool down. Usage of a code phrase means the suspect is not privy to any concern over potential use of force. The “Sergeant Coffee” approach is one of several tactics that officers learn. Other approaches include asking the officer to step away or giving them a different task. The three D’s of the program include: Distract, Delegate, and Direct.

Quantitative Data Collection Modality

Surveys were conducted with all patrol officers in two of the four precincts in the city. The first wave survey addressed constructs including perceptions of officer bystander training, need for bystander intervention, and likelihood of intervention. A second wave survey addressed the same constructs from a retrospective view. All surveys were conducted electronically and anonymously. Officers were emailed a link from the researcher and the results of the survey went to the researcher (not the agency). Survey data were downloaded, reviewed, then analyzed. No incomplete surveys were collected so no participant submissions required rejection. A small number of participants declined to provide demographic data and selected a “Choose not to respond” option. Only aggregated findings were returned to the agency in the form of a summary report. Although each precinct has approximately one hundred sworn officers assigned but, at any given time, some are not available for service (due to injury, time-off, or temporary assignment). This produced a population of potential participants of 192 (across two precincts) during early phase data collection and 186 at late phase data collection. With 77 respondents to the early phase survey and 75 respondents for the late phase survey, the response rate at each phase was just over 40%.

Qualitative Data Collection Modality

Qualitative data were collected from both officers and supervisors via multiple techniques. During the first data collection phase, focus groups were conducted with randomly selected officers from each of the two precincts. Questions addressed the need for officer intervention in escalating situations, variables that could influence bystander intervention, and the likelihood of officer intervention. Officers were ensured anonymity—no names or identifying information were collected. In the post implementation phase, officers were asked similar questions. The officers in the second data collection focus group (approximately one year later) were not the same officers in the initial focus groups—these were convenience samples. The focus groups were not recorded as a step to increase the likelihood of honest responding and minimize any concerns over confidentiality. Additionally, the agency does not support recording of data collection efforts. Interviews with officers were preferred and requested but not approved thus focus groups were the only qualitative option with officers.

Interviews with command staff were also conducted at the first and second phase points. Constructs tapped were similar to those raised in the officer focus groups. Additional questions on policy impact, police academy priorities, and training needs were presented. These meetings were not recorded to increase honest responding and conform to agency preferences against recording data collection steps.

Analysis Plan

Quantitative measures including early phase and late phase implementation officer surveys, and qualitative measures which included early phase and late phase implementation officer focus groups, plus pre and post supervisor interviews, produced multiple data sets. Empirical data were analyzed using a statistics package (SPSS). Descriptive statistics were produced to assess officer support of core concepts. Inferential analyses were conducted to determine whether demographic variables predicted officer perceptions. ANOVA was selected as the appropriate inferential statistic since it identifies statistically significant differences across group means. Group comparison analyses were also conducted. The independent samples t test was identified as the appropriate approach to compare early phase group responses compared with late phase group responses. An independent samples t test is ideal for comparing responses derived from two different groups. Multinomial logistical regression was then run to determine whether there were statistically significant different perceptions between early phase and late phase populations.

Qualitative data were immediately transitioned from interviewer notes to an organized narrative then themes and recurring messages were identified. Overlapping messages between the officer level and supervisor level were particularly noteworthy.

Results

Empirical findings were derived from early and late phase surveys of patrol officers. The following sections detail findings. Qualitative findings were drawn from early and late phase focus groups with officers as well as interviews with supervisors. The following section presents the findings.

Quantitative Data Results—Early Phase

Participants were almost perfectly split in their thoughts on whether an officer intervention program was needed. As described in Table 1, just over half (51%) selected Agree or Strongly Agree when asked if their agency would benefit from an officer intervention training initiative. Perceptions of the utility of de-escalation training were more consistent, with 79% selecting Strongly Agree or Agree that officers benefit from de-escalation training and 72% favoring de-escalation training would benefit all police agencies. Nearly half (44%, or 34 participants) stated they had never thought about intervening with another officer in a use of force situation. Similarly, half of respondents (39, or 51%) said they have never intervened.

Table 1.

Early Phase Officer Perceptions.

Question (n = 77)	Strongly agree	Agree	Disagree	Strongly disagree
Agency needs an officer intervention program	7 (9%)	32 (42%)	23 (30%)	14 (18%)
Agency officers’ benefit from de-escalation training.	14 (18%)	46 (60%)	9 (12%)	7 (9%)
Police everywhere need de-escalation training.	13 (17%)	42 (55%)	11 (14%)	10 (13%)
Officer experience influences whether intervene.	16 (21%)	40 (52%)	14 (18%)	6 (8%)
Officer personality influences whether intervene.	9 (12%)	43 (56%)	16 (21%)	9 (12%)
Citizen attitude influences whether officer intervenes	16 (21%)	38 (49%)	17 (22%)	6 (8%)
Citizen resistance influences whether officer intervenes	17 (22%)	40 (52%)	15 (20%)	5 (7%)

Quantitative Data Results—Late Phase

One year after early phase data collection, perceptions of the officer intervention program changed. As presented in Table 2, nearly two thirds (65% or 49 participants) indicated the agency needed an officer intervention program. A substantial majority (85% or 63 participants) felt that the agency benefitted from de-escalation training and a similar number (75% or 56 respondents) felt all agencies would benefit from de-escalation training.

Table 2.

Late Phase Officer Perceptions.

Question (n = 75)	Strongly agree	Agree	Disagree	Strongly disagree
Agency needs an officer intervention program	17 (23%)	32 (43%)	18 (24%)	8 (11%)
Agency officers benefit from de-escalation training.	26 (35%)	37 (49%)	8 (11%)	3 (4%)
Police everywhere need de-escalation training.	13 (17%)	43 (57%)	17 (23%)	2 (3%)
Officer experience influences whether intervene.	30 (40%)	36 (48%)	7 (9%)	2 (3%)
Officer personality influences whether intervene.	15 (20%)	44 (59%)	13 (17%)	2 (3%)
Citizen attitude influences whether officer intervenes	15 (20%)	41 (55%)	14 (19%)	4 (5%)
Citizen resistance influences whether officer intervenes	33 (44%)	27 (36%)	10 (13%)	2 (3%)
When I have intervened the officer cooperates.	25 (33%)	39 (52%)	3 (4%)	4 (5%)

A question was added to the Late Phase survey asking officers about the success of their interventions. Participants were asked if, when they had intervened with officers, if that officer cooperated. A substantial majority (64 participants, or 85%) selected Agree or Strongly Agree, indicating that officer interventions led to cooperation by those officers involved. Only 19% (15 respondents) stated they had not considered intervening with another officer in an escalating situation over the past year. The vast majority, 81% (60 respondents) said they had considered intervening with an officer at least once, with most (52% or 39 respondents) selecting the option of 6–10 incidents. When asked how often they had actually intervened with another officer, the results were similar with 19% stating zero times and the remainder, 81% indicating at least one intervention.

Inferential Demographic Analysis Results

Several inferential analyses were conducted to determine whether demographic variables influenced or predicted responses. Direct comparisons between early phase and late phase populations are not possible as they were not a matched sample. While they were drawn from the same locations, the early phase population may not be identical to the late phase population.

One-way ANOVA was conducted with the early phase population data set to determine if any of the collected demographic variables (sex, race, experience, education) predicted outcome variables. The outcome variables included: Perceived need for an officer intervention program with this agency; Officers in this agency would benefit from a de-escalation training program; Officers everywhere would benefit from a de-escalation training program; How often have you thought about intervening with another officer, and; How often have you intervened with another officer. There were few statistically significant results of the ANOVA analysis, obviating the need for further inferential analysis.

Similar results emerged with the later phase data set. There were occasional spots of statistical significance in both analyses but only one finding that was statistically significant in both the early phase and late phase populations. Participants with an undergraduate or higher degree were significantly more likely to indicate that police in all agencies would benefit from de-escalation training. This was evident in the early phase group (F = 3.112 Significant at the .032 level) and with the late phase group (F = 4.288 Significant at the .042 level). Given the lack of statistically significant ANOVA findings, no further inferential analyses were conducted.

Group Comparison Analysis

To facilitate analysis, a factor to determine support for officer intervention training was derived in each of the two data sets (early phase and late phase). The factors were then compared analytically with an independent variable representing time (Early or Late). The independent samples T Test produced an F score of 3.753, a T score of −2.890 with 150 degrees of freedom, and a two-tailed significance of .004. A significance score of .004 suggests a high degree of statistical significance. This indicates confidence in concluding that officer intervention program support changed significantly from the early phase period to the late phase period. These findings were confirmed via multinomial logistic regression which produced a Chi-Square score of 92.952 at 23 degrees of freedom with a significance score of .001. These findings indicate that officer perceptions changed in a statistically significant fashion between the two data collection periods, with respondents indicating much higher support for the officer intervention initiative. While there was some skepticism expressed in the early phase regarding the utility of officer intervention, participants clearly indicated a change in perception between early phase and late phase ideology.

Qualitative Data Results—Early Phase

Officers voluntarily participated in focus groups to discuss the officer intervention, de-escalation training they had all recently completed. When asked for their opinion of the de-escalation training, they overwhelmingly found it to be helpful and advocated the use, and expansion, of de-escalation training, particularly with less experienced officers. Their appreciation of de-escalation training was singular in focus—they uniformly liked the process of de-escalation work with civilians and suspects. Conversely, several members of the focus groups felt that de-escalation training directed towards officer intervention was neither necessary nor appropriate. Some officers disagreed but the majority of focus group participants indicated that intervening with another officer was an unlikely and problematic endeavor. When asked if officers should ever intervene with another officer, participants generally felt that was not a good idea. One officer observed “I am never, ever going to do it” referencing intervening with another officer. Although participants clearly steered against intervening with other officers, some said they would do so in an extreme case. They raised George Floyd as an example. When asked how many of them would have intervened with Derek Chauvin if they had been on scene, many (but not all) raised a hand.

One element of the officer intervention training advises officers to interrupt an officer who appears to be on a path towards unnecessary, or excessive, force by telling that officer that “Sgt. Coffee wants to see you.” There is no Sgt. Coffee—it is a code phrase telling one officer to let another officer take over the situation. Several officers felt that a code phrase was not necessary. Those participants said they have told fellow officers to let them handle a situation if they felt the suspect or citizen would be more cooperative with a different officer. Other officers stated that a code phrase was helpful, particularly if a junior officer wants a senior officer to step out of an encounter.

Agency supervisors participated in interviews with research personnel. These interviews produced consistent perspectives—de-escalation training should be a key focus of all officer training. Agency supervisors repeatedly indicated that virtually all sworn officers had completed the officer intervention training in a relatively short period of time and all incoming officers would complete the same training. Supervisors noted that there would likely be problems—junior officers would likely struggle to intervene with senior officers. They contend that these cultural obstacles can change over time and the risk of officers failing to intervene during a problematic encounter are substantial.

Qualitative Data Results—Late Phase

Focus groups conducted with officers approximately one year post the early phase focus groups were conducted. Participants were not the same—these were convenience samples based on officer availability. Participants were asked about their perceptions and utilization of the officer intervention, de-escalation training that they received just over one year prior. Officers were largely supportive of the de-escalation training and felt it had high utility with suspects. Most endorsed the utility of the de-escalation training with officers. When asked for a show of hands regarding how many officers had intervened with other officers since the training, a slight majority of participants raised hands. When asked if any had intervened multiple times, only a few raised their hands. One officer stated that they had intervened subtly, by telling a suspect to “Talk to me instead” as the officer felt an engagement by an officer was becoming overly contentious.

Some officers expressed resistance to the concept of intervening with other officers. One veteran patrol officer asked “Who am I to decide that?” in describing reservations about intervening with another officer. Another officer suggested that sometimes an officer might intentionally escalate a situation with a suspect to create stress and throw the suspect out of their comfort zone. The officer described this as “raising the stakes” and agreed that while this tactic carries risk, it can also produce results.

Participants discussed finesse in dealing with argumentative suspects and agreed that de-escalation tactics can be very helpful with suspects and, sometimes, other officers. One noted how quickly a situation can go from a loud engagement to a use of force situation, then other officers get involved and it quickly leads to a “dogpile” with many officers on a suspect. That officer stated that de-escalating with an angry officer can preclude the dogpile.

Agency supervisors also participated in interviews, addressing similar questions as those posed during early phase meetings. Supervisors affirmed their support for wide-scale officer intervention training. They noted that all new officers, including cadets and transfer officers at any rank, would complete officer intervention training. Additional academy personnel completed the “train the trainers” modules so this training can be conducted with multiple academy classes. Supervisors spoke to the importance of officer de-escalation skills with suspects, citizens, and other officers. Although community relations is a concern, protecting the safety and health of both citizens and officers is the primary motivation for agency-wide adoption of officer de-escalation training. Supervisors were adamant that the death of George Floyd was a catalyst for agencies to examine internal practices, particularly officer culture, and foster officer intervention practices.

Conclusions

Officer misconduct, particularly excessive force, represents an endemic problem for police organizations. The death of George Floyd catalyzed attention on the failure of bystander officers. The culture of policing does not facilitate officer intervention. Since officers are dependent on the support of other officers in potential use of force situations, police are reluctant to defy fellow officers (Alpert et al., 2021; Reuss-Ianni, 1983; Rubenstein, 1973). Foundational policing literature demonstrates the important connection between suspect defiance and subsequent use of force (Skolnick, 1994). Following several incidents in New Orleans in which bystander officers failed to intervene in problematic situations, a training program was developed to help officers identify potentially negative police/citizen encounters and intervene with officers who were building towards overly aggressive behavior or unnecessary force. The ABLE program is predicated on teaching officers the strategies and tactics to intervene in another officer’s behavior, regardless of rank, to prevent misconduct (including excessive force), reduce mistakes, and promote officer health and wellness.

A training program to help officers intervene with other officers to de-escalate situations was implemented in an urban police agency and evaluated at multiple times. The following paragraphs summarize results, discuss the implications of this work, describe limitations to this effort, and consider future research paths.

Early phase, officer level, findings were generally concordant. In the early phase, officer survey responses and focus group outcomes indicated a strong reluctance to intervene with another officer. Although respondents strongly supported de-escalation training, applying de-escalation tactics in officer intervention situations officers was not widely endorsed. Participants were reluctant to do so for a variety of reasons, particularly regarding potential negative interaction outcomes and perceived power disparities. Late phase findings reflected a distinct evolution in participant perceptions. Participants were more likely to endorse the concept of intervening and many more officers reported intervening with fellow officers. This suggests that officer bystander training can have important impacts but they may not be instantaneous.

Supervisors and agency decision-makers were consistent in their views during both early phase and late phase interviews. They overwhelmingly endorsed the utility of officer level de-escalation training for usage both with suspects and intervening when an officer is moving towards a problematic interaction or possible excessive force. If training could prevent a George Floyd style incident, that would have substantial benefits, including the obvious protection of citizen welfare plus enhanced police/community trust. Additionally, administrators pointed to obviating the many negative impacts of a tragedy, including improved officer morale, retention, and mental health.

Broad implementation of an officer de-escalation and intervention training, consistent with the ABLE program, should be considered by law enforcement agency decision makers. Results demonstrate that officers see the utility of these programs, appreciate the de-escalation training (for use with both citizens and officers), and the program facilitated intervention with officers in potentially negative interactions. Benefits were reported at both the officer and agency administrator level. While it is impossible to know how many adverse encounters, with potentially serious outcomes, could be precluded by this kind of training, prevention of any tragedy certainly suggests consideration of these efforts is warranted. Officer intervention programs would likely benefit most types of law enforcement agencies—urban, suburban, rural, state, and federal. Any agency which fields officers that are expected to take suspects into custody will likely face occasional suspect resistance. Decreasing the chances of overly aggressive officer and citizen interaction should be viewed as a priority by all agencies with sworn personnel. Situational exigencies exacerbate complex police citizen interactions. When citizens defy officers, or following a pursuit, officers may have difficulty controlling their emotions and remaining objective. Officer intervention training can have a direct effect of facilitating peer intervention in complex situations. There may be secondary effects such as enhanced understanding of the nature of escalating situations and introspection regarding officer and citizen interactions.

Much of the guidance issued to law enforcement agencies are difficult to implement, inordinately expensive, or both. As an example, most law enforcement agencies struggle to recruit, hire, and retain females and minority persons. This is a very difficult and expensive problem to address. Conversely, adding training for officers is both reasonably priced and achievable. The ABLE program is notably free to most agencies. Since training happens at the agency academy, and is conducted by trainers within the agency (who have completed a train-the-trainer program), the financial and logistical outlay is surmountable. A substantial civil verdict against an agency can cost millions, in addition to the concomitant loss of community trust and consequences on officer morale.

There are a number of law enforcement tactics and strategies that have been eliminated, or are now viewed as commonplace, resulting from the evolving nature of the field. Police academies regularly taught choke holds as a standard means of subduing a suspect but academies no longer teach this strategy. Tasers are now standard issue. Vehicle cameras are commonplace. Body worn cameras are widely accepted and will likely be standard in the near future. Crisis Intervention Training is now a regularly taught strategy at many police academies and will likely be the norm in the near future. Many of these tactics and technologies were implemented (or eliminated) due to liability driven decision making. Officer intervention training may represent a strategy that is necessitated by liability-based decision making, or be implemented pursuant to court demands (as with a consent decree). Ideally, police agencies will voluntarily implement officer intervention training as a means to protect both civilians and law enforcement personnel. Findings from this research suggest that officer perceptions changed significantly regarding the need and utility for officer intervention efforts. Converting officers from bystanders to interveners could save the life of a suspect, change the lives of the officers, and influence public trust of policing.

Limitations

A number of important limitations restrict the generalizability of these findings. This research was conducted in a single urban agency which may not be representative of other agencies. The agency under study experienced numerous protests following the death of George Floyd which may have influenced officer attitudes, morale, and officer perceptions of the public. Additionally, this agency has experienced difficulty in hiring and is currently down approximately 10% of possible sworn personnel. While this is not abnormal with modern urban agencies, this hiring deficit represents a possible confound.

Researchers were provided access to two of the four precincts in the agency and participants in those two precincts may not represent the perceptions of all agency officers. While participation rates were respectable, they were not high enough to generalize findings to all members of the other two precincts. Importantly, participants were not matched across the two data collection phases. This decision was made to provide anonymity and increase the likelihood of honest responding but it significantly limits the sophistication of analyses. Qualitative data collection phases were not recorded to increase honest responding and conform to the agency’s preferences. Since officer level qualitative data collection occurred in focus groups rather than interviews, some participants may have restricted their comments. Focus group participants seemed to have few reservations stating their opinions, even contravening other focus group participants, but some may have governed their perspective. Interviews with officers were preferred and requested but not approved thus focus groups were the only qualitative option with officers. The failure to record qualitative steps limits reliability and validity.

Officer intervention training was not provided to all agency personnel simultaneously—the capacity of the academy restricted the number of officers who could secure training concurrently. The training was therefore provided over a series of weeks. Some officers may have completed training and then shared their thoughts with other, untrained officers, potentially biasing their perspectives. Given these limitations, findings have limited generalizability to any specific agency.

Future Research

Expanding officer bystander intervention training to more agencies, including a variety of agencies (serving urban, suburban, and rural populations) would provide opportunities for evaluative research. Extrapolating findings from any single evaluation should be done with caution. Generalizing findings from multiple studies, which produce concurrent or similar findings, may occur with more confidence. While this study used an early phase and late phase data collection approach, officers were not tracked by any identifier. Future research which connected responses from specific officers over time would allow for more sophisticated inferential analyses and more reliable findings.

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.

ORCID iD

William V. Pelfrey

Author Biography

William V. Pelfrey, Jr. is Professor in the Wilder School of Government and Public Affairs at Virginia Commonwealth University, where he serves in both the Criminal Justice and the Homeland Security/Emergency Preparedness Programs. Dr. Pelfrey’s research revolves around public safety, particularly the intersection between policing and homeland security. He has published on police use of force, community policing, homeland security, school safety, and a variety of other topics. Dr. Pelfrey is a Fellow of the Max Planck Society in terrorism studies. Before arriving at VCU, Dr. Pelfrey worked at the University of Wisconsin—Milwaukee and the University of South Carolina.

References

Alpert

Noble

Rojek

(2021). Solidarity and the code of silence. In Dunham

Alpert

McLean

(Eds.), Critical issues in policing: Contemporary readings (7th ed.). Waveland Press.

Alpert

G. P.

MacDonald

J. M.

(2001). Police use of force: An analysis of organizational characteristics. Justice Quarterly, 18(2), 393–409. https://doi.org/10.1080/07418820100094951

Aronie

Lopez

C. E.

(2017). Keeping each other safe: An assessment of the use of peer intervention programs to prevent police officer mistakes and misconduct, using New Orleans EPIC Program as a potential national model. Police Quarterly, 20(3), 295–321. https://doi.org/10.1177/1098611117710443

Black

(1980). The manners and customs of the police. Academic Press.

Boynton

(Apr. 9, 2012). Why healthcare organizations should adopt a ‘‘No innocent bystander’’ rule for zero tolerance of workplace violence and promoting safe culture. https://www.confidentvoices.com/2012/04/09/why-hospitals.should-adopt-a-no-innocent-bystander-rule-for-zero-tolerance-of-workplace-violence-promoting-safe-culture

Chevigny

(1969). Police power: Police abuses in New York city. Pantheon Books.

Engel

Corsaro

Isaza

G. T.

McManus

H. D.

(2022). Assessing the impact of de-escalation training on police behavior: Reducing police use of force in the Louisville, KY Metro Police Department. Criminology and Public Policy, 21(9), 199–233. https://doi.org/10.1111/1745-9133.12574

Engel

R. S.

McManus

H. D.

Herold

T. D.

(2020). Does de-escalation training work? A systematic review and call for evidence in police use-of-force reform. Criminology and Public Policy, 19(3), 721–759. https://doi.org/10.1111/1745-9133.12467

Farrell

(2022). Use of force during stop and frisks: Examining the role of demeanor and race. Journal of Criminal Justice, 82(C), 102001. https://doi.org/10.1016/j.jcrimjus.2022.102001.

10.

Fischer

Greitemeyer

Pollozek

Frey

(2006). The unresponsive bystander: Are bystanders more responsive in dangerous emergencies? European Journal of Social Psychology, 36(2), 267–278. https://doi.org/10.1002/ejsp.297

11.

Fyfe

J. J.

(1978). Shots fired: An analysis of New York city police firearms discharge. Ph.D. Dissertation, State University of New York at Albany. University Microfilms.

12.

Hickman

M. J.

Piquero

A. R.

Powell

Z. A.

Greene

(2016). Expanding the measurement of police integrity. Policing, 39(2), 246–267. https://doi.org/10.1108/PIJPSM-09-2015-0104

13.

Hinduja

Patchin

(2009). Bullying beyond the schoolyard: Preventing and responding to cyberbullying. Corwin Press.

14.

Hine

K. A.

Porter

L. E.

Westera

N. J.

Alpert

G. P.

(2018). Too much or too little? Individual and situational predictors of police force relative to suspect resistance. Policing and Society, 28(5), 587–604. https://doi.org/10.1080/10439463.2016.1232257

15.

James

Vila

(2018). Testing the impact of citizen characteristics and demeanor on police officer behavior in potentially violent encounters. Policing: An International Journal, 41(1), 24–40. https://doi.org/10.1108/PIJPSM-11-2016-0159

16.

Klinger

D. A.

(1994). Demeanor or crime? Why ‘confrontational’ citizens are more likely to be arrested. Criminology, 32(3), 475–493. https://doi.org/10.1111/j.1745-9125.1994.tb01162.x

17.

New Orleans Police Department . (2012). NOPD consent decree. The city of new Orleans. Retrieved March 17, 2023 from. https://nola.gov/nopd/nopd-consent-decree/

18.

Pelfrey

W. V.

Weber

N. L.

(2015). Student and school staff strategies to combat cyberbulling in an urban student population. Preventing School Failure: Alternative Education for Children and Youth, 59(4), 227–236. https://doi.org/10.1080/1045988X.2014.924087

19.

President’s Task Force on 21st Century Policing . (2015). Final report of the president’s task force on 21st century policing. Office of Community Oriented Policing Services.

20.

Raines

J. B.

Merenda

(2022). Evaluating implementation of the EPIC peer bystander intervention program in basic law enforcement training (BLET). Police Practice and Research, 24(3), 384–396. https://doi.org/10.1080/15614263.2022.2121281

21.

Reuss-Ianni

(1983). Two cultures of policing. Transaction Publishers.

22.

Rubenstein

(1973). City police. Farrar, Strauss and Giroux.

23.

Skolnick

(1994). Justice without trial (3rd ed.). Macmillan.

24.

Skolnick

J. H.

Fyfe

J. J.

(1993). Above the law: Police and the excessive use of force. The Free Press.

25.

Stavert

R. R.

Lott

J. P.

(2013). The bystander effect in medical care. New England Journal of Medicine, 368(1), 8–9. https://doi.org/10.1056/NEJMp1210501

26.

Taniguchi

Vovak

Cordner

Amendola

Yang

Hoogestyn

Bartness

(2022). The impact of active bystander training on officer confidence and ability to address ethical challenges. Policing, 16(3), 508–522. https://doi.org/10.1353/csd.2017.0033

27.

Terrill

Zimmerman

Somers

L. J.

(2023). Applying video-based systematic social observation to police use of force encounters: An assessment of de-escalation and escalation within the context of proportionality and incrementalism. Justice Quarterly. https://doi.org/10.1080/07418825.2023.2222819

28.

Todak

James

(2018). A systematic social observation study of police de-escalation tactics. Police Quarterly, 21(4), 509–543. https://doi.org/10.1177/1098611118784007

29.

Van Maanen

(1978). The asshole. In Manning

P. K.

Van Maanen

(Eds.), Policing: A view from the street (pp. 221–238). Goodyear Publishing.

30.

Worden

R. E.

(1995). Police officers’ belief systems: A framework for analysis. American Journal of Police, 14(1), 49–81. https://doi.org/10.1108/07358549510152979

31.

Yin

R. K.

(2017). Case study research and applications: Design and methods. Sage Publishing.