3

The State of the Empirical Evidence

In Chapter 2, we discussed four basic legitimation expectations that serve as pillars for police legitimacy. Although these pillars—effectiveness, lawfulness, distributive justice, and procedural justice—are not exhaustive, they can offer a useful assessment guide for any given setting or conceived intervention. Building and sustaining police legitimacy requires police managers to pay attention to all four pillars since they each contribute to supporting the rule of law. For example, is a police service particularly good at implementing procedural justice but struggling to reduce crime? In this instance, this police force should focus on integrating effective crime reduction strategies while maintaining procedural justice practices. Does a particular intervention reduce crime but fail to achieve distributive justice? If so, officials should look for adjustments in how the strategies are executed to ensure greater fairness across identity groups. Effectiveness, lawfulness, distributive justice, and procedural justice should be pursued in tandem through complementary strategies, while ensuring that achieving one pillar does not undermine the pursuit of another.

At present, the empirical evidence on effective policing practices for building legitimacy is limited. While the existing science on police legitimacy cannot direct what should be done in all contexts, it can point to promising strategies and potential barriers that deserve consideration and evaluation. In this chapter, we review existing scientific evidence on crime reduction, community policing, procedural justice, and distributive justice. These areas were chosen because of the base of literature available and presented to the committee. We focus on crime reduction

as it is often considered an indicator of police effectiveness. Community policing, procedural justice, and distributive justice represent common strategies employed to improve police legitimacy. The discussion draws on the committee’s previous work (NASEM, 2021, 2022a,b) and builds on input from the workshop about strategies that may enhance legitimacy.

CRIME REDUCTION

Successfully meeting public expectations of police, notably effective attention to crime, is an important pillar of police legitimacy. A large body of research points to policing approaches and tactics known to reduce crime significantly. A challenge with achieving police effectiveness, however, is in translating and institutionalizing this research knowledge on crime and crime reduction into operations through policies and training. Without tracking, supervision, and other organizational changes, adopting and training on crime prevention strategies by themselves may be insufficient in achieving effectiveness as well as the other pillars (NASEM, 2021, 2022b; Sherman, 2013).

The committee’s first and second reports in this series (NASEM, 2021, 2022b) highlighted specific interventions that police can implement to be more effective at crime prevention and deterrence. As shown in Box 3-1, these interventions focus on high-risk offenders and places and on the underlying problems and conditions that create opportunities for crime. It is worth noting that while evidence on these interventions is limited to certain parts of the world, a large body of research does indicate that targeted, focused, problem-oriented, place-based, proactive, and community-centric approaches can be effective in reducing crime.

Directing more resources and police presence to high-risk micro-geographic places (and people within those places) has been a common approach to addressing crime and crime concentration in several country contexts. Blair (2022b) used the term saturation policing to refer to flooding an area with police. Others have referred to the approach as hot-spots policing. This focus on location differs from traditional notions of policing and crime prevention, which have primarily focused on policing individuals (Weisburd, 2002). While the field of policing has never discounted the value of geography, given that institutions must allocate resources and make decisions on how officers respond to the needs of their precincts, hot-spots policing focuses on much smaller areas (e.g., addresses, street blocks, or small groups of addresses or city blocks) than traditional policing (Eck and Weisburd, 1995; Sherman, Gartin, and Buerger, 1989; Weisburd, Groff, and Yang, 2012). There has been no single specified method or practice

that defines place-based policing, and as such, implementation has varied dramatically across interventions.¹

Two prior National Academies reports (NASEM, 2018; NRC, 2004) have determined that a targeted approach to crime control can be effective.

___________________

¹ As noted in the National Academies’ 2018 report Proactive Policing: Effects on Crime and Communities, actual policing practice often combines elements from two or more of the approaches. A hot-spots policing practice that seeks to engage the community could easily become a hybrid of place-based and community-based approaches.

Extensive research suggest that hot-spots policing generates small but noteworthy reductions in crime, which diffuse into areas immediately surrounding the targeted hot spots (Braga et al., 2012, 2019; Sherman, Gartin, and Buerger, 1989).² A systematic review by Braga and colleagues (2012) found that while not every hot-spots study has returned statistically significant findings, the majority of studies support the conclusion that police attention directed at crime hot spots produces a significant beneficial impact on short-term crime in these locations.³ This meta-analysis also concluded that there was little evidence to suggest that spatial displacement⁴ was a major concern; crime did not shift to other nearby locations, but instead reduced crime was observed in immediately adjacent areas (Braga, Papachristos, and Hureau, 2012; NASEM, 2018).

The evidence reviewed above points to some interventions that can help police prevent crime and reduce crime harm. However, an exclusive focus on effectiveness through crime reduction is unlikely to procure true legitimacy on its own, particularly if the crime reduction strategy reduces perceptions of lawfulness, distributive justice, or procedural justice (Bottoms and Tankebe, 2017). It is also crucial that crime reduction policies and interventions, including place-based policing tactics, are not abusive, aggressive, unlawful, or oppressive if they are to improve legitimacy and support the rule of law. As noted in the committee’s second report (NASEM, 2022b),

Therefore, the challenge for any crime reduction policy is ensuring that success in this domain does not erode the other three pillars of legitimacy—lawfulness, procedural justice, and distributive justice. The philosophy and strategy of community policing have been seen as one mechanism to support police legitimacy and procedural justice (NASEM, 2021). The vision for community policing has included enhanced accountability of the police

___________________

² See also Andresen and Malleson, 2011; Braga, Papachristos, and Hureau, 2014; Brantingham and Brantingham, 1999; Crow and Bull, 1975; Curman, Andresen, and Brantingham, 2015; Pierce, Spaar, and Briggs, 1988; Roncek, 2000; Sherman, 1997; Sherman, Gartin, and Buerger, 1989; Weisburd, 2015; Weisburd and Amram, 2014; Weisburd and Green, 1995; Weisburd, Morris, and Groff, 2009; Weisburd, Maher, and Sherman, 1992; Weisburd, Groff, and Yang, 2012.

³ Twenty of 25 tests from 19 experimental or quasi-experimental evaluations reported noteworthy crime or disorder reductions.

⁴ Spatial displacement is the movement of crime from a treatment area to a nearby location (Bowers et al., 2011).

through improved police-citizen interactions and citizen involvement in public safety (Eck and Rosenbaum, 1994; Gill et al., 2014; Mastrofski and Greene, 1988; Mastrofski, Willis, and Kochel, 2007). Unfortunately, these outcomes have yet to be realized on a large scale.

COMMUNITY POLICING

In its first report (NASEM, 2021), the committee preferred the term community-oriented policing over community policing since it emphasizes police practices that are truly concerned with involving the community in public safety and police accountability instead of those activities that just appear to engage the community. However, here and throughout this report, we use the term community policing because it is used in the paper prepared for the committee (Blair, 2022b) and in much of the cited literature. This term reflects the broad range of activities, practices, and programs associated with community policing and the reality that many of these activities and programs have not been well defined or evaluated (NASEM, 2021). Although community policing has been touted as a strategy to improve police legitimacy, research findings about its effectiveness are mixed and the link to legitimacy is weak. Where community policing practices are community oriented and involve citizens in public safety efforts, they can help garner information that police can use to focus their crime reduction strategies such as problem-oriented policing, hot-spots policing, and focused-deterrence strategies (Braga, 2015; NASEM, 2021).

Around the globe, community policing is possibly the most widely adopted policing practice to build trust and legitimacy (Blair, 2022b). Community policing in practice is more about the process of achieving some level of community engagement in police decision making and less about specific crime or social order objectives (Gill et al., 2014). Nonetheless, many articulated goals of community policing are similar to those of building legitimacy. The goals have included reducing fear, improving police-citizen relationships, increasing citizen involvement in public safety, reducing disorder, and increasing accountability and oversight of police by communities (Gill et al., 2014; Mastrofski, 1999).

Community policing is operationalized following a philosophy of community engagement. The practice of community policing encourages the public to act as partners with the police in preventing crime and promoting security based on the specific needs of the community (Mastrofksi and Greene, 1988; Skogan, 2004; Skogan and Hartnett, 1997). This form of policing usually involves some combination of the following:

• Increased police proximity to communities, such as through neighborhood foot patrols and town hall meetings;

• Increased community involvement in policing, such as through civilian advisory groups, neighborhood watch teams, and school-based programs; and
• Increased devolution of decision-making authority to officers (Blair, 2022b).

The committee’s first report (NASEM, 2021, p. 70) notes:

Community policing approaches have been criticized for the range of intended and implemented components, which makes it difficult to establish which elements (if any) were the cause of success—or failure—in attaining hypothesized and measureable outcomes. While this is a valid research challenge, recent research has examined elements of community policing in combination (Blair, 2022a; Blair et al., 2021). A team of researchers carried out a four-year, six-country study of community policing programs, using a problem-oriented policing model and engaging with the local community to share information. This study did not find that any combination of elements was effective, on average. However, they did find sometimes heterogeneous treatment effects: that some combinations under some conditions did show significant effects on some outcomes of interest—changes in crime victimization, perceived future insecurity, citizen perceptions of police, police perceptions of citizens, police abuse, crime reporting, crime tips, or the reporting of police abuse (Blair et al., 2021). The committee notes that interventions like these, which are predicated on relationship building, may need more time to exhibit effects than has been typically given in research studies.

Despite these limitations, community engagement may provide opportunities to identify basic legitimation expectations, and to work toward fulfilling those expectations. Without access to such context-specific expectations, it would be difficult for the police to gain recognition for their authority, policies, and practices. Further, community-oriented approaches improve contact with local communities, which, in turn, can increase information flow from the community to law enforcement organizations (Bell, 2016; Peyton, Sierra-Arévalo, and Rand, 2019). Such information can take

the form of intelligence about suspects and local crime activities, which can boost police effectiveness in crime control. The information may also be in the form of feedback about policing activities, allowing the police to avoid unintended consequences of their actions. Viewed this way, community policing provides law enforcement with the opportunity to better serve the needs of the public and operationalize the four pillars of legitimacy without undermining one pillar in pursuit of another.

PROCEDURAL JUSTICE

As discussed in previous chapters, procedural justice has long been recognized, operationalized, and studied as a pillar of police legitimacy. Research has assessed the use of procedural justice policing and training in several countries. For example, a randomized controlled trial (RCT) in Turkey has found that using procedural justice scripts during traffic stops improves drivers’ perceptions of the arresting officer, though not of the police more generally (Sahin et al., 2017). Experimental and quasi-experimental studies in Mexico, Colombia, and India have found that procedural justice training:

• Induces more procedurally just beliefs and behaviors among officers (Banerjee et al., 2021; Canales et al., 2021);
• Improves officers’ attitudes and sense of accountability toward citizens (Canales et al., 2021; García, Mejia, and Ortega, 2013); and
• Increases victims’ satisfaction with police (Banerjee et al., 2021).

In the United States, procedural justice trainings have been rigorously evaluated in Seattle, Washington (Owens et al., 2018), in Chicago, Illinois (Papachristos and Kirk, 2015), and in Camden, New Jersey (Goh, 2021), and all three of these studies observed reduced rates at which police officers used force. A recent RCT conducted in 120 crime hot spots within three U.S. cities found a reduction in arrests among officers who received the training compared to those who did not (Weisburd et al., 2022). Further, through pre and post-training surveys of households, the same study found that citizens in areas patrolled by officers who received the training were less likely to perceive officers as using unnecessary force; however, no significant changes concerning perceptions of police legitimacy were measured, either in terms of the hot spots themselves or the city overall.

A recent meta-analysis of 56 studies found that procedural justice is positively correlated with citizens’ perceptions of police legitimacy but also that “the causal direction of these relationships could not be tested” (Bolger and Walters, 2019, p. 98). Only a minority of these studies were conducted

in countries with low-to-moderate criminal justice sector capacity (e.g., Bosnia, Ghana, Nigeria, and Turkey). The results of an earlier 2013 meta-analysis conducted by Mazerolle and colleagues returned equally positive results, although the authors note a deficit of randomized experiments in the international literature. This 2013 meta-analysis and systemic review assessed published and unpublished empirical evidence on the impact of police-led interventions aimed at improving legitimacy, and the review shows that the dialogue component of police-citizen interactions is a meaningful mechanism for developing and fostering citizen satisfaction, confidence, compliance and cooperation with the police, and for enhancing perceptions of procedural justice (Mazerolle, Willis, and Kochel, 2013).

There is some evidence that behavioral reactions to procedural justice might depend on the normative context. Norms differ between communities, and the normative context should be considered when assessing interventions. For example, one might be in a community where supplying information to the police is strongly discouraged (e.g., “no snitching”). When the norm dissuades use of the police, procedural justice might have limited impact on people’s willingness to support the police. An alternative context may be one in which corruption characterizes the interactions between police and civilians. It is entirely possible for an officer to act procedurally justly and still take a bribe. Consideration should be given to the effect of such an experience on compliance with the law. It could be that the corrupt behaviors discredit people’s perceptions of the law. Or bribes may be seen as less costly than the hassle of legal proceedings and criminal conviction.

For example, a survey of drivers of commercial vehicles in Ghana found that procedural justice was not associated with self-reported compliance with traffic regulations (Tankebe, Karstedt, and Adu-Poku, 2019). When an interaction effect was explored between procedural justice and personal experiences of police corruption, the interaction term was positive and statistically significant. In other words, in such corrupt settings, procedural justice appears to correlate with more, not less, offending behavior. The reasons could be that police officers were willing to “listen to what drivers have to say, explain decisions, show care for drivers’ well-being, and avoid discriminatory stops of drivers, yet choose not to enforce the law on account of extra-legal considerations” (Tankebe, Karstedt, and Adu-Poku, 2019, p. 11).

DISTRIBUTIVE JUSTICE

There is no one-size-fits-all solution to inequity in policing outcomes. Here, we discuss two approaches that have been implemented widely to address issues of over-policing and under-policing discussed in Chapter 2

and improve distributive justice. These include recruitment efforts to build police units that are more representative (by gender, race, ethnicity, religion) of the communities they protect and the establishment of special units to respond to certain crimes (as has been implemented in India). At this time, there is limited empirical evidence to support the efficacy of these approaches. While there have been some positive outcomes with these approaches, there also have been some negative consequences. One explanation for these mixed results is a central focus on just creating diversity, without an equal focus on addressing any weaknesses in other legitimacy pillars. Another explanation is that too much attention is on fixing problems at the individual officer level and not enough attention is placed on institutional barriers to legitimacy (Epp, Maynard-Moody, and Haider-Markel, 2014).

There are compelling normative reasons to pursue the integration of diverse persons into the police regardless of its effects. As citizens of the democratic polity, minority groups have a right to join police organizations in their local communities. As discussed in the committee’s first report (NASEM, 2021), it is intrinsically valuable for police services to reflect the demographics of the local communities they serve. They can do so through policies that actively encourage the recruitment of underrepresented groups or remove structural obstacles to such recruitment. However, shaping the characteristics of police service personnel requires more than recruitment policies; it also requires a retention and promotions policy, especially for the retention or promotion of historically underrepresented identity groups. There is evidence that such groups often have negative experiences within police departments, suffering disparately higher reports of discrimination and general unfair treatment (Zempi, 2020).

In some contexts, increasing diversity has changed citizen-police interactions. For example, a study out of Chicago shows that relative to white officers, Black and Hispanic officers make far fewer stops and arrests, and use force less often, especially against Black citizens (Ba et al., 2021). Female officers also use less force than their male counterparts, a result that holds for all racial groups. In other contexts (Liberia and India), experimental studies report weak or even adverse effects of diversity efforts on various indicators of police-community relations, including police discrimination against ethnic minorities (Blair and Morse, 2021), police sensitivity and responsiveness to crimes that disproportionately affect women (Karim et al., 2018), and access to justice for women, and—most directly relevant—perceptions of the police among women (Jassal, 2020). As suggested by Blair (2022b), one possible explanation for this finding is that minority police officers feel pressure to conform to an existing discriminatory police subculture. Another possible explanation is that women and racial/ethnic minority officers in newly integrated police forces are pigeonholed into roles

where they are perceived to be less trustworthy and legitimate. It is also the case that “minority” communities are not homogeneous; there are class and other cleavages within them. There is no compelling reason to believe an increase in the number of minority officers will improve legitimacy or confidence in the police among civilians who simply look like those officers, given that simply being from the same minority group does not necessarily equate to shared experiences and expectations.

As discussed at the committee’s workshop, India has increased the participation of women in policing broadly (Jassal, 2020, 2022). In addition to gender quotas and inclusive representation in leadership, this participation has included the creation of specialized units and assignments, such as all-women police stations and women’s help desks. While some specialized assignments, like women’s help desks, have shown improvements to reporting rates of gender-based violence and police responsiveness to these crimes (Sukhtankar, Kruks-Wisner, and Mangla, 2022), these interventions also created physical segregation and occupational segregation (such as the assignment of tasks related to gendered crimes) between men and women police, which can exacerbate existing social devaluation of women (Jassal, 2020, 2022). Women-only police stations, or “enclaves,” were first introduced to India in 1992, and provide an alternate venue in districts to register First Information Reports and investigate crime in certain Indian states (Jassal, 2020). However, empirical findings from the dataset developed to estimate the causal effect of the enclaves opened in the Haryana state show that the intervention did not increase the number of reported crimes (Jassal, 2020). Instead, it lowered the number of cases registered in “standard” stations by justifying the deflection of gendered crimes (or violence against women), reduced the responsibilities of policewomen, and increased travel barriers for victims seeking redress. As summarized in Blair (2022b), the creation of all-women police stations resulted in a reallocation of cases, such that female police officers became more likely to be assigned “gendered” crimes (Jassal, 2020)—assignments that may have diminished their status in the eyes of civilians and potentially type-cast them (Jassal and Barnhardt, 2021).

While approaches to build legitimacy and engender trust among marginalized communities can be well intentioned, there can be unintended consequences with policies that seek to augment the roles and responsibilities for officers from minority groups. This speaks to the importance of rigorously evaluating interventions and regularly monitoring policing outcomes.

SUMMARY

The multi-faceted nature of legitimacy means that a single intervention or program is unlikely to be sufficient for the police to gain legitimacy

from its public. Thus, achieving legitimacy from the public should be conceptualized more broadly than as something attainable by only a single intervention or program. Notably, community policing initiatives have shown limited success when relegated to a single program, a select group of officers, a short time period, or without adequate monitoring. The committee believes that police legitimacy stems from achieving four responsibilities (or “pillars”) of policing in tandem: effectiveness, lawfulness, distributive justice, and procedural justice. Efforts made in only one of these and not the others can cause an imbalance that might threaten the ability to achieve legitimacy. One common example of police practices that may lead to an imbalance is when police use illegal or unjust approaches to quell violence. Another example is when police officers practice procedural justice within a specific interaction, but then select an individual for investigation based on discriminatory practices. Yet another example is when police focus only on following legal guidelines, so much so that they only respond to events procedurally, failing to protect the public.

A consistent view across the committee’s work and overarching charge is that police reform should be driven to protect the public and promote the rule of law. Achieving this overall charge and the four pillars of legitimacy cannot be done only through directives. The reason the committee advocates for an evidence-based policing approach is that achieving (and sustaining) effectiveness, lawfulness, distributive justice, and procedural justice requires a process of constant testing, tracking, and targeting of various interventions, alongside organizational adjustments that facilitate interventions, programs, and monitoring. For example, procedural training in isolation may not be enough to achieve procedural justice; officers need to practice it in the field, they need mentorship and guidance on how and when to exercise procedural justice, and police agencies need systems of accountability to ensure officers are using these skills. Implicit bias training may not be enough to create distributive justice in an agency. Officer actions have to be monitored by supervisors and accountability systems to ensure they are not discriminatorily using force or providing less service to some than others (see the committee’s third report in this series [NASEM, 2022a]).

The committee cautions against interpreting the call to balance effectiveness, lawfulness, procedural justice, and distributive justice as an assertion that a particular policy or strategy must increase one pillar at the expense of another. Instead, departments should be encouraged to actively seek and adopt policies that simultaneously advance all four pillars, and at minimum advance at least one without degrading others. For example, as described in a recent RCT by Weisburd and colleagues (2022), incorporating procedural justice training with hot-spots policing appears to increase both effectiveness and procedural justice.

In summary, the academic literature on police legitimacy has evolved in the past few decades. As empirical studies have generated mixed findings, the perspective on legitimacy has expanded. Current wisdom recognizes the importance of the four pillars presented in Chapter 2 and ongoing monitoring of efforts to achieve these pillars. The committee recommends that foreign assistance donors working with police agencies to build trust and legitimacy in local communities should:

1. Emphasize the attainment of all four pillars of effectiveness, lawfulness, distributive justice, and procedural justice needed to build legitimacy with their publics. To do this, agencies can use all four pillars to assess their strategies to address crime. Four questions might be asked of every intervention:
   1. Is the strategy lawful, constitutional, and following the tenets of human rights?
   2. Is the strategy effective, as determined by sound evaluation and assessment practices and an evidence-based approach?
   3. Is the strategy equitable? Does it benefit or harm specific groups over others?
   4. Can officers implement the strategy in ways that are respectful and that treat people with dignity?
2. Train supervisors and managers on ways to hold officers accountable to behaviors that are effective, lawful, and consistent with the principles of distributive and procedural justice.
3. Support the development of accountability mechanisms and systems. These systems might include tracking approaches using information technologies, accountability mechanisms that use body-worn camera footage, or first-line mentoring or training systems to ensure officer compliance. Tracking differential treatment and outcomes by race, ethnicity, gender, sexual orientation, or other categories is key in detecting distributive injustice in police activity. Donors could provide assistance for developing systems and processes by which to do this.
4. Promote evidence-based approaches to policing using scientific testing to measure outcomes of police practices and interventions and any improvements in perceived legitimacy.
5. Assist in building the capacity of police agencies to instill in their police officers a community-oriented mindset and the disposition to act in accordance with it. This can be accomplished through training on the four pillars of legitimacy and establishment of accountability systems that involve both supervision and tracking to ensure that officer activities are regularly assessed and in line with the four pillars. This will likely involve systematic approaches to

gauging community sentiment (surveys, focus groups, etc.). Donors may fund partnerships with universities or other nongovernmental organizations to carry out these tasks.

6. Encourage agencies to treat their own officers fairly and respectfully. Research indicates that this can affect officer treatment of citizens, and survey findings demonstrate that procedural justice practices improve citizen satisfaction with a police encounter.

This page intentionally left blank.