use of force research paper

An Empirical Analysis of Racial Differences in Police Use of Force

This paper explores racial differences in police use of force. On non-lethal uses of force, blacks and Hispanics are more than fifty percent more likely to experience some form of force in interactions with police. Adding controls that account for important context and civilian behavior reduces, but cannot fully explain, these disparities. On the most extreme use of force – officer-involved shootings – we find no racial differences in either the raw data or when contextual factors are taken into account. We argue that the patterns in the data are consistent with a model in which police officers are utility maximizers, a fraction of which have a preference for discrimination, who incur relatively high expected costs of officer-involved shootings.

This work has benefitted greatly from discussions and debate with Chief William Evans, Chief Charles McClelland, Chief Martha Montalvo, Sergeant Stephen Morrison, Jon Murad, Lynn Overmann, Chief Bud Riley, and Chief Scott Thomson. I am grateful to David Card, Kerwin Charles, Christian Dustmann, Michael Greenstone, James Heckman, Richard Holden, Lawrence Katz, Steven Levitt, Jens Ludwig, Glenn Loury, Kevin Murphy, Derek Neal, John Overdeck, Jesse Shapiro, Andrei Shleifer, Jorg Spenkuch, Max Stone, John Van Reenan, Christopher Winship, and seminar participants at Brown University, University of Chicago, London School of Economics, University College London, and the NBER Summer Institute for helpful comments and suggestions. Brad Allan, Elijah De La Campa, Tanaya Devi, William Murdock III, and Hannah Ruebeck provided truly phenomenal project management and research assistance. Lukas Althoff, Dhruva Bhat, Samarth Gupta, Julia Lu, Mehak Malik, Beatrice Masters, Ezinne Nwankwo, Charles Adam Pfander, Sofya Shchukina and Eric Yang provided excellent research assistance. Financial support from EdLabs Advisory Group and an anonymous donor is gratefully acknowledged. Correspondence can be addressed to the author by email at [email protected]. The usual caveat applies. The views expressed herein are those of the author and do not necessarily reflect the views of the National Bureau of Economic Research.

MARC RIS BibTeΧ

Download Citation Data

• online appendix
• July 5, 2016

Mentioned in the News

More from nber.

In addition to working papers , the NBER disseminates affiliates’ latest findings through a range of free periodicals — the NBER Reporter , the NBER Digest , the Bulletin on Retirement and Disability , the Bulletin on Health , and the Bulletin on Entrepreneurship  — as well as online conference reports , video lectures , and interviews .

The Factors that Motivate Law Enforcement's Use of Force: A Systematic Review

Affiliation.

• 1 School of Social Medicine, Center for Health Equity Research, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA.
• PMID: 32885851
• DOI: 10.1002/ajcp.12460

Media attention relevant to law enforcement use of force in the last decade finally alerted the scientific community to the need for more research regarding law enforcement discretion and decision making. The purpose of this study was to synthesize the existing literature to explore the motivating factors for law enforcement use of force. This study will utilize a social-ecological framework to systematically examine factors that impact officer decisions to use force at the individual and community levels. The paper includes recommendations for research and practice through an equity lens that highlights the disparate use of force against men of color in particular. Interventions, trainings, education, and research to stop the promotion of perceived safety over justice will be highlighted.

Keywords: Discretion; Law enforcement; Police; Police brutality; Social-ecological model; Use of force; Violence.

© 2020 Society for Community Research and Action.

Publication types

• Systematic Review
• Law Enforcement*
• Social Justice

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• NEWS FEATURE
• 19 June 2020
• Update 26 May 2021

What the data say about police brutality and racial bias — and which reforms might work

• Lynne Peeples 0

Lynne Peeples is a science journalist in Seattle, Washington.

You can also search for this author in PubMed   Google Scholar

For 9 minutes and 29 seconds, Derek Chauvin pressed his knee into the neck of George Floyd, an unarmed Black man. This deadly use of force by the now-former Minneapolis police officer has reinvigorated a very public debate about police brutality and racism.

Access options

Access Nature and 54 other Nature Portfolio journals

Get Nature+, our best-value online-access subscription

24,99 € / 30 days

cancel any time

Subscribe to this journal

Receive 51 print issues and online access

185,98 € per year

only 3,65 € per issue

Rent or buy this article

Prices vary by article type

Prices may be subject to local taxes which are calculated during checkout

Nature 583 , 22-24 (2020)

doi: https://doi.org/10.1038/d41586-020-01846-z

Updates & Corrections

Update 26 May 2021 : On 20 April 2021, Derek Chauvin was convicted of causing the death of George Floyd. The text has been modified to include updated information on how long Chauvin knelt on Floyd’s neck.

Edwards, F., Hedwig, L. & Esposito, M. Proc. Natl Acad. Sci. USA 116 , 16793-16798 (2019).

Article   PubMed   Google Scholar  

Nix, J., Campbell, B. A., Byers, E. H. & Alpert, G. P. Criminol. Public Policy 16 , 309–340 (2017).

Article   Google Scholar  

Knox, D., Lowe, W. & Mummolo, J. Am. Polit. Sci. Rev . https://doi.org/10.1017/S0003055420000039 (2020).

Hoekstra, M. & Sloan, C. W. National Bureau of Economic Research Working Paper 26774 (2020).

Goff, P. A. & Rau, H. Ann. Am. Acad. Polit. Soc. Sci. 687 , 67–88 (2020).

Ridgeway, G. Stat. Public Policy 3 , 1–6 (2016).

Zhao, L. & Papachristos, A. V. Ann. Am. Acad. Polit. Soc. Sci. 687 , 89–112 (2020).

Dharmapala, D., McAdams, R. H. & Rappaport, J. University of Chicago Coase-Sandor Institute for Law & Economics Research Paper No. 831 (2018).

Grunwald, B. & Rappaport, J. Yale Law J. 129 , 6 (2020).

Google Scholar  

Baude, W. 106 California Law Review 45; University of Chicago, Public Law Working Paper No. 610 (2018).

Stephens, D. W. Officer Involved Shootings: Incident Executive Summary (National Police Foundation, 2019).

Jennings, J. T. & Rubado, M. E. Public Adm. Rev. 77 , 217–226 (2017).

Ariel, B., Farrar, W. A. & Sutherland, A. J. Quant. Criminol. 31 , 509–535 (2015).

Braga, A., Coldren, J. R., Sousa, W., Rodriguez, D. & Alper, O. The Benefits of Body-Worn Cameras: New Findings from a Randomized Controlled Trial at the Las Vegas Metropolitan Police Department (National Criminal Justice Reference Service, 2017).

Wood, G., Tyler T. R., Papachristos, A. V. Proc. Natl Acad. Sci. USA 117 , 9815–9821 (2020).

O’Brien, T. C., Meares, T. L. & Tyler, T. R. Ann. Am. Acad. Polit. Soc. Sci. 687 , 202–215 (2020).

Sherman, L. W. Annu. Rev. Criminol. 1 , 421–449 (2018).

Del Toro, J. et al. Proc. Natl Acad. Sci. USA 116 , 8261–8268 (2019).

Download references

Reprints and permissions

Related Articles

US halts funding to controversial virus-hunting group: what researchers think

News 16 MAY 24

Real-world plastic-waste success stories can help to boost global treaty

Correspondence 14 MAY 24

A DARPA-like agency could boost EU innovation — but cannot come at the expense of existing schemes

Editorial 14 MAY 24

US TikTok ban: how the looming restriction is affecting scientists on the app

News 09 MAY 24

Judge dismisses superconductivity physicist’s lawsuit against university

News 25 APR 24

European ruling linking climate change to human rights could be a game changer — here’s how

World View 23 APR 24

Neglecting sex and gender in research is a public-health risk

Comment 15 MAY 24

Interpersonal therapy can be an effective tool against the devastating effects of loneliness

Inequality is bad — but that doesn’t mean the rich are

Faculty Positions& Postdoctoral Research Fellow, School of Optical and Electronic Information, HUST

Job Opportunities: Leading talents, young talents, overseas outstanding young scholars, postdoctoral researchers.

Wuhan, Hubei, China

School of Optical and Electronic Information, Huazhong University of Science and Technology

Postdoc in CRISPR Meta-Analytics and AI for Therapeutic Target Discovery and Priotisation (OT Grant)

APPLICATION CLOSING DATE: 14/06/2024 Human Technopole (HT) is a new interdisciplinary life science research institute created and supported by the...

Human Technopole

Research Associate - Metabolism

Houston, Texas (US)

Baylor College of Medicine (BCM)

Postdoc Fellowships

Train with world-renowned cancer researchers at NIH? Consider joining the Center for Cancer Research (CCR) at the National Cancer Institute

Bethesda, Maryland

NIH National Cancer Institute (NCI)

Faculty Recruitment, Westlake University School of Medicine

Faculty positions are open at four distinct ranks: Assistant Professor, Associate Professor, Full Professor, and Chair Professor.

Hangzhou, Zhejiang, China

Westlake University

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Springer Nature - PMC COVID-19 Collection

Police Uses of Force in the USA: a Wealth of Theories and a Lack of Evidence

Kyle mclean.

1 Clemson University, Clemson, SC USA

Seth W. Stoughton

2 University of South Carolina, Columbia, SC USA

Geoffrey P. Alpert

3 Griffith University, Brisbane, Australia

Research Question

How adequate is research in the USA for discovering best policies and practices, and best implementation strategies, for reducing loss of life and injury from police use of force.

This analysis examines police agency policies on the use of force regulation, evaluations of training initiatives, research on supervision, problem officers, and other dimensions of the issues and possible solutions.

The analysis examines both documented correlates and the strength of causal inference about those correlates of reductions in the use of force.

The analysis concludes that while many promising ideas have been offered, there are few tested strategies that have demonstrated substantial effects in reducing the use of force. There are virtually no successful strategies that have been replicated.

Conclusions

The current proliferation of untested programs may divert effort from a coherent and well-funded program of research to test and select effective policies that are urgently needed.

[ Academic editor’s note: this scientific communication is a revised version of a 2021 report that was commissioned by the US National Academies of Sciences, Engineering and Medicine as part of its charge to a Committee to review the “Evidence to Advance Reform in the Global Security and Justice Sectors.” Its focus is on research in the USA, including institutions, laws, and procedures that are unique to the USA. Readers from other countries may find the analysis useful by lateral thinking about the same issues in other national contexts, with different institutions, laws, and procedures. ]

Introduction

The use of force by the police remains a challenging and divisive issue in the USA and around the world. Although relatively rare in police encounters and typically coming in the form of low-level shoving and empty-hand controls (Adams, 1995) , these physical uses of governmental authority can have a devastating impact on how the public perceives the police and how citizens respond to officers (Mourtgos & Adams, 2020 ). Since the murder of George Floyd in May 2020, concern over police use of force has reached a crescendo, spurred on by the blurred information and uninformed opinions that proliferate on social media.

Motivated by entirely understandable concerns that police uses of force occur too frequently, are avoidable, excessive, or disproportionately targeted at specific populations, community members, elected officials, activists, and police leaders have made—and continue to make—a range of claims about when, why, and how much force is used. Unfortunately, many of the descriptive claims are reductive, incomplete, or simply wrong. For example, while many people are familiar with the research findings that demonstrate that African-Americans killed by police are more likely to be unarmed than white Americans killed by the police (Nix et al., 2017 ), the scope of the problem is often exaggerated. A recent study demonstrated a substantial portion of individuals believe that more than 1000 unarmed black Americans are killed every year by the police (McCaffree & Saide, 2021 ) when the number of unarmed individuals of any race killed by the police is relatively small ( n  = 93) (Nix et al., 2017 ). The disconnect between the perceived scope of the issue and the actual scope of the issue has important implications for how optimally to calibrate a public policy response.

Descriptive claims are not proffered in a vacuum; they are provided as the predicates to prescriptive assertions about how best to address the situation: Change agency policy or training! Hire different officers! Make internal oversight more rigorous! Create external oversight mechanisms! Unfortunately, many of the proposed solutions, like the descriptive claims that underlie them, lack supporting evidence.

In large part, this may be because there is substantially less information available about police uses of force than one might expect. Consider that, in less than 3 years since the emergence of the SARS-CoV-2 pandemic, researchers and public health professionals have collected a wealth of data about infections, hospitalizations, deaths, and a number of potential pharmaceutical interventions. In contrast, there is an embarrassing paucity of data about the frequency and severity of officers’ use-of-force decisions. Even looking only at lethal force events, the data reported by official sources—including the Federal Bureau of Investigation and the Centers for Disease Control and Prevention—lag far behind the actual count (Loftin et al., 2017 ). The data on nonlethal incidents are more dismal yet. Only one state, New Jersey, has anything approaching standardized, robust data on (statewide) uses of force, and they exist only as a result of a prolonged effort by a private media organization to obtain and code all uses of force in the state over a 4-year period (NJ Advance Media, 2018 ).

Addressing appropriately important crises—whether the COVID-19 pandemic or the use of excessive force by police—requires scientific evidence that is sufficiently robust, granular, and representative to allow for a complete understanding of the issues and the development of clear communication about successful responses. The bases for public policymaking on technical matters must withstand scientific and social scrutiny and stand above political whims and ideological beliefs.

Of course, the difference between the COVID crisis, on the one hand, and policing, on the other, is painfully simple: billions of dollars. Billions of dollars have been spent to develop the scientific evidence necessary to inform the COVID policy recommendations relating to masking, physical distancing, and vaccinations. To date, no similar investment has been made to understand or resolve the issue of excessive force. Operation Warp Speed—the effort to quickly develop a COVID vaccine—had a reported budget of $12.4 billion (Barone, 2020 ), while the entirety of the research and evaluation budget for the Office of Justice Programs in 2020 was significantly less than 1% of that figure at only $79 million (Office of Justice Programs, 2021 ).

Despite the regrettable lack of investment, social scientists across a range of disciplines have dedicated their careers to understanding and reducing the use of excessive force by the police. Scholars in fields ranging from sociology (e.g., Legewie, 2016 ) to economics (e.g., Fryer, 2019 ) to political science (e.g., Mullinix et al., 2021 ) and, of course, criminology (e.g., Nix et al., 2017 ), have attempted to understand the scope, causes, and possible solutions to the issue of excessive force. Despite this considerable attention, only recently has the use of force been studied with rigorous research methods, and only recently has that research started to have a systemic impact on public policy. 1

This modern effort to develop sound evidence about policing has been linked to a growing public concern, as well as an increasing acceptance of evidence-based methods and practices from police command staff, policymakers, and trainers. Research findings have formed the basis for many decisions, which were previously left to dubious theories, unsupported assumptions, and political whims. Unfortunately, the evidentiary foundation is both thin and full of gaps, incapable of meeting the needs and requests of either the public or police practitioners. If we truly consider the issue of police use of force as being as urgent and pressing as the COVID pandemic, then we must dedicate greater resources to the scientific study of policing as a distinct set of social practices rather than as a topic suited principally for philosophical discussion and theoretical musings.

This is not to say that policies should not be enacted, and practices should not be undertaken in the absence of a body of indisputable scientific evidence. That would be unrealistic. It is to say, however, that public policy should be predicated on robust evidence when available, on reliably developed findings when possible, and on sound theoretical assumptions when needed. To the detriment of police agencies, officers, and communities, it is far too common to see use-of-force policymaking based on nothing more than the historical practices that are grandiosely titled “conventional wisdom” (or, put another way, “how things have always been done”). Even recent reforms reflect good ideas, but few empirical studies of their effectiveness. Sound public policy demands more than optimism and untested assumptions.

Understanding, explaining, and regulating police uses of force and excessive force requires a multifaceted approach that incorporates a thorough understanding of policies, training, supervision, accountability, and their interactions. In this article, we will review the research findings in each of the identified areas to show the evidence, or lack thereof, that drives our knowledge and decisions concerning the uses of force. In each of these substantive areas, scholars and policymakers have generated a lot of good ideas, tested only a few of them, and replicated an even smaller number of them. This leaves us with an abundance of testable ideas, but sizeable gaps in our knowledge of their efficacy.

We can do better. We must do better. Our knowledge base must be broadened and deepened through rigorous evaluation so that the public, elected officials, and police leaders can have faith in research findings and use them to develop rational and reasonable responses to this critical issue.

Agency policies that regulate the use of force represent perhaps the clearest area in which decisions are dominated by the conventional wisdom about what seems like a good idea instead of data about what works. This is dictated, in no small part, by the need for agencies to adopt defensible policies that comply with state and federal law (Stoughton et al., 2020 ), reflect existing jurisprudence on the use of force (Stoughton, 2021 ), and garner both officer and public acceptance. Unfortunately, jurisprudence and public sentiment do not always make for an effective means of properly regulating the use of force.

What has come to be called “best practices” for use-of-force policies are guided by a small number of key organizations, each of which has released so-called “model policies.” The International Association of Chiefs of Police (IACP) has released its National Consensus Policy, the Police Executive Research Foundation (PERF) has released its Guiding Principles on the Use of Force, the College of Policing in the UK has a list of general principles and three core questions for every officer to evaluate during a potential use of force encounter, and the Australia and New Zealand Policing Advisory Agency (ANZPAA) has a list of recommended use of force principles. These documents overlap in several areas, such as emphasizing that officers use force only when it is reasonable, necessary, and proportionate. They each make a number of distinct recommendations that include incorporating a statement on the sanctity of life in agency policy, adopting a range of components of “de-escalation,” articulating that officers have a duty to intervene, and, in the case of PERF’s Guiding Principles, advocating for agencies to provide all patrol officers with riot shields. Regardless of their content, each of these documents aims to build upon clear legal, moral, and philosophical foundations. Notably absent from that foundation, however, is empirical evidence of effectiveness.

Empirical examinations of use-of-force policies are few and far between. Early examinations of force policies used the use of force continuum to create a force factor score for examining the prevalence of potentially problematic use of force incidents (Terrill, 2005 ; Alpert & Dunham, 1997 , 2004 ). This work, while informative, provides little information on whether the use of that continuum—or certain characteristics of the various continua—reduced the uses of force. Similarly, Terrill and Paoline ( 2013 ) conducted a national survey of police departments to assess the prevalence of the use of force continua in the early 2010s. Their study demonstrated that while use of force continua are adopted in most police departments, there is substantial variation in the structure and specification of the continua. Again, however, there was no reliable indication about the substantive impact, if any, of these continua on officers’ use-of-force decision-making.

A small number of studies have determined that more restrictive force policies lead to reductions in the use of force, likely beginning with work by Fyfe ( 1979 ). White ( 2001 ) used interrupted time series analyses to examine the impact of a series of policy changes on officer-involved shootings in the Philadelphia Police Department, concluding that restrictive policies can reduce the number of shootings. Similarly, Terrill and Paoline ( 2017 ) studied less-lethal force policies in eight departments across the country, concluding that more restrictive policies were associated with fewer use of force incidents.

Still, analyses in this area struggle to identify specific causes. Methodologically, studies of agency policies are plagued by endogeneity problems. For legal reasons, different policies cannot be randomly assigned to individual officers. Policies also tend to be revised substantially after major incidents or leadership changes, introducing variables beyond the policy revision itself (such as public outcry). Furthermore, policy changes are often wholesale, with the rewriting of an entire policy or set of policies, rather than with isolated changes to specific policy components.

Policy revisions—of whatever variety—cannot easily be disentangled from the modifications to training or supervisory expectations that accompany those revisions. Accordingly, analyses that examine alterations to use of force policies cannot cleanly determine whether any impact results from not only the policy change, but a host of additional factors, and if so what specific change drives which impact. Similarly, analyses that compare the use of force policies across multiple jurisdictions must contend with the issue that more restrictive force policies may reflect different agency cultures or that training within these departments that may be contributing to any detected effects. To be clear, these observations are not a criticism of the scholars who have worked in this area but rather a statement on how difficult it is to design and conduct this type of research.

In fact, the only recent peer-reviewed study that identified a specific change in use of force policy and measured the impact on reported uses of force in our review was conducted by Shjarback and colleagues (Shjarback et al., 2021 ). These researchers assessed the impact of a change in policy at the Dallas Police Department that required officers to report when they pointed a firearm at a subject, finding that the policy modification resulted in a slow but lasting reduction in officer-involved shootings.

Finally, it should be noted that several journalistic studies of the use of force (e.g., the Washington Post’s Fatal Force Project, Campaign Zero) have provided some recommendations on the contents of use of force policies based on data. These studies are not peer-reviewed and come with many of the same caveats noted above regarding separating the independent impact of individual policies. Overall, their recommendations are largely consistent with the findings that more restrictive policies result in fewer uses of force. Still, it is striking that efforts by advocacy organizations and the media have created more evidence-based suggestions than governmental agencies or many of the professional organizations that spend significant effort promulgating and promoting their model force policies for nationwide adoption.

Compared to agency policies, police training has been the focus of considerably more research. Over the past decade, in-service training programs have been studied for their potential to reduce the use of force. One study examined the impact of a procedural justice training program in the Chicago Police Department on trainees’ attitudes compared to a control group of untrained officers (Skogan et al., 2015 ). This study was one of the few evaluations of police training and advanced the state of the literature, but it was disadvantaged by the lack of experimental design and its failure to specifically measure the use of force. Procedural justice—the outcome of interest in the study—has been linked to citizens’ compliance (Tyler, 2006 ; Tyler & Huo, 2002 ), offering a theoretical connection between procedural justice and the use of force, but the study did not look at the uses of force as a distinct outcome.

Giacomantonio and colleagues (Giacomantonio et al., 2020 ) later pushed this type of evaluation further by examining the impact of a verbal judo training program in the Halifax Regional Police (CAN) on officers’ performance in simulated scenarios. On the one hand, their study used advanced technology—an immersive virtual reality simulator—to expose treatment and control officers to the same stimuli to test for differences in behavior. On the other hand, the study lacked an experimental design and showed no differences in the use of force between trained and untrained officers. Additionally, there is no robust evidence as to the relationship between changes in performance in simulated scenarios, especially using virtual reality simulators, and changes to performance in the field.

Following these initial studies, a number of experimental evaluations have been conducted to test for differences in administrative reports of uses of force. These include an evaluation of Polis Solutions’ T3 training program by McLean and colleagues (McLean et al., 2020 ), an evaluation of PERF’s ICAT program by Engel and colleagues (Engel et al., 2022 ), an evaluation of procedural justice training by Wood and colleagues (Wood et al., 2020 ), and an evaluation of de-escalation training by White and colleagues (White et al., 2021a , b ).

McLean and colleagues’ study randomized individual officers to treatment and control conditions and was strengthened by an experimental design and repetitive training exposure but was limited by potential contamination issues and agency measures of use of force (see McLean et al., 2020 for a discussion). The results showed no significant differences in the number of force incidents involving trained officers compared to the number involving non-trained officers.

Engel and colleagues’ use of a stepped-wedge design limited contamination issues but left unaddressed the possibility that the observed reductions could be attributed to exogenous changes that coincided with the introduction of training, such as distinct policy changes, the occurrence of critical incidents in the jurisdiction, or de-policing (see also Worden and colleagues (Worden et al., 2020 ), for a stepped-wedge study impacted by external factors).

Wood and colleagues studied a procedural justice training program in the Chicago Police Department using a stepped-wedge design. The study had similar issues with the potential for external factors influencing the use of force (see Correction to Wood et al., 2020 ) and found a small reduction in the use of force in the Chicago Police Department.

White and colleagues (White et al., 2021a , b ) recently published their evaluation of a de-escalation training program in the Tempe Police Department (TPD) that found no impact on officers’ uses of force. However, as the authors note:

[A]dministrative use of force and injury data are blunt measures that might not capture the impact of the training. In simple terms, TPD is not a high use of force department and injuries are rare. Since use of force is a rare event, the evaluation needed to take a more nuanced approach to fully capture the impact of de-escalation training. (White et al., 2021a , b:60).

These studies represent the first rigorous reviews of training programs designed to reduce the frequency and severity of uses of force. But several commonalities that dominate the studies create serious issues for providing scientific recommendations about police training. First, each study evaluated a different training program, and no known studies attempt to replicate existing evaluations of police training. Accordingly, the evidence base for any training program is limited to a single study. Even in the most promising of studies—perhaps especially with the most promising of studies—replication is important to improve confidence in the generalizability of the findings.

Second, each study evaluated an in-service training program. In-service training programs are undoubtedly important, but the focus on in-service training in experimental evaluations has highlighted the complete lack of evidence on what programs and techniques are effective in preservice (academy) training of new officers.

Third, each study (except the one that evaluated Polis Solutions’ T3 program) evaluated a one-time training program provided to officers over 1 or 2 days. Best practices for adult learning in training from other fields suggest that high-repetition training over more prolonged periods is important for both retention and understanding (Ericsson, 2004 ; Ericsson & Ward, 2007 ).

To be clear, a number of promising training programs exist that may reduce officers’ uses of force or use of excessive force. Programs like Georgetown University’s Active Bystandership for Law Enforcement promote peer intervention when officers perceive a colleague to be using or about to use excessive force (Active Bystandership for Law Enforcement Project, 2021 ). The Marietta, GA, Police Department has adopted a Brazilian Jiu-Jitsu program that they claim will increase officers’ confidence in their physical abilities, making officers both more willing to communicate before employing physical control techniques and more willing to use empty-hand controls rather than resorting to intermediate weapons such as TASERS (Huddleston, 2021 ). Additionally, the LaGrange, GA, Police Department has developed “Shoot to Incapacitate,” a controversial program that trains officers to shoot at subjects’ pelvic girdles or upper legs in certain deadly force situations to reduce lethality vis-à-vis center mass (Schrade, 2021 ). All three programs reflect innovative ideas and are based on sound theoretical assumptions, but none of the three is backed by rigorous scientific research into their effectiveness in achieving their stated goals.

In sum, our review of studies measuring the impact of police training on the use of force is consistent with our review in each of the other areas. Promising ideas abound, but empirical research continues to lag behind the growing desire to implement evidence-based training programs. Dedicating more money, time, and resources to this critical topic is key to reducing the quantity and improving the legitimacy of police use of force incidents.

Internal Police Oversight

Police oversight focuses on the internal and external mechanisms that police departments use to ensure that officers follow both governing legal rules and their own agency’s administrative policies and training. Internal oversight mechanisms are a critical component of regulation; even the best policies and training programs will be undermined if officers do not believe that the department takes them seriously enough to ensure compliance. In 1981, the U.S. Commission on Civil Rights examined the issue of police oversight in Who’s Guarding the Guardians , a report that is a “must read” for students of police reform and still guides many of the topics that we will review in this section. Discussions of accountability and oversight in early reports on reform highlight a division of internal and external control mechanisms, which we will echo here (see Noble & Alpert, 2009 and Kane & White, 2009 ). However, both the 1981 report and previous reports on police misconduct and associated issues (including the 1973 National Advisory Commission: Task Force on Police, see Walker, 2012 ) suffer from a lack of meaningful data and evidence-based recommendations. Here, too, we see the theme: “good ideas but no data.”

Supervision

Internal oversight represents efforts undertaken within a police department to ensure that its officers are following its policies, training, and relevant legal precedents. Supervision represents one of the more promising areas of research for police oversight when considering the weight of evidence about its impact on the uses of force. Findings that effective supervision affects an officer’s use of force are hardly surprising, given the role that bosses have in encouraging or even requiring particular types of behavior on their shifts. It is often stated that the success of any system of oversight starts with the frontline supervisors and their leadership abilities (see Perez, 1994 ). At the same time, the practical implementation of these findings is unclear, given the limited research on how to identify, promote, and train good supervisors. Practically speaking, we might know that good supervision matters without knowing how to promote it.

Early efforts at police reform focused heavily on the role that supervision played in improving policing. The 1981 Who’s Guarding the Guardians report noted—without explaining how to fix it—that the lack of supervision allowed officers to act outside of training and policies, one consequence of which is the excessive use of force. In the late 1980s, the International Association of Chiefs of Police ( 1989 : 57) published a study on building integrity and concluded that “Supervisors who fail to audit the activities of personnel fail to perform their jobs.” Their point was that first-line supervisors were responsible for knowing their officers’ strengths, weaknesses, and styles of working with the community and for socializing and training officers in their proper roles to reduce any inclination or need to use force unnecessarily. Although this necessary aspect of supervision has been recognized for decades, it remains surprising that very little research has been conducted on the impact of supervisory styles and emphases.

Early studies on the role of supervision in police use of force used surveys to correlate officer attitudes and behaviors to supervisor’s attitudes and behaviors (Engel, 2000 , 2001 , 2002 ; Engel & Worden, 2003 ). Using these techniques, Engel ( 2000 ) classified officers with “active” supervisors as (1) having increased activity in the field; (2) those who take over situations; (3) acting with the dual role of both a street officer and a supervisor; and (4) more likely to use force than officers with less active supervisors. Unsurprisingly, modeling “active” behavior encourages street officers to be more active and engage in more uses of force.

A more recent study examined the impact of procedurally fair supervision on officers’ use of force (Owens et al., 2018 ). In this study, officers identified as working in high-risk circumstances were randomly assigned to either a treatment or control condition. In the treatment condition, officers had a non-disciplinary meeting with their supervisor in which the supervisor modeled procedurally fair behavior, reviewed a recent incident the officer engaged in, and pointed out areas of success and potential areas for improvement. This cognitive debrief was juxtaposed to the control group who were debriefed in a “matter-of-fact” manner, without any emphasis on procedural justice. As a result of the intervention, treated officers were less likely than control officers to engage in uses of force in the 6 weeks following their supervisory meeting. In fact, the officers in the experimental group passed forward the procedurally just treatment they learned from their supervisors to the individuals with whom they subsequently interacted (Owens et al., 2017 ).

This study is particularly promising. It not only reinforces the hypothesis that supervisors have a meaningful impact on officers’ uses of force: it also identifies a specific intervention that can promote this type of supervision. This is promising, but it suffers from many of the same caveats as the studies of police training: the intervention only resulted in a demonstrated reduction in uses of force for 6 weeks; it was studied in only a single police department and has never been subject to replication. There is no evidence on whether the effects of supervisory meetings are purely transient, or on the optimal number, frequency, and duration of such supervisory meetings, or on the opportunity costs of any long-term, repetitive intervention such as the redistribution of supervisor’s time away from other supervisory responsibilities.

Body-Worn Cameras

Another potentially fruitful area of research on supervision comes from the body-worn camera literature (Adams et al., 2021 ; see Lum et al., 2019 for a review of body-worn camera research). Arguments for body-worn cameras typically claim that cameras can increase supervision because supervisors can review footage of police officers performing daily duties (Stoughton, 2018 ). However, the reality of the use of body-worn cameras for internal supervision is likely much weaker.

Koen and Mathna ( 2019 ) found that, in a department with a specific policy that supervisors should review body-worn camera footage proactively to look for potential issues with officers, proactive reviews were rarely (if ever) actually conducted. The management of the body-worn camera program required a considerable investment of time and effort to curate footage for cases that were either proceeding through the judicial system or involved citizen complaints. Very little time was left over for reviewing footage that had not been specifically requested by someone else. This practice substantially weakens the supervision effects of body-worn camera policy as the footage is only reviewed when the department is already aware that something has happened. This effectively uses body-worn camera footage as a source of information about known incidents rather than as a mechanism for random reviews that could uncover otherwise unknown issues, such as low-level uses of force omitted from an officer’s written report.

Despite these unanswered questions, the clear message from this literature is that supervision generally is a promising avenue for reducing the use of force and merits dedicating considerably more resources to evaluating and replicating promising programs.

Focusing on Problem Officers

It is frequently said in policing that 10% of the population causes 90% of the problems that officers are called upon to address (see United States Commission on Civil Rights ( 1981 ). Although the specific numbers may prove imprecise, the concept that a relatively small percentage of people are responsible for a disproportionate amount of a particular behavior is a familiar one. This concept certainly applies in policing, where police leaders and academics have for years been concerned about the small group of officers who are responsible for a large percentage of force incidents or misconduct (bad apples). Data from New Jersey, for example, shows that 10% of officers in the state accounted for 38% of all uses of force, and a small group of just 252 officers used force at a rate five times higher than the state average (NJ Advance Media, 2018 ).

One internal accountability tool noted in Who’s Guarding the Guardians is the Early Warning System, later renamed the Early Identification System (EIS). An EIS uses internally-generated data on various aspects of officers’ performance—ranging from uses of force to use of sick leave—to alert supervisors to officers who are potentially problematic before they become a serious concern. While there is evidence that EIS can be effective if managed properly with moderate goals (Walker et al., 2001 ), a recent systematic review of EIS shows that there were only eight studies that matched the selection criteria for proper methods, six of which were single agency studies (Gullion & King, 2020 ). The results of the reviewed studies are encouraging, but measures of success/effectiveness are inconsistent. Here, as elsewhere in the police research literature, data drawn from single-agency studies provide only weak results.

Chalfin and Kaplan ( 2021 ) recently published a study examining the potential outcomes if EIS were properly implemented including the use of resources to identify and extricate problem officers. Concluding that removing “high-complaint ‘bad apples’” from police departments would result in only small reductions in civilian complaints, Chalfin and Kaplan contend that agencies should focus their resources on broader reforms rather than committing substantial resources to EIS. Sierra-Arévalo and Papachristos ( 2021 ), in turn, challenged the analytical approach that Chalfin and Kaplan relied upon, concluding that the Chalfin and Kaplan estimates are comparable to the modest results seen in other programs. Specifically, Sierra-Arévalo and Papachristos make an argument similar to ours: the existing literature is all but devoid of empirically based recommendations that result in substantial reductions in the use of force or citizen complaints. Rather than committing resources to policies and programs with indeterminate outcomes, Sierra-Arévalo and Papachristos suggest, a modest (4–6%) but proven reduction may be preferable to unsupported optimism.

This argument is worth discussing at greater length because it reflects the conundrum replete in our findings. A policymaker or police chief who wishes to commit resources to reduce the uses of force in their agency and wants to make an evidence-based policy determination must choose among reforms that are all associated with modest reductions. These may include the development of EIS or the implementation of the procedural justice training program evaluated by Wood and colleagues (Wood et al., 2020 ), or programs that are theoretically sound, and which may offer a more robust effect but have yet to be evaluated, such as Active Bystandership for Law Enforcement. Even assuming the policymaker is scientifically literate and well informed, their position in making such choices is not an enviable one (see Fleming & Rhodes, 2018 and Aronie & Alpert, 2020 ).

In the event that a policymaker agrees with Sierra-Arévalo and Papachristos’ ( 2021 ) that these “rotten apples” should be terminated, that decision may be easier to make than to implement. Investigating and terminating an officer can be difficult and time consuming, not least because of the substantive and procedural protections provided by collective bargaining agreements (Rushin, 2017 , 2019 ) or the state Law Enforcement Officers Bill of Rights. Additionally, departmental discipline is often overturned in arbitration (Rushin, 2021 ) or by civil service boards (see Grunwald & Rappaport, 2020 ).

Even assuming that termination is successful, doing so may provide no more than a local solution that displaces problem officers to other agencies in rather the same way that certain crime-fighting strategies may result in the non-reductive displacement of criminal behavior (Johnson et al., 2014 ). Robust decertification programs, which strip someone of their license to be a police officer in the state, are sparse (Goldman, 2013 , 2016 ). Many fired officers find work with other agencies, especially during difficult recruiting periods (Grunwald & Rappaport, 2020 ). Hiring an experienced officer provides some cost savings to agencies, and an agency administrator must balance the immediate reality of putting an officer on the street without the lag time of training against the speculative possibility that the officer may engage in future misconduct.

That said, hiring an officer who resigned in lieu of termination or was fired or asked to resign can create significant liability. A study of such “wandering officers” in Florida found that they were significantly more likely to receive complaints of violent or sexual conduct even when controlling for age, gender, and education (Grunwald & Rappaport, 2020 ). Plaintiffs’ attorneys highly value cases involving an officer who repeated the bad behavior that caused them to leave a prior agency; indeed, such situations can open the door to municipal liability claims that are otherwise difficult to establish because they require evidence that the agency adopted an unconstitutional policy or custom. While the total cost of lawsuits against the police is unknown, news reports and academic assessments suggest they are quite significant (see Schwartz, 2014 ). A 2021 report by the Marshall Project shows that New York and Chicago have paid out more than 2 billion dollars in police misconduct claims over the past 10 years (Thomson-Devaux et al., 2021 ).

External Police Oversight

Like internal oversight, external oversight seeks to ensure that officers follow governing legal rules, agency policies and training, and community expectations. Unlike internal oversight, external oversight goes outside of the traditional command and control mechanisms that have evolved within police agencies by involving agency outsiders in the form of civilian review boards, consent decrees, or individual litigation.

Civilian Review Boards

Civilian Review Boards are often touted as one of the essential tools for police reform. In the late 1990s, the National Institute of Justice funded a project to examine civilian oversight of the police (Finn, 2001 ). The resulting report explained in detail the structures and functions of nine different systems of oversight and described the strengths and weaknesses of each. The nine systems are then aggregated into four types of programs (Finn, 2001 , vii), including.

• Citizens investigate allegations of police misconduct and recommend findings to the chief or sheriff;
• Police officers investigate allegations and develop findings; citizens review them and recommend that the chief or sheriff approve or reject the findings;
• Complainants may appeal findings established by the police or sheriff’s department to citizens, who review them and then recommend their own findings to the chief or sheriff;
• An auditor investigates the process by which the police or sheriff’s department accepts and investigates complaints and reports on the thoroughness and fairness of the process to the department and the public.

Importantly, we should not divorce civilian oversight from other accountability mechanisms, including internal accountability. One of the major conclusions of the 2001 report was the need for a comprehensive system of accountability, as “citizen oversight alone cannot ensure that police will act responsibly” (Finn, 2001 , vii). While the report noted that all oversight efforts should be evaluated (Chapter 7) and included a summary from an assessment of the Albuquerque oversight program (Luna & Walker, 1997 ), there was no methodological roadmap for a systematic review of oversight programs.

In 2016, De Angelis et al. ( 2016 ) surveyed almost 100 police oversight executives and reported on the key issues associated with the implementation and sustainability of civilian oversight of police. As in prior reports, they looked at trends in the implementation of these oversight bodies and discussed objectives and goals but reported no systematic data or evaluations of the various programs.

In a more recent review, Stephens et al. ( 2018 ) reviewed the results of a roundtable discussion and a survey concerning the civilian oversight programs in the major city police departments. The report ( 2018 , p. 20) identifies and reviews multiple objectives of civilian oversight, including.

• Transparency;
• Independent investigations;
• Improving accountability;
• Improving public trust and legitimacy;
• Engaging the community; and
• Demystifying police internal affairs investigations.

While the Stephens et al. ( 2018 ) report updates Finn’s 2001 review regarding the structures and functions of these boards, it concludes (p. 20) that “There is little, if any, empirical research on the effectiveness of civilian oversight of the police—nor are the programs subjected to any systematic evaluation.”

Since these reviews have been published, the National Association for Civilian Oversight of Law Enforcement has provided an impressive array of civilian oversight models ( https://www.nacole.org ). It does not, however, offer any empirical evidence demonstrating the effectiveness of civilian oversight. Similarly, when the Council on Criminal Justice: Task Force on Policing published a Policy Assessment on Civilian Oversight in April 2021 (CCJ Task Force on Policing, 2021 ), it reviewed the variety of models and the potential merits of civilian oversight and concluded (p. 1, 3):

Rigorous empirical research on the impact of civilian oversight and the relative merits of different models does not exist. Other research has yielded mixed findings about the ability of civilian oversight to reduce excessive use of force and other forms of police misconduct. … Research to date has been largely descriptive of existing civilian oversight entities, either taking a deep look at the operations of a single entity or comparing and contrasting across the functions and structures of several entities.

Specifically, the report reviews the most influential studies and notes many methodological problems, including the lack of control over policy changes during the research period. The result is a characteristically mixed message: civilian oversight has promising aspects, it is helpful as a matter of public relations and transparency, but there is no empirical evidence that it works to reduce police misconduct or the use of force. We know that civilian oversight of the police is controversial, that there is not an accepted methodology or set of criteria to measure effectiveness; and that, as a result, there have not been any systematic studies of the effectiveness of any of the various models of civilian review. Once again, we are left with a lot of good ideas and theories but precious little data.

Consent Decrees

One approach to police use of force that is unique to the US stems from post-Civil War legislation that makes deprivation of civil rights under color of law a federal crime, as well as a tort. More recent legislation empowers the US Department of Justice to put state and local police agencies into a de facto state of “receivership” (bankruptcy) in which a court can effectively take over a police agency. This approach can take several specific forms, including Consent Decrees, Settlement Agreements, and Collaborative Reforms. These all offer another model of externally generated reform. Each category of these external reforms differs from the others in terms of degree and oversight requirements.

While consent decrees have been around for a few decades, they and their data have been shrouded in secrecy until recently (Jiao, 2021 and Alpert et al., 2017 ). Unfortunately, the secrecy under which the consent decrees have operated has kept valuable evaluation data out of the hands of researchers or restricted the publication of studies. Monitors’ reports and agency data may be public, but the research behind them is unlikely to be made available. While there are a number of issues to consider, such as federal versus state consent decrees and lawsuits, the lack of access to these data leaves research with lost opportunities.

The research conducted in several agencies under consent decrees could assist reform efforts across the country. On the one hand, a comprehensive evaluation of the paragraphs of the “Findings Letters” and/or decrees would provide a real-time roadmap of where agencies are failing and where they need assistance. These data are available and could be used by other agencies as a framework for reviewing their own policies, training, supervision, and accountability systems in light of what they know about why a Finding Letter was issued. The data collected by the agencies subject to these external controls could also be analyzed. They could, for example, demonstrate the agencies’ capacity for improvement and ability to fulfill the requirements of the consent decree or other reform agreements with the government. These data could be analyzed to develop a roadmap or toolkit for other agencies to follow. This may prove particularly fruitful, given that the process of satisfying a Consent Decree requires input and oversight from agency command staff, external monitors, and a battery of government attorneys. In this way, both the failures and successes of one agency could be leveraged by others. For the most part, though, these data have not been made readily available or used to assist other agencies.

As for the impact or effectiveness of consent decrees, it is difficult to determine whether that impact justifies the significant amount of time, money, and effort that they inevitably require. The relative merits of a consent decree are difficult to assess with any precision. Several of the best descriptive studies of consent decrees are by Goh ( 2020 ), Rushin ( 2015 ), Chanin ( 2015 , 2017 ), and Harmon ( 2009 ). There is some evidence that consent decrees and other forms of structural reform litigation have been effective in reducing misconduct (Rushin, 2015 ), but we do not know if consent decrees were an efficient or cost-effective mechanism for reform. Chillar ( 2021 : 7) conducted an extensive and thorough review of the evaluation research conducted on consent decrees and published an important article that concluded as follows:

While some consent decree evaluations find compelling evidence regarding the outcomes used to track compliance and the positive impacts police and citizens experienced, it is important to note that the studies were conducted during the period in which the department was under consent decree or was recently released from federal oversight. This ensured that the departments had no time to revert to the problematic policies and practices that warranted the implementation of the consent decree. Following the exit of the monitoring team in the cities of Pittsburgh, Washington D.C., Cincinnati, Detroit, and Prince George’s County in Maryland, Chanin ( 2014 , 2015 ) used focus groups and administrative data to investigate the effect of consent decrees. While results indicate that consent decrees offered departments the opportunity to bring about organizational change, such change faltered and reverted to business as usual upon the absence of DOJ oversight.

Chillar’s observation reflects an issue raised in several of the articles; whether the reforms that are measured in the short-term term persist over the long term. His own research examined the short-term impacts on police stops of community members by the Newark, NJ Police Department. Overall, Chillar found that the consent decree reforms likely explained improvements in data quality collected in field interrogations but that there were unexplained decreases in reported field interrogations of blacks and Latinos. Significantly, he concluded that regardless of specific reforms and oversight efforts, cultural change in the agency was necessary to improve police behavior and actions toward black and Latino community members.

The best available evidence shows that consent decrees tend to improve departments in the short term, but that, in the absence of ongoing oversight, agencies will revert to many of the problematic behaviors that resulted in the consent decree in the first place.

Qualified Immunity

Perhaps, no aspect of the legal regulation of policing has been subject to more contentious public debate than qualified immunity: a legal doctrine that protects officers who violate the Constitution from civil liability unless it was “clearly established” at the time that their action was a constitutional violation. In other words, so long as a “reasonable officer” could have thought that a particular action did not violate the Constitution, an officer will not be liable even if they did violate the Constitution. And the unconstitutionality of an officer’s actions is only “clearly established” if it is particularly egregious or if a court in the jurisdiction has decided a factually similar case. As the leading qualified immunity scholar described it, “Current Supreme Court doctrine suggests that an officer violates clearly established law only if there is a prior court of appeals or Supreme Court decision holding virtually identical facts to be unconstitutional” (Schwartz, 2021 ). Under this standard, qualified immunity has become a significant bar to vindicating civil rights claims.

In no small part, this is because the constitutional standards that regulate policing—such as Fourth Amendment standards of proof like “probable cause” and “reasonable suspicion”—are themselves often indistinct (Stoughton et al., 2021 ). The result is a series of questions about the substantive standard (i.e., whether an officer violated the constitution) and the applicability of qualified immunity (i.e., whether the violation was clearly established) that are both framed in ways that often favor and defer to officers.

In part because of these ambiguities and the barriers to accountability that they create, there are arguments both to maintain and abolish qualified immunity. On the one hand, it is argued that the doctrine helps weed out weak cases, develop acceptable practices, develop the law, and protect the purses of municipalities. On the other hand, as Professor Joanna Schwartz has noted ( 2017 : 1):

The United States Supreme Court appears to be on a mission to curb civil rights lawsuits against law enforcement officers, and appears to believe qualified immunity is the means of achieving its goal. The Supreme Court has long described qualified immunity doctrine as robust—protecting ‘all but the plainly incompetent or those who knowingly violate the law.’ (citing Mallory v Briggs, 475 U.S. 335, 341 (1986)).

Her work, along with that of Kinports ( 2016 ), explores the arguments and judicial opinions that have used qualified immunity as a shield for the police, protecting them from abusing or mistreating community members. Schwartz presents an analysis of the role of qualified immunity in constitutional litigation that leads her to conclude that qualified immunity has been seen as both unclear and unreasonably protective of government actors.

The evidence on qualified immunity suggests that the law, as it now stands, not only shields officers from liability for abuses, it also shields their employers. Private employers are liable for the actions of their employees under a legal theory known as “vicarious liability” or “respondeat superior,” but public employers, including police agencies, are not. Nevertheless, police agencies virtually always cover the costs of defending their officers in litigation as well as the settlement or judgment costs, known in law as indemnification (Schwartz, 2014 ). Thus, when individual officers are insulated from liability, it is almost always employers, rather than the officers themselves, that benefit.

What is far less clear is whether eliminating qualified immunity—or revising municipal liability doctrines or tort reform, for that matter—would have any significant effect on officer action. Schwartz has predicted that there would be more civil rights litigation if qualified immunity were abolished, but that the average time, cost, and complexity per case would decrease and that the success rates for plaintiffs and defendants would remain roughly stable (Schwartz, 2020 ). Here, too, we have a sound theory but little empirical information. Schwartz’s predictions are well informed, based on doctrinal analysis and interviews with attorneys, but there is no supporting analytic model. As the 2021 Marshall Project concluded its report finding that police misconduct cost New York and Chicago more than 2 billion dollars over 10 years, “If not for inconsistent and shoddy record-keeping, we might know if settlements make a difference in police misconduct” (Thomson-Devaux et al., 2021 ).

The current state of the literature on police use of force can best be summarized as a wealth of ideas with a lack of support. Our review suggests that quality studies of policies, training, selection, supervision, and accountability exist in small numbers, with few of the many policies and programs implemented by police agencies having been evaluated properly. Existing studies of policy changes, training, selection or supervision programs, or accountability measures never contained more than a few agencies and were frequently limited to single departments.

Moreover, replications of existing studies are virtually nonexistent. Creating a robust evidence base requires replication. Yet replication is generally rare across social sciences. One analysis, for example, examined articles in 100 psychological journals with the highest impact scores, (Makel et al., 2012 ), finding that only 1% of published studies are replications.

There are a number of reasons for this replication crisis. First, journal editors who place a higher price on originality and positive results are reluctant to publish replications. In fact, a study of journal editors by Madden et al. ( 1995 , p. 85) concluded:

A paradox of replication in the social sciences is that a researcher who operationally replicates and finds nonsupport for previous work may be accused of not being true to the original method, but if the researcher finds support for the previous work, then the argument will be made that nothing new has been learned.

The difficulty of getting replications published reduces the desirability of conducting them, although it is worth noting that the more controversial the topic the more likely a replication study will be published (Madden et al., 1995 ).

The second major reason for the lack of replication studies is the difficulty of obtaining funding for major applied research projects in policing. As noted in the introduction, the research and evaluation budget for the Office of Justice Programs—the major national agency that is responsible for funding these types of evaluations—is relatively small, especially considering that it covers not only policing but other areas of criminal justice (e.g., courts, corrections, and forensic sciences). The annual Research and Evaluation on Policing solicitation funds evaluations of policing programs but usually fewer than 5 per year, as in 2021. In a competitive environment where only a select number of evaluations of policing programs are funded every year, it is nearly impossible to obtain funding to replicate an evaluation of a program in a different context. Even www.CrimeSolutions.gov , a program sponsored by the National Institute of Justice (NIJ) that rates research to determine the certainty of the outcomes and effectiveness, does not use replications as an evaluation criterion. The National Institute for Justice (or another federal agency) should create a solicitation dedicated to replications of studies that show promise by prior evidence, as already exists in the private sector (Arnold Ventures). 2

In the end, the strongest recommendation we can provide is that more resources—in the form of time, money, effort, and commitment from police researchers and agencies alike—must be dedicated to studying police behavior and uses of force. More detailed policies, different equipment, better communication skills, the extended use of mental health professionals, improved supervision, changes to accountability strategies, and other reforms could all work—or, more likely, could all marginally advance the ultimate goal as components of comprehensive reform—but it will take time and diligent research to determine how to best make them work and whether each particular effort is individually or collectively worth the time, effort and cost.

Generating an evidence base for dealing with the use of force in policing is exceedingly difficult. Reducing unnecessary and excessive force is an objective of vital importance to the public. National attention to this issue has generated the immediate need for departments to “do something.” Unfortunately, from a scientific perspective, immediate action is often not supported by or conducive to generating solid research findings. Just as responsible public health officials did not immediately roll out untested SARS-CoV-2 treatments nationwide, we should resist the urge to rush out untested programs to reduce the use of force nationwide despite the urgent need for improvement.

Indeed, the importance of the issue is precisely why we must take a deliberative approach. Getting it right slowly comes with real costs, but those costs are far less than continuing to get it wrong indefinitely. This is not to suggest that no one should take action. Rather, those actions that are taken should be, at a minimum, predicated on sound theory and then evaluated to determine their efficacy. The implementation of new programs and policies must be accompanied by evaluations. Resources should be expended to replicate those evaluations and ensure the accuracy of the findings. In this area, the USA, in particular, lags considerably behind other democratic countries, yet it never seems to learn its lesson. Each new use of force controversy is accompanied by a plethora of new ideas but little support for evaluating them.

Developing a truly evidence-based approach to the regulation of police use of force will take time and resources. But if this is—properly, in our view—regarded as an urgent social issue, then society should make funding more readily available. We should all demand that police agencies commit to not only relying on theoretically sound and evidence-based reforms but also to taking part in research that includes their own policing agencies.

Acknowledgements

This manuscript is based on a report submitted to the National Academies of Sciences, Engineering, and Medicine for their project entitled, “Evidence to Advance Reform in the Global Security and Justice Sectors: A Workshop-based Consensus Study Series. 2021.” It primarily reviews uses of force in US police agencies, but its assessment of the dearth of evidence-based policing practices designed to reduce frequency and severity is equally relevant around the world.

1 For our purposes, excessive force refers to force that exceeds what the governing legal standards and/or what the public deems reasonable in a given society. When we refer to the study of the use of force, this includes excessive force, as well as reasonable uses of force. Academic studies in this area tend to focus on the use of force rather than excessive force as excessive force requires a determination of unreasonableness that is unavailable in most datasets. Accordingly, we focus on the study of the use of force with the intent of reducing the problem of excessive force.

2 https://craftmediabucket.s3.amazonaws.com/uploads/RFP-for-RCTs-in-Criminal-Justice-CJ-EBP.pdf

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

• Active Bystandership for Law Enforcement. (2021). Active bystandership for law enforcement (ABLE) project fact sheet. Retrieved from: https://drive.google.com/file/d/1PPImHwxyl8nYP8E38FIXVNKSXoQbCjf7/view
• Adams K. Measuring the prevalence of police abuse of force. In: Geller WA, Toch H, editors. And Justice for All: Understanding and Controlling Police Abuse of Force. Police Executive Research Forum; 1995. pp. 61–98. [ Google Scholar ]
• Adams I, Mourtgos S, Mastracci S. High-stakes administrative discretion: What drives body-worn camera activations? Public Administration Review. 2021; 81 :691–703. doi: 10.1111/puar.13339. [ CrossRef ] [ Google Scholar ]
• Alpert GP, Dunham RG. The force factor: Measuring police use of force relative to suspect resistance. Police Executive Research Forum; 1997. [ Google Scholar ]
• Alpert GP, Dunham RG. Understanding police use of force: Officers, suspects, and reciprocity. Cambridge University Press; 2004. [ Google Scholar ]
• Alpert G, McLean K, Wolfe S. Consent decrees: An approach to police accountability and reform. Police Quarterly. 2017; 20 :239–249. doi: 10.1177/1098611117709591. [ CrossRef ] [ Google Scholar ]
• Aronie, J, & Alpert, G. (2020). 16 to a dealer’s 10: Could blackjack odds help inform police pursuit policies? PoliceOne. March 2. Retrieved from: https://www.police1.com/police-products/pursuit-managementtechnology/articles/16-to-a-dealers-10-could-blackjack-odds-help-inform-police-pursuit-policies-Y9x2Avr8SWzJG1c4/
• Barone, E. (2020). The Trump administration’s ‘Operation Warp Speed’ has spent $12.4 billion on vaccines. How much is that, really? Time. Retrieved from:  https://time.com/5921360/operation-warpspeed-vaccine-spending/ . Accessed 11 Aug 2021
• Chalfin A, Kaplan J. How many complaints against police officers can be abated by incapacitating a few “bad apples?” Criminology & Public Policy. 2021; 20 :351–370. doi: 10.1111/1745-9133.12542. [ CrossRef ] [ Google Scholar ]
• Chanin J. On the implementation of pattern or practice police reform. Criminology, Criminal Justice, Law & Society. 2014; 15 :38–56. [ Google Scholar ]
• Chanin J. Examining the sustainability of pattern or practice police misconduct reform. Police Quarterly. 2015; 18 :163–192. doi: 10.1177/1098611114561305. [ CrossRef ] [ Google Scholar ]
• Chanin J. Early stages of pattern or practice police misconduct reform: An examination of the Department of Justice’s investigation and negotiation processes. Police Quarterly. 2017; 20 :250–274. doi: 10.1177/1098611117709268. [ CrossRef ] [ Google Scholar ]
• Chillar VF. The racial divide at micro places: A pre/post analysis of the effects of the Newark Consent Decree on field inquiries (2015–2017) Journal of Research in Crime and Delinquency. 2021 doi: 10.1177/00224278211030964. [ CrossRef ] [ Google Scholar ]
• CCJ Task Force on Policing. (2021). Civilian oversight: Policy assessment. Retrieved from:  https://assets.foleon.com/eu-west-2/uploads-7e3kk3/41697/civilian_oversight.2690411fd370.pdf . Accessed 11 Aug 2021
• De Angelis, J., Rosenthal, R., & Buchner, B. (2016). Civilian oversight of law enforcement: A review of the strengths and weaknesses of various models NCJ Number 250265 Downloaded December 20, 2021).  https://www.ojp.gov/ncjrs/virtual-library/abstracts/civilian-oversight-law-enforcement-assessing-evidence . Accessed 11 Aug 2021
• Engel RS. The effects of supervisory styles on patrol officer behavior. Police Quarterly. 2000; 3 :262–293. doi: 10.1177/1098611100003003003. [ CrossRef ] [ Google Scholar ]
• Engel RS. Supervisory styles of patrol sergeants and lieutenants. Journal of Criminal Justice. 2001; 29 :341–355. doi: 10.1016/S0047-2352(01)00091-5. [ CrossRef ] [ Google Scholar ]
• Engel RS. Patrol officer supervision in the community policing era. Journal of Criminal Justice. 2002; 30 :51–64. doi: 10.1016/S0047-2352(01)00122-2. [ CrossRef ] [ Google Scholar ]
• Engel RS, Worden RE. Police officers’ attitudes, behavior, and supervisory influences: An analysis of problem solving. Criminology. 2006; 41 :131–166. doi: 10.1111/j.1745-9125.2003.tb00984.x. [ CrossRef ] [ Google Scholar ]
• Engel RS, Corsaro N, Isaza GT, McManus HD. Assessing the impact of de-escalation training on police behavior: Reducing police use of force in the Louisville, KY Metro Police Department. Criminology & Public Policy. 2022; 21 :199–233. doi: 10.1111/1745-9133.12574. [ CrossRef ] [ Google Scholar ]
• Ericsson KA. Deliberate practice and the acquisition and maintenance of expert performance in medicine and related domains. Academic Medicine. 2004; 79 :570–581. doi: 10.1097/00001888-200410001-00022. [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Ericsson KA, Ward P. Capturing the nautrally occurring superior performance of experts in the laboratory: Toward a science of expert and exceptional performance. Current Directions in Psychological Science. 2007; 16 :346–350. doi: 10.1111/j.1467-8721.2007.00533.x. [ CrossRef ] [ Google Scholar ]
• Finn, P. (2001). Citizen review of police: Approaches and implementtion. Washington, DC: National Insitute of Justice. Retrieved from: https://www.ojp.gov/pdffiles1/nij/184430.pdf
• Fleming J, Rhodes R. Can experience be evidence? Craft knowledge and evidence-based policing. Policy and Politics. 2018; 46 :3–26. doi: 10.1332/030557317X14957211514333. [ CrossRef ] [ Google Scholar ]
• Fryer, R.G. (2019). An empirical analysis of racial differences in police use of force. Journal of Political Economy .  127 , 1210–1261.
• Fyfe J. Administrative interventions on police shooting discretion: An empirical examination. Journal of Criminal Justice. 1979; 7 :309–323. doi: 10.1016/0047-2352(79)90065-5. [ CrossRef ] [ Google Scholar ]
• Giacomantonio C, Goodwin S, Carmichael G. Learning to de-escalate: Evaluating the behavioural impact of Verbal Judo training on police constables. Police Practice and Research. 2018; 21 :401–417. doi: 10.1080/15614263.2019.1589472. [ CrossRef ] [ Google Scholar ]
• Goldman RL. A model decertification law. Saint Louis University Public Law Review. 2013; 32 :147–156. [ Google Scholar ]
• Goldman RL. The importance of state law in police reform. Saint Louis University Law Journal. 2016; 60 :363–390. [ Google Scholar ]
• Goh LS. Going local: Do consent decrees and other forms of federal intervention in municipal police departments reduce police killings? Justice Quarterly. 2020; 37 :900–929. doi: 10.1080/07418825.2020.1733637. [ CrossRef ] [ Google Scholar ]
• Grunwald B, Rappaport J. The wandering officer. Yale Law Journal. 2020; 129 :1676–1782. [ Google Scholar ]
• Gullion C, King W. Early intervention systems for police: A state-of-the-art review. Policing: An International Journal. 2020; 43 :643–658. doi: 10.1108/PIJPSM-02-2020-0027. [ CrossRef ] [ Google Scholar ]
• Harmon R. Promoting civil rights through proactive policing reform. Stanford Law Review. 2009; 62 :1–68. [ Google Scholar ]
• Huddleston, D. (2021). Marietta police officers learning Brazilian Jiu Jitsu to help keep the streets safe. WSB-TV 2. Retrieved from:  https://www.wsbtv.com/news/local/cobb-county/marietta-police-officers-learning-brazilian-jiu-jitsu-help-keep-streets-safe/CCYMTFR47FFD7B4QU5WOUQGC6Q/ . Accessed 11 Aug 2021
• International Association of Chiefs of Police . Building integrity and reducing drug corruption in police departments. IACP; 1989. [ Google Scholar ]
• Jiao A. Federal consent decrees: A review of policies, processes, and outcomes. Police Practice and Research. 2021; 22 :793–804. doi: 10.1080/15614263.2020.1722664. [ CrossRef ] [ Google Scholar ]
• Johnson S, Guerette RT, Bowers K. Crime displacement: What we know, what we don’t know, and what it means for crime reduction. Journal of Experimental Criminology. 2014; 10 :549–571. doi: 10.1007/s11292-014-9209-4. [ CrossRef ] [ Google Scholar ]
• Kane RJ, White MD. Bad cops: A study of career-ending misconduct among New York City police officers. Criminology & Public Policy. 2009; 8 :737–769. doi: 10.1111/j.1745-9133.2009.00591.x. [ CrossRef ] [ Google Scholar ]
• Kinports K. The Supreme Court’s quiet expansion of qualified immunity. Minnesota Law Review. 2016; 100 :62–78. [ Google Scholar ]
• Koen M, Mathna B. Body-worn cameras and internal accountability at a police agency. American Journal of Qualitative Research. 2019; 3 :1–22. doi: 10.29333/ajqr/6363. [ CrossRef ] [ Google Scholar ]
• Legewie J. Racial profiling and use of force in police stops: How local events trigger periods of increased discrimination. American Journal of Sociology. 2016; 122 :379–424. doi: 10.1086/687518. [ CrossRef ] [ Google Scholar ]
• Loftin C, McDowall D, Xie M. Underreporting of homicides by police in the United States, 1976–2013. Homicide Studies. 2017; 21 (2):159–174. doi: 10.1177/1088767917693358. [ CrossRef ] [ Google Scholar ]
• Lum C, Stultz M, Koper CS, Scherer JA. Research on body-worn cameras: What we know, what we need to know. Criminology & Public Policy. 2019; 18 :93–118. doi: 10.1111/1745-9133.12412. [ CrossRef ] [ Google Scholar ]
• Luna, E., & Walker, S. (1997). A report on the oversight mechanisms of the Albuquerque Police Department, prepared for the Albuquerque City Council. Retrieved from: https://www.cabq.gov/council/documents/Walker%20Luna%20Report.pdf
• Madden C, Easley R, Dunn M. How journal editors view replication research. Journal of Advertising. 1995; 24 :77–87. doi: 10.1080/00913367.1995.10673490. [ CrossRef ] [ Google Scholar ]
• Makel MC, Plucker JA, Hegarty B. Replications in psychology research: How often do they really occur? Perspectives on Psychological Science. 2012; 7 :537–542. doi: 10.1177/1745691612460688. [ PubMed ] [ CrossRef ] [ Google Scholar ]
• McCaffree, K. & Saide, A. (2021). How informed are Americans about race and policing? Skeptic Research Center, CUPES-007.
• McLean K, Wolfe SE, Rojek J, Alpert GP, Smith MR. Randomized controlled trial of social interaction police training. Criminology & Public Policy. 2020; 19 :805–832. doi: 10.1111/1745-9133.12506. [ CrossRef ] [ Google Scholar ]
• Mourtgos SM, Adams IT. Assessing public perceptions of police use-of-force: Legal reasonableness and community standards. Justice Quarterly. 2020; 37 :869–899. doi: 10.1080/07418825.2019.1679864. [ CrossRef ] [ Google Scholar ]
• Mullinix KJ, Bolsen T, Norris RJ. The feedback effects of controversial police use of force. Political Behavior. 2021; 43 :881–898. doi: 10.1007/s11109-020-09646-x. [ CrossRef ] [ Google Scholar ]
• Nix J, Campbell BA, Byers EH, Alpert GP. A bird's eye view of civilians killed by police in 2015: Further evidence of implicit bias. Criminology & Public Policy. 2017; 16 :309–340. doi: 10.1111/1745-9133.12269. [ CrossRef ] [ Google Scholar ]
• NJ Advance Media. (2018). The force report. NJ.com. Retrieved from  https://force.nj.com/ . Accessed 11 Aug 2021
• Noble J, Alpert G. Managing accountability systems for police conduct: Internal affairs and external oversight. Waveland Press; 2009. [ Google Scholar ]
• Office of Justice Programs. (2021). FY 2021 performance budget. US DOJ. Retrieved from:  https://www.justice.gov/doj/page/file/1246736/download . Accessed 11 Aug 2021
• Owens E, Weisburd D, Amendola KL, Alpert GP. Research in brief: Passing it forward: Modeling procedural jutice in the workplace to improve public-police interactions. Police Chief. 2017; 79 :14–15. [ Google Scholar ]
• Owens E, Weisburd D, Amendola KL, Alpert GP. Can you build a better cop? Experimental evidence on supervision, training, and policing in the community. Criminology & Public Policy. 2018; 17 :41–87. doi: 10.1111/1745-9133.12337. [ CrossRef ] [ Google Scholar ]
• Perez DW. Common Sense About Police Review. Temple University Press; 1994. [ Google Scholar ]
• Rushin S. Structural reform litigation in American police departments. Minnesota Law Review. 2015; 99 :1343–1422. [ Google Scholar ]
• Rushin S. Police union contracts. Duke Law Journal. 2017; 66 :1191–1266. [ Google Scholar ]
• Rushin S. Police disciplinary appeals. University of Pennsylvania Law Review. 2019; 167 :545–610. [ Google Scholar ]
• Rushin S. Police arbitration. Vanderbilt Law Review. 2021; 74 :1023–1078. [ Google Scholar ]
• Schrade, B. (2021). In Georgia agency, police train to shoot, not kill. The Atlanta Journal-Constitution. Retrieved from: https://www.ajc.com/news/crime/in-georgia-agency-police-train-to-shoot-notkill/IJNVJCHXBRHJHKPFHLEXQ672YI/
• Schwartz J. Police indemnification. New York University Law Review. 2014; 89 :885–1006. [ Google Scholar ]
• Schwartz J. How qualified immunity fails. The Yale Law Journal. 2017; 127 :2–76. [ Google Scholar ]
• Schwartz J. After qualified immunity. Columbia Law Review. 2020; 120 :309–388. [ Google Scholar ]
• Schwartz J. Qualified immunity’s boldest lie. The University of Chicago Law Review. 2021; 88 :605–684. [ Google Scholar ]
• Shjarback JA, White MD, Bishopp SA. Can police shootings be reduced by requiring officers to document when they point firearms at citizens? Injury Prevention. 2021 doi: 10.1136/injuryprev-2020-043932. [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Sierra-Arévalo M, Papachristos A. Bad apples and incredible certitude. Criminology & Public Policy. 2021; 20 :371–381. doi: 10.1111/1745-9133.12545. [ CrossRef ] [ Google Scholar ]
• Skogan WG, Van Craen M, Hennessy C. Training police for procedural justice. Journal of Experimental Criminology. 2015; 11 :319–334. doi: 10.1007/s11292-014-9223-6. [ CrossRef ] [ Google Scholar ]
• Stephens D, Scrivner E, Cambareri J. Civilian oversight of the police in major cities. Office of Community Oriented Policing Services; 2018. [ Google Scholar ]
• Stoughton SW. Police body-worn cameras. North Carolina Law Review. 2018; 96 :1363–1423. [ Google Scholar ]
• Stoughton SW. How the Fourth Amendment frustrates the regulation of police violence. Emory Law Journal. 2021; 70 :521–585. [ Google Scholar ]
• Stoughton SW, Noble JJ, Alpert GP. Evaluating police uses of force. NYU Press; 2020. [ Google Scholar ]
• Stoughton S, McLean K, Nix J, Alpert G. Policing suspicion: Qualified immunity and “clearly established” standards of proof. Journal of Criminal Law & Criminology. 2021; 112 :1–42. [ Google Scholar ]
• Terrill W. Police use of force: A transactional approach. Justice Quarterly. 2005; 22 :107–138. doi: 10.1080/0741882042000333663. [ CrossRef ] [ Google Scholar ]
• Terrill W, Paoline EA. Examining less lethal force policy and the force continuum: Results from a national use-of-force study. Police Quarterly. 2013; 16 :38–65. doi: 10.1177/1098611112451262. [ CrossRef ] [ Google Scholar ]
• Terrill W, Paoline EA. Police use of less lethal force: Does administrative policy matter? Justice Quarterly. 2017; 34 :193–216. doi: 10.1080/07418825.2016.1147593. [ CrossRef ] [ Google Scholar ]
• Thomson-Devaux, A., Bronner, L., & Sharma, D. (2021). Police misconduct costs cities millions every year. But that’s where the accountability ends. The Marshall Project. Retrieved from: https://www.themarshallproject.org/2021/02/22/police-misconduct-costs-cities-millions-every-year-but-that-s-where-the-accountability-ends . Accessed 11 Aug 2021
• Tyler TR. Why people obey the law. Princeton University Press; 2006. [ Google Scholar ]
• Tyler TR, Huo YJ. Trust in the law: Encouraging public cooperation with the police and courts. Russell Sage; 2002. [ Google Scholar ]
• United States Commission on Civil Rights . Who’s Guarding the Guardians. USGPO; 1981. [ Google Scholar ]
• Walker S. Institutionalizing police accountability reforms: The problem of making police reforms endure. Saint Louis University Public Law Review. 2012; 32 :57–92. [ Google Scholar ]
• Walker, S., Alpert, G.P., & Kenney, D.J. (2001). Early warning systems: Responding to the problem police officer. National Institute of Justice: Research in Brief.  https://www.ojp.gov/pdffiles1/nij/188565.pdf
• White MD. Controlling police decisions to use deadly force: Reexamining the importance of administrative policy. Crime & Delinquency. 2001; 47 :131–151. doi: 10.1177/0011128701047001006. [ CrossRef ] [ Google Scholar ]
• White MD, Mora VJ, Orosco C, Hedberg EC. Moving the needle: Can training alter officer perceptions and use of de-escalation? Policing: An International Journal. 2021; 44 :418–436. doi: 10.1108/PIJPSM-08-2020-0140. [ CrossRef ] [ Google Scholar ]
• White, M. D., Orosco, C., Pooley, M., & Sorenson, D. (2021b). Testing the impact of de-escalation training on officer behavior: The Tempe (AZ) Smart Policing Initiative. Retrieved from:  https://www.smart-policing.com/sites/default/files/inline-files/Tempe%20SPI%20Final%20Report%2012-21.pdf . Accessed 11 Aug 2021
• Wood G, Tyler TR, Papachristos AV. Procedural justice training reduces police use of force and complaints against officers. Proceedings of the National Academy of Sciences of the United States of America. 2020; 117 :9815–9821. doi: 10.1073/pnas.1920671117. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Worden, R., McLean, S. J., Engel, R. S., Cochran, H., Corsaro, N., Reynolds, D., Najdowski, C. J., & Isaza, G. T. (2020). The impacts of implicit bias awareness training in the NYPD. IACP/UC Center for Police Research and Policy & the John F. Finn Institute for Public Safety. Retrieved from: https://www1.nyc.gov/assets/nypd/downloads/pdf/analysis_and_planning/impacts-of-implicit-bias-awareness-training-in-%20the-nypd.pdf . Accessed 11 Aug 2021

• Follow us on Facebook
• Follow us on Twitter
• Criminal Justice
• Environment
• Politics & Government
• Race & Gender

Expert Commentary

Excessive or reasonable force by police? Research on law enforcement and racial conflict

Updated review of studies and reports that provide insights into law enforcement actions and recent patterns in America.

Republish this article

This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License .

by John Wihbey and Leighton Walter Kille, The Journalist's Resource July 28, 2016

This <a target="_blank" href="https://journalistsresource.org/criminal-justice/police-reasonable-force-brutality-race-research-review-statistics/">article</a> first appeared on <a target="_blank" href="https://journalistsresource.org">The Journalist's Resource</a> and is republished here under a Creative Commons license.<img src="https://journalistsresource.org/wp-content/uploads/2020/11/cropped-jr-favicon-150x150.png" style="width:1em;height:1em;margin-left:10px;">

Allegations of the use of excessive force by U.S. police departments continue to generate headlines more than two decades after the 1992 Los Angeles riots brought the issue to mass public attention and spurred some law enforcement reforms . Recent deaths at the hands of police have fueled a lively debate across the nation in recent years.

In a number of closely watched cases involving the deaths of young black men, police have been acquitted, generating uproar and concerns about equal justice for all. On Staten Island, N.Y., the July 2014 death of Eric Garner because of the apparent use of a “chokehold” by an officer sparked outrage . A month later in Ferguson, Mo., the fatal shooting of teenager Michael Brown by officer Darren Wilson ignited protests, and a grand jury’s decision not to indict Wilson triggered further unrest . In November, Tamir Rice was shot by police in Cleveland, Ohio. He was 12 years old and playing with a toy pistol. On April 4, 2015, Walter L. Scott was shot by a police officer after a routine traffic stop in North Charleston, S.C. The same month, Freddie Gray died while in police custody in Baltimore, setting off widespread unrest . The policeman in the South Carolina case, Michael T. Slager, was charged with murder based on a cellphone video. In Baltimore, the driver of the police van in which Gray died, Caesar Goodson, was charged with second-degree murder , with lesser charges for five other officers. There have been no indictments in the earlier cases.

These follow other recent incidents and controversies, including an April 2014 finding by the U.S. Department of Justice (DOJ), following a two-year investigation, that the Albuquerque, N.M., police department “engages in a pattern or practice of use of excessive force, including deadly force, in violation of the Fourth Amendment,” and a similar DOJ finding in December 2014 with regard to the Cleveland police department . In March 2015, the DOJ also issued a report detailing a pattern of “clear racial disparities” and “discriminatory intent” on the part of the Ferguson, Mo., police department.

As the Washington Post reported in July 2015 , a pervasive problem that is only now beginning to be recognized is the lack of training for officers dealing with mentally ill persons, a situation that can often escalate to violent confrontations.

The events of 2014-2016 have prompted further calls by some police officials, politicians and scholars for another round of national reforms, in order to better orient “police culture” toward democratic ideals.

Two sides, disparate views

Surveys in recent years with minority groups — Latinos and African-Americans , in particular — suggest that confidence in law enforcement is relatively low, and large portions of these communities believe police are likely to use excessive force on suspects. A 2014 Pew Research Center survey confirms stark racial divisions in response to the Ferguson police shooting, as well, while Gallup provides insights on historical patterns of distrust. According to a Pew/ USA Today poll conducted in August 2014, Americans of all races collectively “give relatively low marks to police departments around the country for holding officers accountable for misconduct, using the appropriate amount of force, and treating racial and ethnic groups equally.” Social scientists who have done extensive field research and interviews note the deep sense of mistrust embedded in many communities.

Numerous efforts have been made by members of the law enforcement community to ameliorate these situations, including promising strategies such as “community policing.” Still, from a police perspective, law enforcement in the United States continues to be dangerous work — America has a relatively higher homicide rate compared to other developed nations, and has many more guns per capita. Citizens seldom learn of the countless incidents where officers choose to hold fire and display restraint under extreme stress. Some research has shown that even well-trained officers are not consistently able to fire their weapon in time before a suspect holding a gun can raise it and fire first; this makes split-second judgments, even under “ideal” circumstances, exceptionally difficult. But as the FBI points out, police departments and officers sometimes do not handle the aftermath of incidents well in terms of transparency and clarity, even when force was reasonably applied, fueling public confusion and anger.

In 2013, 49,851 officers were assaulted in the line of duty, with an injury rate of 29.2 percent, according to the FBI. Twenty-seven were murdered that year.

FBI Director: No “reliable grasp” of problem

How common are such incidents of police use of force, both lethal and non-lethal, in the United States? Has there been progress in America? The indisputable reality is that we do not fully know. FBI Director James B. Comey stated the following in a remarkable February 2015 speech :

Not long after riots broke out in Ferguson late last summer, I asked my staff to tell me how many people shot by police were African-American in this country. I wanted to see trends. I wanted to see information. They couldn’t give it to me, and it wasn’t their fault. Demographic data regarding officer-involved shootings is not consistently reported to us through our Uniform Crime Reporting Program. Because reporting is voluntary, our data is incomplete and therefore, in the aggregate, unreliable.

I recently listened to a thoughtful big city police chief express his frustration with that lack of reliable data. He said he didn’t know whether the Ferguson police shot one person a week, one a year, or one a century, and that in the absence of good data, “all we get are ideological thunderbolts, when what we need are ideological agnostics who use information to try to solve problems.” He’s right.

The first step to understanding what is really going on in our communities and in our country is to gather more and better data related to those we arrest, those we confront for breaking the law and jeopardizing public safety, and those who confront us. “Data” seems a dry and boring word but, without it, we cannot understand our world and make it better.

How can we address concerns about “use of force,” how can we address concerns about officer-involved shootings if we do not have a reliable grasp on the demographics and circumstances of those incidents? We simply must improve the way we collect and analyze data to see the true nature of what’s happening in all of our communities.

The FBI tracks and publishes the number of “justifiable homicides” reported by police departments. But, again, reporting by police departments is voluntary and not all departments participate. That means we cannot fully track the number of incidents in which force is used by police, or against police, including non-fatal encounters, which are not reported at all.

Without a doubt, training for police has become more standardized and professionalized in recent decades. A 2008 paper in the Northwestern University Law Review provides useful background on the evolving legal and policy history relating to the use of force by police and the “reasonableness” standard by which officers are judged. Related jurisprudence is still being defined, most recently in the 2007 Scott v. Harris decision by the U.S. Supreme Court. But inadequate data and reporting — and the challenge of uniformly defining excessive versus justified force — make objective understanding of trends difficult.

A 2015 report  conducted for the Justice Department analyzed 394 incidents involving deadly police force in Philadelphia from 2007-2014. It found that “officers do not receive regular, consistent training on the department’s deadly force policy”; that early training among recruits is sometimes inadequate in regard to these issues; that investigations into such incidents are not consistent; and that officers “need more less-lethal options.”

For perhaps the best overall summary of police use-of-force issues, see “A Multi-method Evaluation of Police Use of Force Outcomes: Final Report to the National Institute of Justice,” a 2010 study conducted by some of the nation’s leading criminal justice scholars.

Available statistics, background on use of force

The federal Justice Department releases statistics on this and related issues, although these datasets are only periodically updated: It found that in 2015, among the 53.5 million U.S. residents aged 16 or older who had any contact with police, 985,300 of them — 1.8 percent — experienced threats or use of force .  Law enforcement officials were more likely to threaten or use force on black people and Hispanics than white people, according to an October 2018 report . “ When police initiated the contact , blacks (5.2 percent) and Hispanics (5.1 percent) were more likely to experience the threat or use of physical force than whites (2.4 percent), and males (4.4 percent) were more likely to experience the threat or use of physical force than females (1.8 percent).” Of those who experienced a threat or use of force, 84 percent considered it to be excessive. In terms of the volume of citizen complaints, the Justice Department also found that there were 26,556 complaints lodged in 2002; this translates to “33 complaints per agency and 6.6 complaints per 100 full-time sworn officers.” However, “overall rates were higher among large municipal police departments, with 45 complaints per agency, and 9.5 complaints per 100 full-time sworn officers.” In 2011, about 62.9 million people had contact with the police.

In terms of the use of lethal force, aggregate statistics on incidents of all types are difficult to obtain from official sources. Some journalists are trying to rectify this; and some data journalists question what few official national statistics are available. The Sunlight Foundation explains some of the data problems , while also highlighting databases maintained by the Centers for Disease Control (CDC). The available data, which does not paint a complete national picture, nevertheless raise serious questions, Sunlight notes:

[A]ccording to the CDC, in Oklahoma the rate at which black people are killed per capita by law enforcement is greater than anywhere else in the country. That statistic is taken from data collected for the years 1999-2011. During that same time period, Oklahoma’s rate for all people killed by law enforcement, including all races, is second only to New Mexico. However, Oklahoma, the District of Columbia, Nevada and Oregon are all tied for the rate at which people are killed. (The CDC treats the District of Columbia as a state when collecting and displaying statistics.) In Missouri, where Mike Brown lived and died, black people are killed by law enforcement twice as frequently as white people. Nationwide, the rate at which black people are killed by law enforcement is 3 times higher than that of white people.

As mentioned, the FBI does publish statistics on “justifiable homicide” by law enforcement officers: The data show that there have been about 400 such incidents nationwide each year. However, FiveThirtyEight, among other journalism outlets, has examined the potential problems with these figures. News investigations suggest that the rates of deadly force usage are far from uniform. For example, Los Angeles saw an increase in such incidents in 2011, while Massachusetts saw more officers firing their weapon over the period 2009-2013.

The Bureau of Justice Statistics did publish a report in 2016 that found that about 1,900 people had died while in police custody  during the prior year. That report, which offered details about the cause of death during a three-month period, found that nearly two-thirds of deaths in police custody between June and August of 2015 were homicides — including justifiable homicides by a law enforcement officer — while nearly one-fifth were suicides and just over one-tenth were accidental deaths.

The academic community has also provided some insights in this area. A 2008 study from Matthew J. Hickman of Seattle University, Alex R. Piquero of the University of Maryland and Joel H. Garner of the Joint Centers for Justice Studies reviewed some of the best studies and data sources available to come up with a more precise national estimate for incidents of non-lethal force. They note that among 36 different studies published since the 1980s, the rates of force asserted vary wildly, from a high of more than 30 percent to rates in the low single digits. The researchers analyze Police-Public Contact Survey (PPCS) data and Bureau of Justice Statistics Survey of Inmates in Local Jails (SILJ) data and conclude that an estimated 1.7 percent of all contacts result in police threats or use of force, while 20 percent of arrests do.

A 2012 study in the Criminal Justice Policy Review analyzed the patterns of behavior of one large police department — more than 1,000 officers — and found that a “small proportion of officers are responsible for a large proportion of force incidents, and that officers who frequently use force differ in important and significant ways from officers who use force less often (or not at all).” A 2007 study in Criminal Justice and Behavior, “Police Education, Experience and the Use of Force,” found that officers with more experience and education may be less likely to use force, while a review of case studies suggests that specific training programs and accountability structures can lower the use of violence by police departments.

A 2016 working paper from the National Bureau of Economic Research (NBER) came to a conclusion that surprised some observers. Across the U.S., though blacks are 21.3 percent more likely to be involved in an altercation with police where a weapon is drawn, the researchers found no racial differences in police shootings: “Partitioning the data in myriad ways, we find no evidence of racial discrimination in officer-involved shootings. Investigating the intensive margin – the timing of shootings or how many bullets were discharged in the endeavor – there are no detectable racial differences.”

Researchers continue to refine analytical procedures in order to make more accurate estimates based on police reports and other data. 

Characteristics of suspects

A widely publicized report in October 2014 by ProPublica concluded that young black males are 21 times more likely to be shot by police than their white counterparts: “The 1,217 deadly police shootings from 2010 to 2012 captured in the federal data show that blacks, age 15 to 19, were killed at a rate of 31.17 per million, while just 1.47 per million white males in that age range died at the hands of police.”

Research has definitively established that “racial profiling” by law enforcement exists — that persons of color are more likely to be stopped by police. FBI Director James Comey’s 2015 comments are again relevant here:

[P]olice officers on patrol in our nation’s cities often work in environments where a hugely disproportionate percentage of street crime is committed by young men of color. Something happens to people of good will working in that environment. After years of police work, officers often can’t help but be influenced by the cynicism they feel.

A mental shortcut becomes almost irresistible and maybe even rational by some lights. The two young black men on one side of the street look like so many others the officer has locked up. Two white men on the other side of the street—even in the same clothes—do not. The officer does not make the same association about the two white guys, whether that officer is white or black. And that drives different behavior. The officer turns toward one side of the street and not the other. We need to come to grips with the fact that this behavior complicates the relationship between police and the communities they serve.

While the cases of Rodney King in 1991 and Amadou Diallo in 1999 heightened the country’s awareness of race and policing, research has not uniformly corroborated the contention that minorities are more likely, on average, to be subject to acts of police force than are whites. A 2010 paper published in the Southwestern Journal of Criminal Justice reviewed more than a decade’s worth of peer-reviewed studies and found that while many studies established a correlation between minority status and police use of force, many other studies did not — and some showed mixed results.

Of note in this research literature is a 2003 paper, “Neighborhood Context and Police Use of Force,” that suggests police are more likely to employ force in higher-crime neighborhoods generally, complicating any easy interpretation of race as the decisive factor in explaining police forcefulness. The researchers, William Terrill of Northeastern University and Michael D. Reisig of Michigan State University, found that “officers are significantly more likely to use higher levels of force when encountering criminal suspects in high crime areas and neighborhoods with high levels of concentrated disadvantage independent of suspect behavior and other statistical controls.” Terrill and Reisig explore several hypothetical explanations and ultimately conclude:

Embedded within each of these potential explanations is the influence of key sociodemographic variables such as race, class, gender, and age. As the results show, when these factors are considered at the encounter level, they are significant. However, the race (i.e., minority) effect is mediated by neighborhood context. Perhaps officers do not simply label minority suspects according to what Skolnick (1994) termed “symbolic assailants,” as much as they label distressed socioeconomic neighborhoods as potential sources of conflict.

In studying the Seattle and Miami police departments, the authors of the 2010 National Institute of Justice report also conclude that “non-white suspects were less likely to be injured than white suspects … where suspect race was available as a variable for analysis. Although we cannot speculate as to the cause of this finding, or whether it is merely spurious, it is encouraging that minority suspects were not more likely to be injured than whites.”

Use of Tasers and other “less lethal” weapons

A 2011 report from the National Institute of Justice, “Police Use of Force, Tasers and Other Less-Lethal Weapons,” examines the effectiveness and health outcomes of incidents involving CEDs (conducted energy devices), the most common of which is the Taser. The report finds that: (1) Injury rates vary widely when officers use force in general, ranging from 17 percent to 64 percent for citizens and 10 percent to 20 percent for officers; (2) Use of Tasers and other CEDs can reduce the statistical rate of injury to suspects and officers who might otherwise be involved in more direct, physical conflict — an analysis of 12 agencies and more than 24,000 use-of-force cases “showed the odds of suspect injury decreased by almost 60 percent when a CED was used”; and (3) A review of fatal Taser incidents found that many involved multiple uses of the device against the suspect in question.

A 2011 study, “Changes in Officer Use of Force Over Time: A Descriptive Analysis of a National Survey,” documents trends in the use of non-lethal force by law enforcement officers (LEAs). The results indicate that CED use has risen significantly (to about 70 percent of LEAs), while baton use is down to 25 percent in 2008. “CED use was ranked among the most-used tactics from 2005 to 2008,” the scholars conclude. “Excessive-force complaints against LEAs, internally generated, have more than doubled from 2003 to 2008. Officer injuries varied little from 2003 to 2008, but they are still only about half as common as suspect injuries. Also, only 20 percent of LEAs collect injury data in a database, complicating future research.”

Meanwhile, a 2018 study published in the Security Journal , “Smart Use of Smart Weapons: Jail Officer Liability for the Inappropriate Use of Tasers and Stun Guns on Pretrial Detainees,”  offers insights on how some these weapons are used on individuals who are incarcerated while awaiting trial. The paper demonstrates that although the U.S. Supreme Court has ruled it inappropriate to use tasers on pretrial detainees who do not follow verbal commands, the practice is common. It suggests that correctional officers should be reminded “of the distinction between convicted inmates and pretrial detainees who cannot be punished, so that they can abide by the recent Supreme Court decision and avoid liability in the future.”

Potential impact of body cameras

Video recordings of interactions between the police and the public have increased significantly in recent years as technology has improved and the number of distribution channels has expanded. Any standard smartphone can now make a video — as was the case in the Walter L. Scott shooting — and dash-mounted cameras in police cars have become increasingly common.

The mandatory adoption of body cameras by police has been suggested to increase transparency in interactions between law-enforcement officials and the public. A 2014 study from the U.S. Department of Justice, “Police Officer Body-Worn Cameras: Assessing the Evidence,” reviews available research on the costs and benefits of body-worn camera technology. The author, Michael D. White of Arizona State University, identified five empirical studies on body cameras, and assesses their conclusions. In particular, a year after the Rialto, Calif., police deparment began requiring all officers to wear body cameras, use of force by officers fell by 60 percent and citizen complaints dropped by nearly 90 percent. The searcher notes:

The decline in complaints and use of force may be tied to improved citizen behavior, improved police officer behavior, or a combination of the two. It may also be due to changes in citizen complaint reporting patterns (rather than a civilizing effect), as there is evidence that citizens are less likely to file frivolous complaints against officers wearing cameras. Available research cannot disentangle these effects; thus, more research is needed.

The studies also noted concerns about the cost of the required devices, training and systems for storing video footage; potential health and safety effects; and especially privacy concerns, both for citizens and the police. In April 2015, a bill being considered in the Michigan State legislature would exempt some body-camera footage from the state’s Freedom of Information (FOI) laws. Those who spoke in favor of the law included a conservative Republican legislator and an ACLU representative.

Public opinion and media

The coverage of such incidents by mass media has been studied by researchers, some of whom have concluded that the press has often distorted and helped justify questionable uses of force. Finally, survey data continue to confirm the existence of undercurrents of racism and bias in America, despite demonstrable social progress; a 2014 Stanford study shows how awareness of higher levels of black incarceration can prompt greater support among whites for tougher policing and prison programs.

Keywords: crime, local reporting, racism, violence, police enforcement, police brutality, body cameras, technology , policing

About the Authors

John Wihbey

Leighton Walter Kille

PODCAST: The FRONTLINE Dispatch

Inside the investigation into police use of force.

More places to listen

A new investigation reveals that over nearly a decade, more than 1,000 people died following encounters where police employed tactics known as “less-lethal force,” which ranged from Tasers or physical restraint to forced sedation and other methods meant to stop people without killing them. Police say they are often responding to volatile and sometimes violent situations, and deaths are rare.

Drawing on police records, autopsy reports, and footage from cellphones and body-worn cameras, The Associated Press, in collaboration with FRONTLINE and the Howard Centers for Investigative Journalism, compiled a database that serves as the most extensive accounting ever of deaths following such police encounters.

Serginho Roosblad, director and producer of the joint documentary Documenting Police Use of Force , and Justin Pritchard, a reporter and editor with the AP, join host Raney Aronson-Rath on The FRONTLINE Dispatch to discuss their findings.

The investigation also includes an interactive story and database .

Stream Documenting Police Use of Force on FRONTLINE’s website , FRONTLINE’s YouTube channel , or the PBS App .

Want to be notified every time a new podcast episode drops? Sign up for The FRONTLINE Dispatch newsletter .

More From this Series

Sign Up For The FRONTLINE Dispatch Newsletter

More podcasts.

Un(re)solved

I’m Not A Monster

FRONTLINE Film Audio Track

Follow frontline.

The FRONTLINE Dispatch is made possible by the Abrams Foundation and by the GBH Catalyst Fund. Additional support for The FRONTLINE Dispatch comes from the Massachusetts General Hospital Cancer Center.

• Apple Podcasts
• Spotify Podcasts
• RadioPublic
• Google Podcasts

FRONTLINE Newsletter

We answer to no one but you.

You'll receive access to exclusive information and early alerts about our documentaries and investigations.

I'm already subscribed

The FRONTLINE Dispatch

Don't miss an episode. sign-up for the frontline dispatch newsletter., sign-up for the unresolved newsletter..

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

•  We're Hiring!
•  Help Center

Use of Force

• Most Cited Papers
• Most Downloaded Papers
• Newest Papers
• Save to Library
• Last »
• Use of force in international law and international relations Follow Following
• Law of Armed Conflict Follow Following
• State Responsibility Follow Following
• Military Law Follow Following
• Irregular Warfare Follow Following
• The Cyprus Problem Follow Following
• National Security Law Follow Following
• International Courts And Tribunals Follow Following
• Less Lethal Weapon Follow Following
• Police Reform Follow Following

Enter the email address you signed up with and we'll email you a reset link.

• Academia.edu Publishing
•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024

Female labor force participation

Across the globe, women face inferior income opportunities compared with men. Women are less likely to work for income or actively seek work. The global labor force participation rate for women is just over 50% compared to 80% for men. Women are less likely to work in formal employment and have fewer opportunities for business expansion or career progression. When women do work, they earn less. Emerging evidence from recent household survey data suggests that these gender gaps are heightened due to the COVID-19 pandemic.

Women’s work and GDP

Women’s work is posited to be related to development through the process of economic transformation.

Levels of female labor force participation are high for the poorest economies generally, where agriculture is the dominant sector and women often participate in small-holder agricultural work. Women’s participation in the workforce is lower in middle-income economies which have much smaller shares of agricultural activities. Finally, among high-income economies, female labor force participation is again higher, accompanied by a shift towards a service sector-based economy and higher education levels among women.

This describes the posited  U-shaped relationship  between development (proxied by GDP per capita) and female labor force participation where women’s work participation is high for the poorest economies, lower for middle income economies, and then rises again among high income economies.

This theory of the U-shape is observed globally across economies of different income levels. But this global picture may be misleading. As more recent studies have found, this pattern does not hold within regions or when looking within a specific economy over time as their income levels rise.

In no region do we observe a U-shape pattern in female participation and GDP per capita over the past three decades.

Structural transformation, declining fertility, and increasing female education in many parts of the world have not resulted in significant increases in women’s participation as was theorized. Rather, rigid historic, economic, and social structures and norms factor into stagnant female labor force participation.

Historical view of women’s participation and GDP

Taking a historical view of female participation and GDP, we ask another question: Do lower income economies today have levels of participation that mirror levels that high-income economies had decades earlier?

The answer is no.

This suggests that the relationship of female labor force participation to GDP for lower-income economies today is different than was the case decades past. This could be driven by numerous factors -- changing social norms, demographics, technology, urbanization, to name a few possible drivers.

Gendered patterns in type of employment

Gender equality is not just about equal access to jobs but also equal access for men and women to good jobs. The type of work that women do can be very different from the type of work that men do. Here we divide work into two broad categories: vulnerable work and wage work.

The Gender gap in vulnerable and wage work by GDP per capita

Vulnerable employment is closely related to GDP per capita. Economies with high rates of vulnerable employment are low-income contexts with a large agricultural sector. In these economies, women tend to make up the higher share of the vulnerably employed. As economy income levels rise, the gender gap also flips, with men being more likely to be in vulnerable work when they have a job than women.

From COVID-19 crisis to recovery

The COVID-19 crisis has exacerbated these gender gaps in employment. Although comprehensive official statistics from labor force surveys are not yet available for all economies,  emerging studies  have consistently documented that working women are taking a harder hit from the crisis. Different patterns by sector and vulnerable work do not explain this. That is, this result is not driven by the sectors in which women work or their higher rates of vulnerable work—within specific work categories, women fared worse than men in terms of COVID-19 impacts on jobs.

Among other explanations is that women have borne the brunt of the increase in the demand for care work (especially for children). A strong and inclusive recovery will require efforts which address this and other underlying drivers of gender gaps in employment opportunities.

• Research and Publications
• Projects and Operations
• Privacy Notice
• Site Accessibility
• Access to Information
• REPORT FRAUD OR CORRUPTION