Urban policing in America — the practice and organization of law enforcement in cities, the strategies that departments deploy to reduce crime and maintain order, and the contested relationship between police and the communities they serve — occupies a central position in both urban criminology and American political life. It is the most visible expression of the state’s claim to authority over public space, the most frequent point of contact between citizens and criminal justice institutions, and the primary mechanism through which the concentrations of disadvantage and violence documented throughout this series of articles receive institutional response. Its effectiveness in reducing crime, its fairness in distributing enforcement attention, and its legitimacy in the eyes of the communities it most intensively polices are simultaneously among the most important questions in applied criminology and among the most contested in American political discourse.
Urban Criminology engages policing as both a practical and a theoretical topic: practical in that the evidence on which policing strategies reduce crime most cost-effectively has direct operational implications; theoretical in that policing is simultaneously a response to the social conditions that produce crime and a factor in the legitimacy dynamics and community trust relationships that affect crime through the compliance and cooperation mechanisms that procedural justice research has documented. Understanding urban policing requires understanding both what police do and what effects those actions have — on crime rates, on community trust, on racial equity, and on the long-term social conditions that determine whether policing addresses or compounds the concentrated disadvantage it most intensively encounters.
Introduction
American urban policing is organized through approximately 18,000 independent law enforcement agencies — city police departments, county sheriffs, transit authorities, park police, and other specialized agencies — that collectively employ approximately 700,000 full-time sworn officers. This organizational fragmentation is unique among comparable democracies, most of which organize policing through regional or national agencies with more standardized training, equipment, and accountability standards. American policing’s fragmentation produces extraordinary variation in police-community relations, training quality, tactical approaches, racial composition, and accountability structures across agencies of vastly different sizes, resources, and political environments.
The Bureau of Justice Statistics Law Enforcement Management and Administrative Statistics (LEMAS) survey provides the most comprehensive data on American police agency characteristics, documenting that the overwhelming majority of American police agencies are small — more than 80% employ fewer than 25 sworn officers — while a handful of large agencies in major cities employ the majority of the sworn officer workforce. The New York City Police Department with approximately 36,000 sworn officers is the largest in the country and has been the most intensively studied, but its scale and resources make it an atypical research site for the smaller departments that serve most Americans.
The evolution of American urban policing over the past five decades — from the community relations focus of the 1970s through the hard enforcement era of the 1980s and 1990s through the community policing movement of the Clinton era through the evidence-based policing movement of the 2000s and 2010s — reflects both changes in crime rates and public demands and changes in the research evidence about what works. The current period — characterized by the simultaneous recognition that police legitimacy is essential for crime control, that racial disparities in enforcement are both unjust and counterproductive, and that evidence-based operational strategies can substantially reduce crime — represents the most sophisticated state of police practice and police research in American history, even as it confronts the most intense public scrutiny and contested legitimacy that American policing has experienced in decades.
Evidence-Based Policing Strategies
Hot Spots Policing
Hot spots policing — the strategic concentration of patrol resources at the small number of locations responsible for disproportionate crime volume — is the most evidence-supported operational policing strategy available, with a body of experimental and quasi-experimental evaluation evidence that no other police tactic can match. The theoretical foundation — that crime concentrates at specific places in patterns stable enough to guide resource deployment — was established by the hot spots research reviewed in the urban crime patterns article; the operational application translates that concentration finding into a deployment strategy that concentrates patrol attention where crime actually occurs rather than distributing it proportionately across the beat geography that historical deployment practices have favored.
The evidence base for hot spots policing includes multiple randomized controlled trials — the gold standard of evaluation evidence — that find significant reductions in crime at targeted hot spots without proportional displacement to nearby areas. Lawrence Sherman and David Weisburd’s Minneapolis Hot Spots Patrol Experiment (1995) established the experimental foundation; subsequent randomized trials in Kansas City, Sacramento, Philadelphia, and multiple other cities have consistently found that directed patrol at hot spots reduces crime during patrol periods and for some time afterward, with average effect sizes suggesting 20–30% reductions in crime at treated hot spots relative to control hot spots.
The crime displacement concern — that hot spots policing merely moves crime to adjoining areas — has been extensively tested and largely refuted. Systematic review evidence finds that displacement occurs in some implementations but that “diffusion of benefits” — crime reductions spreading to areas adjacent to treated hot spots — is equally or more common than displacement, producing a positive net effect on total crime rather than merely geographic redistribution. The mechanism for diffusion of benefits — that potential offenders perceive elevated risk in a broader area than the specifically targeted locations — is consistent with deterrence theory’s prediction that uncertainty about enforcement locations deters beyond the specific locations where enforcement is concentrated.
Problem-Oriented Policing
Problem-oriented policing (POP) — Herman Goldstein’s proposal that police should analyze the underlying conditions generating recurring crime problems at specific locations rather than responding to calls for service and moving on — extends hot spots policing’s concentration on specific places by adding the analytical component that diagnoses and addresses the situational and social conditions enabling crime at those places. Where standard hot spots patrol concentrates enforcement at high-crime locations, POP uses the SARA model (Scanning, Analysis, Response, Assessment) to identify the specific feature — a drug market, a liquor establishment, a homeless encampment, a specific group of offenders — responsible for the crime concentration and to tailor responses to that specific feature.
Research comparing POP to standard patrol at hot spots has found that POP produces larger and more durable crime reductions than patrol alone, because problem-oriented responses modify the conditions enabling crime rather than simply displacing offenders during patrol periods. The Campbell Collaboration systematic review by Weisburd and colleagues found significant effects of problem-oriented policing across multiple studies and crime types, with the strongest evidence for property crime and drug market problems where the underlying conditions are most amenable to the environmental and situational modifications that POP prescribes. The Center for Problem-Oriented Policing at Arizona State University maintains the most comprehensive practitioner-accessible library of problem analysis guides and evaluation evidence available, making POP one of the best-supported in both evidence and practice guidance.
Table 1. Evidence-Based Policing Strategies: Summary
| Strategy | Core Mechanism | Evidence Quality | Average Effect Size | Key Limitation | Best Application |
|---|---|---|---|---|---|
| Hot spots patrol | Deterrence and incapacitation through concentrated patrol | Strong — multiple RCTs | 20–30% crime reduction at treated spots | Requires sustained deployment; some displacement | High-crime addresses and street segments |
| Problem-oriented policing | Diagnosis and removal of crime-enabling conditions | Strong — systematic review | Larger than patrol alone; more durable | Requires analytical capacity; SARA training | Recurring problems at specific locations |
| Focused deterrence (GVI) | Targeted deterrence + social service + community norms | Moderate-strong — quasi-experiments | 20–60% gang violence reduction | Requires law enforcement / community / service partnership | Gang-involved violence |
| Community policing | Trust-building; information sharing; legitimacy | Moderate — mixed results | Variable | Poorly defined; implementation varies widely | Legitimacy-deficit neighborhoods |
| Predictive policing | Algorithmic deployment optimization | Weak to moderate; contested | Modest when positive; equity concerns | Reproduces historical bias; limited crime reduction | Limited; equity oversight essential |
| Body cameras | Accountability; behavior change | Moderate — mixed RCTs | Modest and inconsistent | Implementation quality critical; not always effective | Use-of-force reduction; accountability |
Focused Deterrence in Policing
The Group Violence Intervention (GVI) and focused deterrence strategies that have shown the most compelling evidence for urban homicide reduction require a specific and often challenging form of police engagement: law enforcement agencies must credibly commit to swift and certain consequences for targeted individuals while simultaneously partnering with social service providers and community leaders rather than pursuing enforcement alone. This partnership model represents a significant departure from traditional law enforcement culture, which tends to see social service provision as outside the police role and community partnership as operationally compromising.
Research on the police component of focused deterrence — evaluating what aspects of the law enforcement engagement are most critical for producing violence reductions — has found that credibility is paramount: the enforcement promises made in group notification meetings must be kept, targeted individuals must see that the consequences communicated are actually delivered, and the social service offers must be genuine rather than nominal. Police departments that have implemented GVI with the highest fidelity — where a dedicated team of officers maintains sustained engagement with targeted individuals, tracks case outcomes, and coordinates with social services and community partners — have shown the largest violence reductions, while departments that treat GVI as an intermittent program rather than a sustained strategy show more modest and less durable effects.
Policing and the Crime Decline
Research on the contribution of policing to the dramatic American crime decline of the 1990s and 2000s has found evidence that police personnel levels and deployment strategies contributed meaningfully to falling crime rates, alongside the other factors reviewed in the urban violence article. Research by Chalfin and McCrary using instrumental variable methods found that additional police officers produced crime reductions with benefit-cost ratios substantially greater than one, establishing police hiring as a cost-effective crime control investment. Research on CompStat — the data-driven police management system pioneered in New York City in the early 1990s that systematically tracked crime at the precinct level and held commanders accountable for crime trends — found evidence that CompStat adoption was associated with crime reductions in implementing cities, suggesting that management and accountability systems that align officer behavior with crime reduction objectives produce better crime outcomes than management systems that do not.
The contribution of policing to the crime decline does not, however, establish that aggressive enforcement tactics were responsible. Research attempting to identify whether the New York stop-and-frisk program — which reached approximately 685,000 annual stops at its peak — contributed to New York’s crime decline has found little evidence that it did: a federal court found the program unconstitutional in 2013 and stops declined dramatically thereafter, but crime rates continued declining rather than increasing, establishing that the program was not responsible for the crime reductions attributed to New York policing during the period. This natural experiment — the near-elimination of stop-and-frisk without crime increases — has been one of the most important findings in the empirical case for police reform, demonstrating that a program whose racial disparities were documented and contested was not the mechanism responsible for crime reduction even while it was being defended on crime control grounds.
Community Policing and Legitimacy
Procedural Justice and Police Legitimacy
Tom Tyler’s procedural justice theory — which proposes that compliance with law and cooperation with legal institutions depends more on perceptions of fair and respectful treatment than on deterrence calculations based on sanction severity — has become the dominant framework for understanding police-community relations and its effects on crime. Research applying procedural justice theory to policing has documented that police are more likely to receive cooperation, crime information, and voluntary compliance with the law from citizens who perceive police treatment as fair, respectful, and unbiased — and less likely to receive it from citizens who perceive police as treating them as suspects rather than as community members deserving respectful service.
The implications for policing practice are direct: enforcement tactics that are perceived as disrespectful, racially biased, or indifferent to the dignity of the people being policed undermine the legitimacy foundation on which effective crime control depends. The aggressive order maintenance policing of the stop-and-frisk era — with New York City conducting approximately 685,000 stops in 2011, 88% involving Black or Latino individuals and over 90% resulting in no arrest — produced documented legitimacy deficits in affected communities that reduced crime reporting, hindered investigation of the very crimes the strategy purported to prevent, and contributed to the police-community tensions that the Floyd killing and subsequent protests made nationally visible.
Research by Desmond, Papachristos, and Kirk documenting that police killings of Black Americans substantially reduce crime reporting in Black communities provides the clearest empirical evidence for the mechanism through which illegitimate policing undermines crime control: not through any direct effect on criminal motivation but through the deterrence of victims from reporting crimes to police whose fairness and accountability they do not trust (Desmond et al., 2016). The policy implication — that racially equitable, procedurally just policing produces better crime control outcomes than racially disparate enforcement regardless of the short-term enforcement benefits of aggressive tactics — connects the justice and the effectiveness arguments for police reform in ways that neither argument alone achieves.
Alternative Crisis Response and Mental Health
The growing recognition that a substantial proportion of calls for police service involve mental health crises, substance use emergencies, and welfare checks that police are not optimally equipped to handle has driven the development of alternative crisis response systems — mobile crisis teams, co-responder models, and non-police dispatch programs — that address these calls through clinical rather than enforcement responses. Research on the CAHOOTS program in Eugene, Oregon — which dispatches mental health workers and medics to mental health calls without police — has found that the program successfully handles approximately 20% of all calls for service without police involvement, with better outcomes for the individuals served and no evidence of increased public safety risk.
The development of alternative crisis response reflects a broader recognition — consistent with the evidence-based policing framework — that police are not the optimal responders to every public safety problem and that deploying specialized responders for specialized problems produces better outcomes than deploying police as all-purpose first responders. The 988 Suicide and Crisis Lifeline national infrastructure, activated in 2022, and the growing network of mobile crisis teams connected to it represent the emerging policy infrastructure for diverting appropriate calls from police response to clinical response, with implications for both the individuals served and the police deployment capacity available for the crime response functions that law enforcement is uniquely equipped to perform.
Racial Disparities in Urban Policing
Documentation and Mechanisms
Racial disparities in police stops, searches, use of force, and arrest rates are among the most thoroughly documented findings in criminological research on policing, consistent across datasets, methodological approaches, and decades of study. Research by Roland Fryer, Jonathan Mummolo, and others using police administrative data has documented that Black Americans are stopped at rates approximately two to five times those of white Americans in comparable situations, that force is more likely to be used in encounters with Black suspects than with comparably situated white suspects for non-lethal force, and that police killings show consistent racial disparities even after controlling for neighborhood crime rates and encounter characteristics.
The mechanisms through which racial disparities in policing arise include both explicit racial discrimination — conscious decisions motivated by racial animus — and the implicit racial associations that cognitive psychology research has documented operate in ways that decision-makers do not experience as bias. Research by Jennifer Eberhardt and colleagues has documented that police officers more quickly categorize crime-related objects when primed with Black faces and more readily perceive weapons in the hands of unarmed Black targets, establishing that implicit racial bias affects perceptual processing in policing contexts in ways that produce discriminatory outcomes without conscious discriminatory intent.
Structural mechanisms also produce racial disparities independent of individual officer bias: patrol deployment concentrated in high-crime neighborhoods — which are disproportionately minority neighborhoods because of concentrated disadvantage — produces higher arrest rates for minority residents not because they are more criminal but because they are more policed. The enforcement of drug laws whose race-neutral application in racially unequal contexts produces racially disparate outcomes — Black Americans arrested for marijuana at approximately four times the rate of white Americans despite similar use rates — is the most documented example of this structural mechanism, and one that the legalization movement has addressed partly through its removal of the most dramatically disparate enforcement from the legal framework entirely.
The Ferguson Effect Debate
The “Ferguson effect” — the hypothesis that increased scrutiny of police following high-profile use-of-force incidents reduces police proactivity and produces crime increases — generated intense political debate following the 2014 Michael Brown killing in Ferguson, Missouri, and the subsequent protests. Research evaluating the Ferguson effect hypothesis has produced mixed results: some studies found short-term crime increases following high-profile killings in specific cities, while others found no significant increases or found increases too modest and brief to constitute a durable Ferguson effect. The most methodologically rigorous analyses — those using comparison groups, synthetic controls, or interrupted time series designs that account for secular trends — have generally found limited evidence for a sustained Ferguson effect on crime rates.
What research has found more consistently is that police killings of Black Americans reduce crime reporting in Black communities — through the legitimacy mechanism documented by Desmond, Papachristos, and Kirk — producing the paradoxical consequence that events that might be expected to produce crime increases (through reduced police aggressiveness) may actually produce them through a different mechanism (reduced community cooperation with police). This finding does not support the Ferguson effect narrative — it is not police de-policing that produces the crime increase but the community’s reduced willingness to seek police assistance following high-profile police killings — but it does establish that police conduct and community trust have bidirectional effects on crime outcomes that simple enforcement-centered accounts cannot capture.
Police Reform: Evidence on What Works
Consent Decrees and Their Consequences
Federal Department of Justice consent decrees — negotiated following pattern-and-practice investigations that find systemic constitutional violations in police departments — represent the most comprehensive externally-imposed police reform mechanism available and have been implemented in departments including Los Angeles, New Orleans, Seattle, and Ferguson. Research by Stephen Rushin and colleagues evaluating crime rates following consent decree implementation has found no evidence of crime increases in consent decree cities and some evidence of crime reductions, directly contradicting the “Ferguson effect” narrative that police reform undermines crime control.
The consent decree evidence is important because it addresses the central political objection to police reform — that accountability-oriented reform will produce crime increases by reducing police aggressiveness — with empirical evidence from actual reform implementations rather than theoretical arguments. If departments can reduce use of force, reduce racial disparities, and improve community relations through consent decree oversight without increasing crime rates, the political case for reform is substantially strengthened: the equity and legitimacy improvements that reform achieves come without the public safety cost that reform opponents predict.
Body-Worn Cameras
Body-worn cameras (BWCs) have been adopted by approximately 80% of large American police departments following high-profile police killings, with claims of both accountability and officer safety improvements. The evidence on BWCs’ effects on use of force and civilian complaints is more mixed than early claims suggested. The most rigorous randomized trials — including the Washington D.C. Metropolitan Police Department trial — found null effects on use of force in some implementations, while other trials found significant reductions. Meta-analyses suggest that BWC effects depend critically on implementation decisions — review policies, officer accountability for camera deactivation, public access to footage — rather than on the technology itself. Cameras without accountability policies do not reliably change officer behavior; cameras embedded in accountability frameworks show more consistent positive effects.
Predictive Policing: Promise and Peril
Predictive policing — the use of algorithmic tools that forecast where crimes are likely to occur or who is likely to commit them, allowing preemptive deployment of patrol resources — represents the most technologically ambitious application of hot spots policing’s concentration logic, and the one with the most serious equity concerns. Place-based predictive policing tools — which forecast high-crime locations based on historical crime patterns — are relatively well-supported by the hot spots evidence base: crime does concentrate stably at specific locations, and tools that identify those locations more precisely than analyst judgment alone should improve deployment efficiency. Person-based predictive policing — which forecasts which individuals are most likely to commit future crimes, drawing on prior criminal history, social network proximity to prior offenders, and other individual-level predictors — raises civil liberties concerns about pre-crime enforcement and algorithmic discrimination that place-based tools do not generate with the same force.
Research on the equity consequences of predictive policing algorithms has found that tools trained on historically biased arrest data reproduce and amplify the racial disparities in their training data, directing patrol resources toward minority communities at rates that exceed what crime rates alone would predict (Richardson et al., 2019). The feedback loop of prediction-driven enforcement — where prediction directs patrol to minority neighborhoods, which generates arrests that confirm the prediction and justify continued concentration — can produce self-fulfilling prophecies of racial disparity that are difficult to interrupt once established. The Los Angeles Police Department’s discontinuation of its PredPol predictive policing system in 2020 following advocacy and internal review reflects the practical conclusion that the equity costs of person-based prediction exceeded its demonstrated crime reduction benefits — a policy decision that the research literature supports even though the evidence base for specific tools remains limited.
Police Staffing, Deployment, and Crime
The relationship between the number of police officers and crime rates has been studied through multiple research designs that provide converging evidence of a positive crime-reduction effect from additional officers, particularly for violent crime. Research by Chalfin and McCrary using instrumental variable methods estimated that each additional police officer prevents approximately 0.3–0.5 violent crimes per year, producing benefit-cost ratios substantially above one given officer salary costs and the substantial welfare value of prevented violent crime. Research on the COPS program — the Clinton-era federal grant program that funded the hiring of approximately 100,000 additional officers — found consistent reductions in violent crime in cities that received COPS grants, with the effect concentrated in grants that funded community-oriented policing activities rather than traditional patrol enhancement alone.
The specific deployment of officers — what they do and where they go — matters as much as their number for crime control effectiveness. Research by Nathan Hess and colleagues comparing crime control outcomes across patrol strategies — directed patrol at hot spots versus undirected patrol across entire beats — found that directed patrol produced substantially larger crime reductions per officer-hour than undirected patrol, establishing that redeployment of existing officers toward evidence-based strategies can produce crime reductions comparable to those achievable through additional hiring. This finding is operationally important for the many departments whose budgets preclude significant personnel expansion but whose deployment practices do not reflect the best available evidence.
Conclusion
The evidence base for urban policing has matured substantially since the early experimental policing research of the 1970s and 1980s, progressing from the skeptical findings of the Kansas City Preventive Patrol Experiment — which found that varying patrol intensity had no effect on crime — through the hot spots revolution that established where directed patrol works, through the focused deterrence framework that established when and how specific deterrence produces violence reductions, to the current integrated framework that connects crime control effectiveness to procedural justice and community legitimacy. Each stage of this evolution has produced more specific, more actionable, and more theoretically grounded guidance for police practice than the stage before.
Urban policing in America is at an inflection point — a moment when the evidence for effective, equitable, and legitimate policing is stronger than at any previous point in the discipline’s history, while the gap between that evidence and actual police practice, and between the communities most affected by crime and the institutions supposed to serve them, is as wide as it has ever been politically. The evidence reviewed in this article consistently supports a set of conclusions: hot spots policing and problem-oriented approaches reduce crime more effectively than undirected patrol; focused deterrence with genuine community partnership reduces gang violence more effectively than enforcement alone; procedurally just policing that treats community members with dignity and fairness produces better cooperation and better long-term crime control than aggressive enforcement that prioritizes short-term deterrence over long-term legitimacy; and racial disparities in enforcement undermine both justice and effectiveness by eroding the community trust that effective policing requires.
Translating this evidence into changed practice requires overcoming both the organizational inertia of policing institutions shaped by decades of different priorities and the political pressures that reward tough-on-crime signals regardless of their empirical basis. The communities that bear the heaviest burden of urban crime — and who simultaneously experience the most intensive and least accountable policing — deserve both the protection that effective policing can provide and the dignity that procedurally just policing must express. Achieving both simultaneously, as the evidence increasingly suggests is possible, is the defining challenge for urban policing in the decades ahead. The articles that follow in this section address the specific dimensions of urban crime control to which policing contributes: the prevention strategies that address crime’s ecological conditions before police encounter them, and the policy frameworks that connect policing strategy to the broader urban criminology and criminal justice policy landscape. Effective urban policing is not policing in isolation — it is policing as part of an integrated approach to urban public safety that combines enforcement, prevention, community engagement, and structural investment in the proportions that the evidence supports rather than those that political convenience and institutional inertia have historically produced.
References
- Desmond, M., Papachristos, A. V., & Kirk, D. S. (2016). Police violence and citizen crime reporting in the Black community. American Sociological Review, 81(5), 857–876. https://doi.org/10.1177/0003122416663494
- Rushin, S., & Edwards, G. (2017). An empirical assessment of consent decrees. Stanford Law Review, 69, 1555–1618.
- Sherman, L. W., & Weisburd, D. (1995). General deterrent effects of police patrol in crime “hot spots”: A randomized, controlled trial. Justice Quarterly, 12(4), 625–648. https://doi.org/10.1080/07418829500093221
- Tyler, T. R. (1990). Why people obey the law. Yale University Press.
- Weisburd, D. (2015). The law of crime concentration and the criminology of place. Criminology, 53(2), 133–157. https://doi.org/10.1111/1745-9125.12070
- Weisburd, D., Telep, C. W., Hinkle, J. C., & Eck, J. E. (2010). Is problem-oriented policing effective in reducing crime and disorder? Criminology & Public Policy, 9(1), 139–172. https://doi.org/10.1111/j.1745-9133.2010.00617.x