Race and criminal justice policy in America presents criminology with both its most consequential empirical questions and its most politically charged policy debates — a domain where the research evidence is extensive, the stakes for millions of people are immediate, and the distance between what the evidence establishes and what public discourse acknowledges is among the largest in the discipline. Racial disparities in criminal justice contact, processing, and outcomes are among the most thoroughly documented findings in American criminology: Black Americans are arrested at approximately twice their share of the population, convicted at higher rates than comparable white defendants, sentenced to longer terms for comparable offenses, and incarcerated at approximately five times the rate of white Americans. These disparities have been documented across datasets, jurisdictions, time periods, and crime types with a consistency that statistical artifact cannot explain and that constitutes one of the field’s most robust empirical findings.
Criminology and Public Policy engages race and criminal justice policy as a domain requiring both empirical precision and normative clarity — empirical precision about what the research establishes regarding the sources and mechanisms of racial disparities, and normative clarity about what standards of justice are relevant to evaluating those disparities and what they imply for policy. The debate about whether racial disparities in criminal justice reflect racial discrimination, differential criminal behavior, or the accumulated consequences of historical and structural racism is not merely academic — it determines whether addressing those disparities requires reform within the criminal justice system, social investment addressing structural inequality, or both. The research evidence, examined carefully, supports a substantially more complex answer than any of these positions alone.
Introduction
Understanding racial disparities in criminal justice requires distinguishing among the multiple sources that contribute to them and the multiple stages of the criminal justice process at which they accumulate. Disparities in arrest rates reflect both police deployment decisions — which concentrate patrol in high-crime communities that are disproportionately Black — and actual differences in criminal conduct rates that are themselves the product of the concentrated disadvantage, historical discrimination, and limited opportunity that has shaped the social conditions of many Black communities. Disparities in prosecution, conviction, and sentencing reflect both the legally relevant factors (prior record, offense severity) that predict harsher treatment for defendants with more extensive criminal histories — which Black defendants disproportionately have as a result of more intensive prior enforcement — and the legally irrelevant factors (implicit bias, social distance between prosecutor or judge and defendant) that research consistently finds influence criminal justice outcomes net of legally relevant characteristics.
The academic debate between those who emphasize racial discrimination within the criminal justice system and those who emphasize differential criminal behavior as the primary driver of racial disparities reflects a genuine empirical complexity that neither position fully captures. The most careful research — controlling for legally relevant case characteristics, decomposing disparities across justice system stages, and using audit study and implicit bias methodologies to identify discrimination — finds evidence for racial discrimination at multiple points in the criminal justice process alongside evidence that some portion of arrest disparities reflects actual conduct differences that are themselves products of structural racial inequality. The policy implications of this complex picture require reforming criminal justice processes that embed discriminatory practices while simultaneously addressing the structural racial inequality that elevates criminal justice risk in Black communities through pathways that individual-level anti-discrimination measures cannot address.
Three policy domains that the research most clearly implicates for reform are police deployment and use of force, prosecutorial charging and plea bargaining, and sentencing policy — each of which shows racial disparities whose magnitude, consistency, and independence from legally relevant factors provides sufficient evidence for the discrimination hypothesis to motivate reform regardless of the total quantitative contribution of discrimination to overall racial disparities.
Racial Disparities in Policing
Stop and Frisk, Surveillance, and Police Deployment
Racial disparities in police stops — the most frequently studied manifestation of racially unequal policing — have been documented in virtually every city that has made stop data publicly available and in every national survey of police-public encounters. The New York City stop-and-frisk program at its peak — when police conducted approximately 685,000 stops in 2011 — stopped Black and Latino New Yorkers at rates approximately five times their white counterparts, with the overwhelming majority of stops producing no arrest and no contraband, consistent with a program driven by racial profiling rather than individualized reasonable suspicion (Floyd v. City of New York, 2013). Federal Judge Scheindlin’s ruling finding the program unconstitutional under the Fourth and Fourteenth Amendments — a ruling that also found the New York City Police Department had engaged in a pattern of indirect racial profiling — represented the most significant federal court intervention in racially disparate policing in the stop-and-frisk era.
Research by Jonathan Mummolo and others using administrative data on police deployments has documented that Black neighborhoods receive more intensive patrol than the crime rates of those neighborhoods would predict, independent of residents’ preferences for police presence and service — a finding consistent with the hypothesis that racial considerations influence patrol deployment beyond the crime-based rationale that deployment decisions nominally reflect. Research by Roland Fryer using the Police-Public Contact Survey and Houston Police Department data reached more contested conclusions about racial disparities in officer-involved shootings, finding disparities in non-lethal uses of force but not lethal force — findings that generated substantial methodological criticism and illustrate the genuine complexity of studying racial discrimination in policing contexts where the relevant comparison group (who would have been stopped or shot in a race-neutral policing environment) is not directly observable.
Use of Force and Police Killings
Research on racial disparities in police killings — which the Washington Post, The Guardian, and the Mapping Police Violence project have documented more comprehensively than the FBI’s previously incomplete statistics — consistently finds that Black Americans are killed by police at approximately 2.5–3 times the rate of white Americans. The interpretation of this disparity is contested in the research literature because the appropriate benchmark — the population from which lethal force encounters are drawn — can be operationalized as the general population (supporting the finding of large racial disparities), the population of people who have police contact (reducing but not eliminating the disparity), or the population of people engaged in violent encounters with police (where some studies find limited residual disparities). The methodological choice among these benchmarks reflects theoretical positions about the locus of racial disparities — whether they emerge primarily in who has police contact or in how those who have contact are treated — that the data alone cannot resolve.
The Mapping Police Violence database’s finding that unarmed Black Americans are killed by police at substantially higher rates than unarmed white Americans — a comparison that controls for the armed status of the individual at the time of the fatal encounter — provides evidence most directly relevant to the question of whether racial discrimination in the application of lethal force is occurring independent of criminal threat, since armed status is among the most important proximate determinants of whether lethal force is legally justified. The finding that even among individuals presenting no weapon, racial disparities in fatal police encounters persist suggests racial discrimination in the deployment of lethal force beyond what legally relevant threat characteristics can explain.
Table 1. Racial Disparities in U.S. Criminal Justice: Selected Evidence
| Stage | Disparity Measure | Black/White Ratio | Data Source | Key Finding |
|---|---|---|---|---|
| Arrest | Arrest rate per 100,000 | ~2:1 overall; ~3:1 violent crime | FBI UCR | Disparities largest for drug and violent offenses; vary substantially by jurisdiction |
| Marijuana arrest | Arrest rate | ~3.73:1 | ACLU (2020) | Persists despite similar use rates; reduced in legalization states |
| Police stops | Stop rate per 100,000 | ~5:1 (NYC 2011 peak) | Floyd v. NYC; multiple city datasets | Low hit rates suggest profiling rather than crime-based selection |
| Pretrial detention | Bail denial / detention rate | ~1.4:1 controlling for charge | Various; Dobbie et al. (2018) | Racial gap partly but not fully explained by legally relevant factors |
| Conviction | Conviction rate | ~1.2:1 controlling for charge | Federal data; state studies | Smaller disparity than arrest; partially explained by prior record |
| Sentencing | Sentence length controlling for charge | ~1.19:1 federal (USSC 2017) | U.S. Sentencing Commission | Persistent disparity after legal controls; partly explained by criminal history |
| Incarceration | Incarceration rate per 100,000 | ~5:1 | Bureau of Justice Statistics | Reflects cumulative disparities from all upstream stages |
| Death by police | Killings per 100,000 | ~2.5:1 overall; higher for unarmed | Mapping Police Violence; Washington Post | Unarmed disparity more difficult to explain by legally relevant factors |
Implicit Bias and Structural Racism in Criminal Justice
Psychological Mechanisms of Racial Disparity
The mechanisms through which racial disparities are produced in criminal justice decision-making include both explicit racial discrimination — conscious decisions motivated by racial animus — and the implicit biases and structural features that produce racial disparities in the absence of conscious discriminatory intent. Research in social psychology has documented that implicit racial associations — the unconscious mental associations between Black individuals and concepts of danger, criminality, and threat — influence perceptual and behavioral responses in ways that decision-makers are not aware of and do not experience as racial bias. Studies by Jennifer Eberhardt and colleagues at Stanford University have documented that police officers more quickly categorize crime-related objects when primed with Black faces than white faces, that individuals more often see weapons in the hands of unarmed Black targets than unarmed white targets, and that implicit race-crime associations predict real-world criminal justice decisions in samples of practicing officers and judges (Eberhardt et al., 2004).
The policy implications of implicit bias research are contested: if racial disparities are partly produced by unconscious mental processes that bias training can address, then implicit bias training for police officers, prosecutors, and judges is a legitimate reform. But research on the effectiveness of implicit bias training has found at best modest short-term effects on measured implicit bias and no demonstrated effects on actual decision-making outcomes, suggesting that awareness-based training may be insufficient to produce the behavioral change that reducing discrimination requires. The more structurally oriented policy response — removing individual discretion at high-disparity decision points by requiring standardized criteria, documentation of reasoning, and algorithmic or committee review — reduces the opportunity for both conscious and unconscious bias to operate, though at the cost of the individualized justice that discretion is intended to enable.
Historical Legacies and Contemporary Consequences
The contemporary racial disparities in criminal justice cannot be understood without the history of racially motivated law enforcement that produced them: the enforcement of Black codes and vagrancy laws against Black Americans in the post-Civil War South, which criminalized normal economic activity by Black workers; the use of convict leasing systems that effectively re-enslaved Black men through criminal prosecution for minor offenses; the racially selective enforcement of alcohol prohibition and narcotics laws that established the pattern of racially concentrated drug law enforcement that the War on Drugs replicated and amplified; and the systematic exclusion of Black Americans from the civic, economic, and geographic opportunities that would have reduced criminal justice risk through the labor market, housing, and social capital pathways that criminological research has documented.
This historical account is not merely a morally important background to contemporary research — it is an empirically necessary part of the explanation for contemporary racial disparities, because many of the legally relevant factors that appear to justify race-neutral criminal justice disparities — prior criminal records, residential location, socioeconomic status — are themselves products of this history. When prior record predicts harsher treatment and Black defendants have more extensive prior records because historical enforcement concentrated on Black communities, using prior record as a legally relevant sentencing factor reproduces historical racial discrimination even in individual decisions that are individually race-neutral. This argument — that facially neutral criminal justice criteria can constitute structural racial discrimination when they encode the consequences of prior historical discrimination — is one of the most important theoretical contributions of critical race theory to criminal justice policy debate.
Prosecutorial Discretion and Racial Disparities
Charging, Pleading, and the Race of the Defendant
Prosecutorial discretion — the broad authority of prosecutors to decide whom to charge, what to charge, whether to offer plea agreements, and on what terms — creates substantial opportunity for racial disparities to enter criminal justice outcomes at a stage with minimal external oversight and accountability. Research documenting racial disparities in prosecutorial decision-making has found that Black defendants are more likely to be charged with offenses carrying mandatory minimum sentences and less likely to benefit from prosecutorial motions for sentence reductions than comparable white defendants, with disparities that persist after controlling for legally relevant case characteristics including offense severity and prior criminal history (Rehavi & Starr, 2014). The United States Sentencing Commission‘s analyses of federal sentencing data have consistently documented racial disparities in sentence length that are only partially explained by legally relevant factors, with unexplained residuals attributable either to racial discrimination in the exercise of prosecutorial and judicial discretion or to unmeasured legally relevant factors that the available data cannot capture.
The plea bargaining system — through which approximately 97% of state and 90% of federal convictions are obtained without trial — creates particularly acute racial fairness concerns because the discretionary nature of plea offers, the information asymmetries between prosecutors and defendants, and the coercive pressures created by the sentencing disparity between trial conviction and guilty plea all operate in contexts where racial disparities in the quality of legal representation amplify any existing racial disparities in prosecutorial willingness to offer favorable terms. Research by Cassia Spohn and others has found that the race of the defendant influences plea offer quality even after controlling for legally relevant case characteristics, with Black defendants receiving less favorable plea terms than comparable white defendants in some jurisdictions, contributing to the racial disparities in conviction and sentencing that accumulate from the charging stage forward.
Prosecutorial Reform and Accountability
The election of reform-oriented prosecutors in jurisdictions including Philadelphia, Los Angeles, San Francisco, and Chicago — many of whom explicitly ran on platforms of reducing racial disparities in prosecution and incarceration — has created natural experiments in prosecutorial reform whose crime and racial equity consequences criminologists are actively studying. Research on the consequences of Philadelphia District Attorney Larry Krasner’s sentencing reform policies — including a significant reduction in the use of cash bail and in prosecution of low-level drug offenses — found no significant increases in crime alongside substantial reductions in incarceration rates, providing early evidence that prosecutorial reform can achieve racial equity goals without the public safety costs that opponents predicted (Doherty, 2019).
The Fair and Just Prosecution network, which convenes elected prosecutors committed to reform, and the Vera Institute of Justice‘s work on prosecutorial accountability have been instrumental in developing both the reform agenda and the evidence base for prosecutorial change. The growing investment in prosecutorial data infrastructure — tools that enable prosecutors to analyze their own charging and disposition data for racial disparities — represents a transparency accountability mechanism that the discretionary nature of prosecutorial decision-making has historically lacked.
Sentencing Policy and Racial Equity
Mandatory Minimums, Sentencing Guidelines, and Race
The sentencing policies most directly associated with the mass incarceration era — mandatory minimum sentences, sentencing enhancements, and prosecutorial guidelines that constrain judicial discretion in ways that increase sentence severity — have been disproportionately applied to offense categories concentrated in Black communities, producing racial disparities in incarceration that persist even when individual charging and sentencing decisions are race-neutral. The crack-powder cocaine sentencing disparity — which required equal weights of crack and powder cocaine to trigger identical mandatory minimums at a 100:1 ratio, partially corrected to an 18:1 ratio by the Fair Sentencing Act of 2010 and further corrected to 1:1 in prospective application by the First Step Act of 2018 — is the most documented example of a sentencing policy producing racially disparate outcomes through its differential application to offense types concentrated in different racial communities.
The First Step Act of 2018 — the most significant federal criminal justice reform legislation in decades, enacted with bipartisan support — reduced mandatory minimum sentences for certain drug offenses, expanded the safety valve allowing judges to sentence below mandatory minimums in appropriate cases, and applied the crack-powder cocaine corrections retroactively to individuals sentenced under the old disparity. Research evaluating the First Step Act’s implementation has documented meaningful reductions in sentence lengths for the affected populations, with racial equity consequences from the crack-powder correction that the United States Sentencing Commission has documented in its evaluation reports. The First Step Act’s limited scope — addressing federal sentencing while leaving state sentencing policies untouched — means that its racial equity consequences, while real and documented, affect only a fraction of the racial disparities in incarceration that exist across the full criminal justice system.
Reform Strategies: Evidence and Assessment
Police Reform Approaches
The reform strategies with the most direct evidence of reducing racial disparities in police conduct include several approaches that criminological and organizational research has evaluated with varying degrees of rigor. Body-worn cameras — which have been adopted by approximately 80% of large American police departments — have shown inconsistent effects on racial disparities in use of force and civilian complaints across evaluations, with the most rigorous randomized trials finding smaller effects than observational studies had suggested. Research by Ariel and colleagues in a large-scale police department randomized trial found significant reductions in use-of-force incidents and complaints in shifts with body cameras — but a subsequent meta-analysis of multiple randomized trials found more modest and inconsistent effects, suggesting that camera effectiveness depends on departmental implementation decisions including footage review practices, officer training, and accountability consequences for camera deactivation.
Civilian oversight boards — independent civilian bodies with authority to investigate police misconduct, review disciplinary decisions, and make policy recommendations — have been established in over 150 American cities and have been identified by reform advocates as mechanisms for external accountability that internal affairs investigations cannot provide. Research evaluating civilian oversight’s effectiveness at reducing racial disparities in police conduct has found modest positive effects in cities with strong civilian oversight models — those with genuine investigative authority, adequate staff, and access to departmental records — while finding little effect for boards without these structural features. The difference between strong and weak oversight models suggests that the institutional design of oversight, not merely its existence, determines its effectiveness as an accountability mechanism.
Consent decrees negotiated by the Department of Justice’s Civil Rights Division following pattern-and-practice investigations of police departments have been evaluated in research by MacDonald and colleagues finding reductions in excessive force complaints, civilian injuries, and in some cases reductions in crime rates following consent decree implementation. The consent decree evidence — while limited by evaluation methodology challenges that prevent clean causal attribution — is among the most promising available for comprehensive police reform that addresses both racial disparities and public safety simultaneously.
Community Perspectives and Legitimacy
The Ferguson Effect and Community Trust
The concept of the “Ferguson effect” — proposed following the 2014 police killing of Michael Brown in Ferguson, Missouri and the subsequent protests — posited that increased scrutiny of police conduct following high-profile killings would reduce police proactivity and produce crime increases. The research evidence for this hypothesis is mixed at best: while some studies found increases in violent crime following specific police-community flashpoint events, other rigorous evaluations found no such increase or found increases that were concentrated in the days immediately following events rather than sustained over time, and the causal attribution of crime trends to the Ferguson effect specifically requires separating it from the many other factors simultaneously affecting crime rates.
What the Ferguson effect debate obscured is the more important and more empirically robust finding that police killings of Black Americans reduce the willingness of Black community members to report crimes and cooperate with police investigations — findings documented by Desmond, Papachristos, and Kirk using shooting data as natural experiments in the consequences of police killings for community cooperation (Desmond et al., 2016). If police killings reduce crime reporting by deterring victimized community members from contacting police, the net crime consequence of racially unequal policing is not the zero-sum deterrence-legitimacy tradeoff that the Ferguson effect narrative posits but a compounding of harm: police killings produce both direct harm to the individuals killed and their families and indirect harm to community members victimized by the increased criminal impunity that reduced reporting enables.
The Abolition-Reform Debate
The sustained movement for abolition of police and prisons — associated with Angela Davis’s Are Prisons Obsolete?, Ruth Wilson Gilmore’s Golden Gulag, and the Movement for Black Lives’ policy platform — represents a challenge to the reform framework that has characterized most criminological engagement with racial disparities in criminal justice. Where reform advocates seek to reduce racial disparities within existing institutions through training, oversight, and accountability mechanisms, abolitionists argue that those institutions are constitutively racially oppressive — that they cannot be reformed into racially just institutions because racial inequality is not an incidental feature of their operation but a structural function they serve — and that genuine racial justice requires imagining and building alternative institutional arrangements for addressing harm. The criminological engagement with abolitionist thought has been uneven, with mainstream criminology largely treating abolition as a philosophical position rather than an empirical one, while critical criminologists have taken seriously the abolitionist argument that the evidence on racial disparities calls into question not only specific institutional practices but the institutional framework within which those practices occur.
The practical abolitionist agenda — building up what Gilmore calls “life-affirming institutions” through investment in housing, healthcare, education, and community infrastructure that reduce the conditions producing harm — is compatible with much of the empirical evidence on the structural determinants of crime reviewed throughout this series of articles, even for criminologists who do not accept the full abolitionist theoretical framework. The investment in community health, education, housing stability, and economic opportunity that the evidence supports as effective crime prevention is precisely the “life-affirming institutions” investment that abolitionists call for, suggesting a potential common ground between abolitionist and evidence-based reform frameworks at the level of policy investment even where theoretical disagreement about the appropriate long-term institutional framework persists.
Research on decarceration experiments — the reduction of incarceration in specific jurisdictions through sentencing reform, prosecutorial diversion, and supervision reform — has generally found that moderate reductions in incarceration rates do not increase crime rates, and in some cases are associated with crime reductions through the mechanisms that mass incarceration critics have predicted: reducing the criminogenic consequences of incarceration itself, reducing the social disorganization that concentrated incarceration produces in sending communities, and freeing resources for social investments with higher crime-prevention returns. The Vera Institute of Justice‘s analysis of jurisdictions that have reduced incarceration substantially over the past decade while maintaining or improving public safety outcomes provides the most systematic documentation of this pattern, making the empirical case that racial equity and public safety are more complementary than the “tough on crime” framework assumes.
Research on Second Look sentencing review — which allows individuals who have served 15–20 years of long sentences to petition courts for reconsideration of their sentences in light of their prison record and changed circumstances — has documented that individuals who receive second-look reviews have extremely low recidivism rates following release and that the people most affected by these reviews are disproportionately Black Americans sentenced under the mandatory minimum and three-strikes policies of the 1980s and 1990s. Second look legislation, adopted in the District of Columbia and under consideration in multiple states, represents a specific racial equity reform whose crime prevention costs are minimal — because the individuals eligible for review are by definition very long-serving and typically among the lowest recidivism risks in the prison population — and whose racial justice benefits are substantial.
The Marshall Project and the Sentencing Project‘s sustained data journalism and policy research on racial disparities in criminal justice has created the public information infrastructure that effective reform advocacy requires — documenting state-by-state variations in racial disparities, tracking the racial consequences of specific sentencing reforms, and communicating research findings to audiences that academic journals do not reach. Building this information infrastructure — making racial disparity data publicly available, easily interpretable, and regularly updated — is itself a reform with direct accountability consequences, because transparency about racial outcomes creates pressure on institutional actors to address disparities that opacity allows to persist unexamined.
Conclusion
The evidence on racial disparities in American criminal justice is clear enough that its honest summary requires direct statement: Black Americans are systematically treated more harshly than comparable white Americans at every stage of the criminal justice process, the disparities are only partially explained by legally relevant factors, and the unexplained residuals reflect a combination of discrimination, structural inequality, and the encoding of historical racial injustice in facially neutral legal criteria that individual-level anti-discrimination measures cannot adequately address. This is not a politically convenient conclusion for any side of the political debate about criminal justice reform — it requires both the reformers who emphasize discrimination and the conservatives who emphasize behavior as the primary driver to grapple with a more complex reality than either position fully acknowledges.
Racial disparities in American criminal justice are among the best-documented phenomena in criminological research and among the most consequential for the communities most directly affected. The evidence documents disparities at every stage of the criminal justice process — police contact, arrest, charging, bail, conviction, sentencing, and incarceration — that are partially but not fully explained by legally relevant factors, with the unexplained residuals attributable to discrimination, to unmeasured case characteristics, or both, in proportions that available methods cannot definitively determine for the overall system while individual studies clearly document discrimination at specific decision points.
Policy reform adequate to addressing racial disparities in criminal justice requires interventions at multiple levels: reforming police deployment, use of force, and stop-and-frisk practices that concentrate enforcement in Black communities beyond crime-rate justifications; reducing prosecutorial discretion that produces racially disparate charging and plea bargaining outcomes; reforming sentencing policies that produce racially disparate incarceration through their differential application to offense categories concentrated in Black communities; and addressing the structural racial inequality that elevates criminal justice risk in Black communities through pathways that criminal justice reform alone cannot address. These are complementary requirements, not competing alternatives — the evidence that structural inequality elevates criminal justice contact does not diminish the obligation to reform criminal justice processes that embed discriminatory practices, and the evidence that those processes embed discrimination does not diminish the obligation to address the structural conditions that produce the concentrated disadvantage that criminal justice enforcement targets.
References
- Alexander, M. (2010). The new Jim Crow: Mass incarceration in the age of colorblindness. New Press.
- 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
- Rehavi, M. M., & Starr, S. B. (2014). Racial disparity in federal criminal sentences. Journal of Political Economy, 122(6), 1320–1354. https://doi.org/10.1086/677255
- U.S. Sentencing Commission. (2017). Demographic differences in sentencing: An update to the 2012 Booker Report. https://www.ussc.gov/research/research-reports/demographic-differences-sentencing
- Western, B. (2006). Punishment and inequality in America. Russell Sage Foundation.