To minimise confounding bias and disentangle warranted from unwarranted disparities, researchers examining sentencing discrimination have traditionally sought to control for as many legal factors as possible. However, over the past decade, a growing number of scholars have questioned this strategy, noting that many legal factors are themselves subject to judicial discretion and that controlling for them can introduce post-treatment bias. Here, we use directed acyclic graphs (DAGs) to provide a formal and comprehensive assessment of the different types of bias that may arise from different choices of controls. In addition, we propose a new modelling framework to facilitate the selection of controls and reflect the model uncertainty created by the trade-off inherent in judicially-defined legal factors and other factors with a similar dual causal role. We apply this framework to examine race disparities in US federal courts and gender disparities in the England and Wales magistrates’ court. We find substantial model uncertainty for gender disparities and for race disparities affecting Hispanic offenders, rendering estimates of the latter inconclusive. Disparities against black offenders are more consistent and — under specific conditions — could be interpreted as evidence of direct discrimination.

Artificial intelligence (AI) is increasingly utilized in criminal justice systems to support decisions related to bail, sentencing, and parole. However, these systems often perpetuate historical biases, particularly racial disparities embedded in the training data. Many existing models lack effective mechanisms for verifying fairness and mitigating bias, resulting in unequal outcomes for legally protected groups. This research aims to develop a structured machine learning (ML) framework that integrates Fairness Verification Algorithms with mitigation strategies to enhance fairness in criminal risk assessment. The framework employs Extreme Gradient Boosted K-Means Clustering (XGBoost-KMC), trained exclusively on data from White offenders, to minimize bias during model development. The dataset comprises 200,000 arraignment records from a metropolitan court system, including demographic and criminal history variables. Pre-processing steps involve removing sensitive attributes such as race and zip code, applying Z-score normalization, and addressing class imbalance using the Synthetic Minority Over-sampling Technique (SMOTE). Principal Component Analysis (PCA) is used for feature extraction, reducing dimensionality while preserving key predictive information. To mitigate bias, the model applies optimal transport to align the feature distribution of Black offenders with that of White offenders. It generates calibrated and uncertainty-aware risk forecasts using conformal prediction sets. Fairness Verification Algorithms assess the model’s outputs using statistical metrics such as prediction parity and classification parity across demographic groups. The model achieves an average accuracy of above 93% with 5-fold cross-validation, demonstrating both fairness and predictive capability. Results demonstrate that the proposed framework achieves a Pareto improvement, enhancing fairness for disadvantaged groups without compromising predictive accuracy. This approach provides a scalable and transparent solution for building equitable AI systems in criminal justice applications.

Life sentences without the possibility of parole (LWOP) are an important feature of the modern American penal system, but relatively little research examines the factors that shape the imposition of these sentences. In the federal system, life sentences—all of which are LWOP sentences due to the abolition of federal parole—often result from mandatory minimum sentencing statutes that require life imprisonment. We use USSC data from 2010 to 2017 to investigate the imposition and circumvention of mandatory LWOP sentences in federal court. We consider the role of structural inequalities in defining who is convicted under statutes requiring life sentences and examine who ultimately receives these sentences. We also investigate the unique role that substantial assistance departures play in circumventing mandatory minimum penalties, examine how they potentially contribute to racial inequality in life imprisonment, and assess how these processes differ for drug and non-drug crimes. Our results indicate that almost half of all individuals convicted under statutes requiring LWOP did not receive the mandatory life sentence; moreover, almost all (95%) of those who avoided the mandatory sentence received a substantial assistance departure. We also find that Black, but not Hispanic, offenders are more likely than White offenders to be convicted of an offense carrying a mandatory life sentence, that Black offenders who are eligible for a mandatory life term are more likely than similarly-situated White defendants to actually receive it, and that Black mandatory life offenders are less likely than their White counterparts to benefit from downward departures for providing substantial assistance. We conclude that substantial assistance departures are an important mechanism that prosecutors and judges use to “demandatorize” life sentences and that both mandatory minimums and substantial assistance departures shape patterns of inequality in the federal sentencing process, but we find limited evidence to suggest that these differences are due to the sentencing of federal drug offenses.

In the US context, critics of court use of algorithmic risk prediction algorithms have argued that COMPAS involves unfair machine bias because it generates higher false positive rates of predicted recidivism for black offenders than for white offenders. In response, some have argued that algorithmic fairness concerns, either also or only, calibration across groups–roughly, that a score assigned to different individuals by the algorithm involves the same probability of the individual having the target property across different groups of individuals–and that, for mathematical reasons, it is virtually impossible to equalize false positive rates without impairing the calibration. I argue that in standard non-algorithmic contexts, such as hirings, we do not think that lack of calibration entails unfair bias, and that it is difficult to see why algorithmic contexts, as it were, should differ fairness-wise from non-algorithmic ones in this respect. Hence, we should reject the view that calibration is necessary for fairness in an algorithmic context.

ABSTRACT During the COVID-19 pandemic, reports of violence against Asian Americans surged in traditional and social media, often depicting an Asian victim attacked by a Black assailant and suggesting heightened Black animosity toward Asian Americans. Scholars have challenged this Black-Asian conflict narrative by highlighting that White perpetrators commit over 75% of anti-Asian hate crimes. However, the large share of White offenders in anti-Asian hate crime may simply reflect American racial demographics, and Black offenders could be disproportionately victimizing Asian Americans relative to the size of the Black population or offender pool. The current study examined the share of Black offenders in violence against Asian victims relative to the racial composition of local populations or offender pools. Analysis using 2020 NIBRS and Census data indicated overrepresentation of Black offenders in violence against Asian victims relative to the weighted average of the Black population proportion of the 351 cities in which such incidents occurred. However, Black offenders’ share of violence against Asian victims closely matched the proportion Black in the average local potential offender pool. Comparison of 2019 and 2020 data indicated an increase in Black offenders’ share of violence against Asian victims, but this difference in the distributions was not statistically significant.

The Gladue report, named after R. v. Gladue, is a landmark Supreme Court of Canada case which emphasizes the need to consider unique circumstances faced by Indigenous individuals when determining appropriate sentences. Given the overrepresentation of Black identities at all levels in the justice system, it is argued that the use of pre-sentencing reports referred to as Impact of Race and Culture Assessments (IRCAs), also needs to be comprehensively implemented for Black offenders in Canada. IRCAs are pre-sentencing reports that help sentencing judges better understand the effect of poverty, marginalization, racism, and social exclusion on the offender and their life experiences, and how those factors inform the circumstances of the offender, the offence committed, and the offender’s experience with the justice system. This is significant as it goes beyond a one-size-fits-all punitive justice system that has been ineffective in reducing recidivism. By recognizing the intersections of race, culture, and justice, IRCAs enable judges to make more informed decisions contributing to an equitable consequence for the accused. More importantly, we argue that the insights from IRCAs should be used to connect offenders with culturally reflective wraparound social services upon return into the community to address the root causes in areas of employment, education, and housing that gravitate people towards criminality. By acknowledging historical and systemic biases and tailoring supports to individual identities, life experiences, and community conditions, IRCAs have the potential to transform the criminal justice system through promotion of alternatives to custody that correlates with reductions in recidivism.

While much is written about racialization of street criminals in the American media, racial dimensions of the media framing of white-collar crime remain underexplored. To address this issue, we analyze the coverage of bribery, electoral fraud, tax evasion, and insider trading in five national newspapers between 1950 and 2010. Drawing on John Hagan’s (2012) work, we trace the racialization of white-collar crime in the press back to Richard Nixon’s presidency and the beginnings of the War on Drugs. We also find that race is a significant predictor of offenders’ individualization, or the length of description accorded to them by writers. We argue that by individualizing black offenders significantly more than white perpetrators, reporters connote their oddity in the context of white-collar criminality and contribute to their collective framing as an exception. Finally, we find that black perpetrators receive significantly more positive coverage than white offenders, which serves to further underscore their distinctiveness from stereotypical black criminals and their similarity to nonthreatening (white) Americans. These findings support Hagan’s (2012) argument that racialization of street crime is mirrored by the collective framing of elite economic crime as white and, by extension, a nonthreatening side effect of American capitalism.

Abstract Public belief in redeemability reduces punitiveness and increases support for policy measures such as rehabilitation, expungement, and housing and employment opportunities. Although racial attitudes are known to influence a wide range of criminal justice policy opinions, their effects on beliefs about redeemability and condemnation have not been fully explored. Using data from a 2019 YouGov survey of a national sample of White U.S. adults (N = 766), the current study estimates the effects of three distinct racial attitudes—racial resentment, racial sympathy, and White nationalism—on three measures of belief in redeemability: 1) a race‐neutral measure, 2) a measure of belief in redeemability of Black offenders, and 3) a measure of condemnation of Black offenders. The results indicate that belief in redeemability is high—for offenders in general and for Black offenders. These findings are supported by a second 2022 YouGov survey of White U.S. adults (N = 1,505). Racial sympathy and White nationalism have significant effects across all three outcomes, with the positive effect of White nationalism on condemnation of Black offenders being the largest across the three models. These findings suggest that although most Whites agree that formerly incarcerated people are redeemable, racial attitudes influence these beliefs, especially for Black offenders.

ABSTRACT The historically complex relationship between Blacks and Whites is believed to play an important role in the way offenders select their victims in violent crimes. This study tests two theories that have been used to explain interracial crimes: racial animosity theory and racial threat theory. Using data from the National Crime Victimization Survey (NCVS) from 1992 through 2016, we conduct two analyses of the racial dyads of White and Black offenders and victims to examine the role racial characteristics play in interpersonal crimes. Multinomial logistic regression is used to first analyze the likelihood of weapon use before narrowing our analysis specifically to handguns use. The findings provide minimal support for both racial threat theory and racial animosity theory.

Systemic racism in data practices

Abstract We explore the annual number of death sentences imposed on black and white offenders within each US state from 1989 through 2017, with particular attention to the impact of aggregate levels of racial resentment. Controlling for general ideological conservatism, homicides, population size, violent crime, institutional and partisan factors, and the inertial nature of death sentencing behavior, we find that racial hostility translates directly into more death sentences, particularly for black offenders. Racial resentment itself reflects each state’s history of racial strife; we show powerful indirect effects of a history of lynching and of racial population shares. These effects are mediated through contemporaneous levels of racial resentment. Our findings raise serious questions about the appropriateness of the ultimate punishment, as they show its deep historical and contemporary connection to white racial hostility toward blacks.

Criminal history plays a prominent role in violence risk assessments. For people in nondominant groups, disproportionate criminal justice involvement may unfairly and inaccurately elevate violence risk in evaluations. Criminal justice reports continue to document higher rates of arrest and convictions for those in minoritized racial groups. Bureau of Justice surveys have found that ethnic minorities are more likely to serve time when crime is violent than are Whites. Black males ages 18 and 19 were 12.7 times more likely to be imprisoned than White males of the same ages. In fact, across all age bands, from 18 to 65 and older, Blacks had higher rates of imprisonments than Whites. Racial inequities in incarceration rates can translate into a Black offender receiving higher risk scores on actuarial instruments than a White offender and thus a biased misclassification as high risk. Awareness of the impact of structural biases that may be embedded in violence risk assessments is critical to fair assessments. We highlight sources of potential systemic racial bias embedded within existing violence risk assessment methods and conclude with potential methods to enhance structural competency and reduce the risk of biased assessments.

Between Empirical Data and Anti-BlacknessA Critical Perspective on Anti-Asian Hate Crimes and Hate Incidents Janelle Wong (bio) and Rossina Zamora Liu (bio) After a US president (with connections to white nationalists), raised the specter of Yellow Peril and white shooters engaged in mass killings of Asian Americans in Atlanta and Indianapolis in the spring of 2021, white violence toward Asian Americans was difficult to ignore. Yet one leading story of anti-Asian violence in the wake of the pandemic is of an Asian American senior, often termed an "elder" in reporting, or young woman brutally beaten by a person who "appears to be Black." This story and others like it have circulated throughout the Asian American community via viral videos. The story has been the subject of calls for attention to "Black-Asian conflict" in the recent past.1 In March 2021, for example, Vox reporters noted that "Many of the attacks that have gained widespread attention have featured Black assailants, and have threatened to inflame tensions between Asian Americans and Black Americans."2 In April of 2021, a story by an NBC local affiliate in Seattle observed that "There have also been widely circulated videos that show Black men attacking Asian Americans."3 Meanwhile, survey and crime data suggest a different trend. Empirical data, for instance, shows that, compared to their share of the population, Asian American elders (over age 65) are underrepresented among victims of Asian American hate crimes and hate incidents. While women are more likely to report a hate incident to the StopAAPIHate reporting site, multiple sources of data show that men are as likely or more likely to experience a hate incident than women. Further, the vast majority of violence against Asian Americans [End Page 387] consists not of physical assaults but of verbal harassment and "shunning." The data also shows that Black offenders make up a minority of offenders. And, comparatively, Black Americans are up to ten times more likely to report being the victim of a hate crime than Asian Americans, and this pattern persists even in places like California, where Asian Americans comprise nearly double the population of Black Americans.4 This is true despite the fact that people of all racial backgrounds indicate that they are reluctant to report hate crimes. The point here is not to minimize the disturbing incidents, crimes, and even killings that have been widely circulated as part of anti-Asian hate media coverage; rather, placing these incidents in a broader context allows them to be better understood and ultimately addressed by well-informed policy. As two Asian American women and non-Black educators of Color, we seek to better understand the disconnection between the empirical data and the many Black-Asian conflict narratives of anti-Asian violence. We note that, despite a wealth of compelling empirical data, the media arc of anti-Asian violence—historically a symptom of white supremacy—quickly turned from the China-focused rhetoric of a white president and the heinous actions of white mass shooters to a focus on Black individuals physically assaulting Asian American elders. What is disturbing about this second narrative, which we describe as the "Black-on-Asian crime" narrative, is that it eclipses systematic racism captured by data, while gaining a widely accepted place in the discourse of Black aggression as a root cause of anti-Asian violence. The Black-on-Asian crime narrative has not only (re)ignited the Black-Asian conflict trope but seems to have also illuminated an undercurrent of anti-Blackness in narratives of Asian American victimization and perceptions of safety. In this paper, we present data regularly ignored in widely circulating Black-on-Asian crime narratives around anti-Asian violence, followed by a theoretically grounded reflection on the gap between empirical data and viral videos that emphasize Asian American vulnerability against the threat of Black violence. To be clear, we acknowledge that the anti-Asian incidents shown on viral videos are not only real and abhorrent but they have understandably elicited anger and fear in our community. What we hope to illustrate is the way in which these incidents have become prominent in discourses around anti-Asian violence, even though...

When will racial bias in blame and punishment emerge? Here, we focus on White people's willingness to “un-blame” Black and White offenders upon learning of their unfortunate life histories or biological impairments. We predicted that personal mitigating narratives of Black (but not White) offenders would be ignored by Whites who are societal-level anti-historicists. Societal-level anti-historicists deny that a history of oppression by Whites has shaped current societal-level intergroup disparities. Thus, our prediction centers on how societal-level beliefs relate to bias against individuals. Our predictions were confirmed in three studies. In one of those studies, we also showed how racial bias in willingness to un-blame can be removed: Societal-level anti-historicists became open to mitigation for Black offenders if they were reminded that the offender began as an innocent baby. Results are discussed in terms of how the rich literature on blame and moral psychology could enrich the study of racial bias.

One goal of incarceration is offender rehabilitation. We examined whether characteristics of an offender affect beliefs about rehabilitation capacity. In three studies using large samples, we investigated inferences about criminal offenders who were described as juveniles or adults (15 or 30 years old). Participants read about or were shown a picture of a White or Black actor. They judged the offender’s maturation, intentionality, and long-term goals, and indicated their rehabilitation capacity. Black offenders, regardless of age, were seen as more capable of rehabilitation, seen as possessing less intentionality, and having more positive long-term goals than White offenders. Discussion focuses on potential explanations for the data including system justification and attitudinal influence.

A substantial body of research focuses on racial disparity in the criminal justice system, with mixed results due to difficulty in disentangling differential offending from racial bias. Additionally, some research has demonstrated that victim characteristics can exacerbate racial disparity in outcomes for offenders, but little research has focused on the arrest stage. We use a quasi-experimental approach that examines incidents involving co-offending pairs to isolate the influence of offender race on arrest, beyond any characteristics of the incident itself, and we test for moderating effects of victim race and sex on racial disparities in arrest. Our findings reveal that, on average, when two offenders of different races commit the same offense together against the same victim, Black offenders are significantly more likely to be arrested than their White co-offending partners, especially for assault offenses. More importantly, this effect-for both assaults and homicides-is particularly strong when the victim is a White woman. Because these differences are between two offenders who commit the same offense together, we argue that the most plausible explanation for the differences is the presence of racial bias or discrimination.

Color, Loan Approval, and Crimes: The Dark Side of Mortgage Market Deregulation

Research on when alternatives to incarceration are available – and for whom – is underdeveloped. In this study we introduce the concept of “salvageability” as a fourth focal concern guiding the decisions of court actors. In assessing salvageability, actors must consider the casual reasons behind offenders’ criminal involvement and the extent to which those causes can be ameliorated through rehabilitative programming. This process of causal attribution likely exacerbates racial disparity in sentences. We test whether offender race, gender, and prior histories of substance use affect assessments of salvageability as indicated by prosecutor decisions to pre-screen offenders for admittance into an intensive rehabilitation alternative sentence in Pennsylvania using linked data from the Pennsylvania Commission on Sentencing and the Pennsylvania Department of Corrections. Results suggest that black offenders are substantially less likely to be referred for the rehabilitative program even if that referral would be an upward departure from a guidelines-conforming sentence.

Abstract Most convicted offenders serve their sentences under “community supervision” at home instead of in prison. Under supervision, however, a technical rule violation, such as not paying fees, can result in incarceration. Rule violations account for 25% of prison admissions nationally and are significantly more common among black offenders. I test whether technical rules are effective tools for identifying likely reoffenders and deterring crime and examine their disparate racial impacts using administrative data from North Carolina. Analysis of a 2011 reform reducing prison punishments for technical violations on probation reveals that 40% of rule breakers would go on to commit crimes if spared harsh punishment. The same reform also closed a 33% black-white gap in incarceration rates without substantially increasing the black-white reoffending gap. These effects combined imply that technical rules target riskier probationers overall but disproportionately affect low-risk black offenders. To justify black probationers’ higher violation rate on efficiency grounds, their crimes must be roughly twice as socially costly as that of white probationers. Exploiting the repeat spell nature of the North Carolina data, I estimate a semiparametric competing risks model that allows me to distinguish the effects of particular types of technical rules from unobserved probationer heterogeneity. Rules related to the payment of fees and fines, which are common in many states, are ineffective in tagging likely reoffenders and drive differential effects by race. These findings illustrate the potentially large influence of ostensibly race-neutral policies on racial disparities in the justice system.

PULL THE TRIGGER: United States v. Corner and Judicial Discretion in Drug Sentencing Under the First Step Act