Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N. — Bias in the Law: A Definitive Look at Racial Prejudice in the U.S. Criminal Justice System

639 published claims from this book.

• 911 dispatchers mediate all civilian information reaching first responders
• A bias-removing model should predict counterfactual race-neutral outcomes, not likely ones.
• A GAO meta-review confirmed race predicts death penalty judgments
• A Milwaukee DA decriminalized drug and prostitution cases after data showed racial disparity
• A name alone can distort eyewitnesses’ recall of an incident’s violence
• A name alone can trigger false memories of aggression
• A prison record harms Black job seekers more than Whites
• A suggestive model removes race to predict race-neutral case outcomes
• A warmth display by a White prosecutor can read as disrespect
• A white ex-felon outcompetes an equally qualified innocent Black applicant in hiring.
• Academic research on defense attorney bias lags far behind prosecution-side studies
• African-featured defendants get death penalties mainly when victims are White.
• Aggregate racial disparities in prosecution can reverse direction when filtered by offense category
• AI adoption requires bilingual role for engineers and users
• AI fairness requires domain experts who understand the social context of the training data.
• Algorithmic fairness requires stakeholder-defined meaning plus expert contextual vetting
• Algorithmic impact assessments decouple bias monitoring from stated goals
• Algorithmic tools can reduce racial bias by excluding race and context from legal decisions
• Algorithms can be reweighted to equality-neutral treatments that correct documented bias.
• Algorithms can reduce human bias in legal decision-making
• Algorithms can reduce legal-system bias by making decisions free of human judgment foibles.
• Algorithms correcting real sentencing disparities could violate equal protection law
• Algorithms that correct documented bias may themselves violate equal protection law
• All-White juries can convict Black defendants despite convincing exculpatory evidence
• Ambiguity in trials triggers racial bias that clarity suppresses
• American Black prison overrepresentation actually lags the average subordinate-rate across OECD
• An emotional shock can activate racial bias in judges who otherwise show none
• Anonymity removes the social inhibitions that normally suppress overt racial prejudice
• Anti-bias interventions can backfire by threatening the group they regulate.
• Anti-bias training mostly changes knowledge, not behavior, and only when sustained.
• Anti-discrimination jury reforms have failed to reduce race-based strikes.
• Apparent judicial racial bias can stem from prosecutors’ bias, not judges’ own.
• Appointed attorneys face client distrust because they belong to the prosecuting system’s world
• Arrest statistics reflect surveillance bias, but homicide victimization is a cleaner measure.
• Asking callers whether a gun is involved primes weapon perception
• Attorney type outweighs race in predicting detention and sentencing outcomes
• Attorney type predicts outcomes more strongly than race does
• Attorneys function as family, priest, and therapist for their clients
• Autoencoders can reweight training features to correct proportional bias
• Automated record clearance corrected historical bias by removing procedural barriers
• Automated record-clearing algorithms corrected racial disparities from past convictions.
• Banning bias-auditing algorithms while permitting the biased humans who produced the data is incoherent.
• Base-rate training data can mislead models about causality
• Base-rate-trained models yield false causal conclusions without social-context experts
• Batson challenges rarely succeed because courts resist finding strikes discriminatory
• Batson lets attorneys give any race-neutral reason for striking jurors
• Batson’s reform left a loophole allowing race-neutral pretexts for biased strikes.
• Belief in social hierarchy predicts belief that Blacks get fair trials
• Benevolent sexism operates below conscious awareness even in egalitarians
• Bennett’s crime-logic applies to any group, exposing its racial targeting
• Bias affects even well-intentioned egalitarian defense attorneys who represent the accused
• Bias against Black defendants emerges only after judges suffer emotional shocks
• Bias can transmit between agents when threats are ambiguous.
• Bias compounds because real encounters add violent-suspect information absent in laboratory tests
• Bias concentrates where police and prosecutors wield the most discretion.
• Bias makes people interpret identical behaviors differently by the actor’s race
• Bias operates without intention or awareness
• Bias risk rises under emotional, ambiguous, and low-effort decision conditions
• Bias trainings change self-reported knowledge but rarely alter behavior
• Bias-based 911 calls often bypass racial intent and evade legal remedy.
• Bias-override interventions beat bias reduction for urgent decision contexts
• Bias-reducing legal AI requires ongoing human translators to iterate between models and practice
• Bias-reduction training shifts awareness and action more than measured implicit bias itself
• Biased human prosecutors carry an unearned exemption from nondiscrimination scrutiny.
• Big data and AI outperform human judgment in high-stakes decisions.
• Black defendants face death sentences most when victims are White.
• Black drivers are stopped for minor offenses requiring more officer discretion
• Black drivers are stopped less after dark when race is harder to see
• Black drug and murder defendants receive harsher punishment than White ones
• Black juveniles are perceived as older and more often charged as adults
• Black men face a 1 in 1,000 lifetime risk of being killed by police
• Blinding judges to race fails because correlated proxies still leak bias.
• Body camera footage can verify actual compliance with formal policing policies
• Body-camera footage can audit policy compliance like consent-search notification
• Both Black and White jurors show out-group bias in verdicts.
• Caller accents stereotype social class and may bias emergency dispatch priority
• Caller vocal cues like accent could bias emergency dispatchers’ response decisions.
• Calling out racial bias inside a public defender office invites retaliation
• Case outcomes depend on demographics that a blind justice ideal says must not matter
• Certainty of apprehension deters drug crime but severity of punishment does not
• Charge reductions favor White and Asian defendants over Black and Hispanic ones
• Charging data disaggregated by race can flag prosecutorial bias
• Chicago police complaints from White citizens are twice as likely to yield discipline
• Civilians recruit police power when intergroup contact feels threatening
• Client trust in counsel is functionally necessary for effective defense work
• Clients must feel represented or they withhold facts and act destructively
• Closed case files can anchor prosecutors to race-neutral decision baselines
• Closed case files can seed race-neutral predictive models for prosecutorial decisions
• Cognitive-science bias insights demand granular institutional examination to be useful
• Community calls weaponize police as racial-bias enforcers even when officers act fairly.
• COMPAS scored Black defendants as falsely high-risk nearly twice as often
• Competing statistical fairness definitions can conflict irreconcilably within one algorithm
• Complex AI models can be explained by decomposing them into subspaces
• Computer models of officer speech reveal racial disparities in respect
• Concrete reweighting strategies exist for bias-correcting legal algorithms.
• Conflicting fairness definitions make algorithmic fairness a policy decision
• Congress legislated a deliberate 100-to-1 crack-to-powder sentencing disparity in the 1980s
• Consciously egalitarian values fail to prevent racially disparate behavior
• Contemporary American racism no longer operates through explicit or legal means.
• Controlling for crime rates can reverse or erase racial use-of-force disparities
• Correlated defendant attributes, not discrimination alone, can drive apparent bias in verdicts
• Correlated omitted variables like income make racial bias estimates unreliable.
• Correlational agency data but not experimental trials show body cameras reduce force
• Courts demand proof of discriminatory intent, not mere racial imbalance, in jury challenges
• Courts rarely reject race-neutral excuses for juror strikes
• Courts recognized racially discriminatory jury strikes as constitutional violations
• Crime priming makes officers misremember Black faces as more stereotypically Black
• Crime-rate differences cannot fully explain racial disparities in police stops.
• Criminal defense attorneys are overlooked as contributors to racial bias in sentencing
• Criminal defense attorneys are overlooked contributors to racial disparities in the justice system.
• Criminal defense attorneys exhibit implicit racial bias affecting client outcomes
• Criminal justice bias reflects universal dominance hierarchies, not US peculiarity
• Criminal justice discrimination against subordinates recurs across all stratified societies
• Criminal justice maintains order in ways systematically biased toward inequality
• Criminal justice systems maintain social order in a way that systematically favors existing inequalities
• Criminal-justice bias is inherent to all stratified societies.
• Cross-racial interactions fail when groups want respect versus warmth
• Custody settings undermine confidential attorney-client communication
• Darker skin and Afrocentric features predict longer drug sentences and death sentences
• Darker skin tone lengthens prison sentences even after legal factors are controlled.
• Darker skin tone within a racial group yields longer prison sentences
• Darker-skinned Blacks and Latinos face higher risks of police stops and arrests.
• Data accuracy concerns block prosecutors from adopting data-driven decisions
• Data collection intended as feedback requires continuous analysis to avoid backlash.
• Data-driven policy analysis can measurably reduce racial enforcement disparities.
• Debiasing algorithms risk exacerbating bias without a careful implementation framework
• Decisions made by legal algorithms need contestation rights comparable to human appeals
• Defendants sentenced on their birthdays get shorter sentences
• Defendants’ race reliably predicts negative case outcomes
• Defense attorneys avoid discussing racial bias because they fear it as a weakness
• Defense attorneys harbor bias blind spots that worsen racial disparities
• Defense attorneys must confront their own bias before fixing the system
• Defense attorneys show measurable race bias despite their professional role defending clients.
• Defining fairness in AI requires policy decisions only leadership can make
• Different base rates across groups make fairness definitions mathematically incompatible
• Discretion, not policy content, drives persistent racial drug disparities
• Discrimination is greater at earlier criminal justice stages because of official discretion
• Disparity in drug sanctions stems from policing practices, not the drug policies
• Dispatch protocol compels first responders to react to unvalidated racial accusations
• Dispatchers are barred by protocol from vetting callers’ claims of danger
• Dispatchers must send responders even for racially biased false reports
• Displayed emotion and regret actively legitimize rather than undermine discriminatory verdicts
• Distinguishing legitimate from race-proxy criminal predictors is genuinely difficult
• Diverse groups face cohesion costs and anxiety that can impair performance
• Diverse juries eliminate the tendency to evaluate Black defendants’ cases less carefully
• Diverse juries erase the disparity in how carefully evidence is weighed
• Diverse juries produce fewer but more thoughtful juror responses
• Diversifying prosecutors may not reduce bias because in-group bias rivals out-group bias
• Diversifying the judiciary can worsen in-group bias against defendants
• Diversity improves deliberations partly by making White jurors work harder
• Diversity improves group decisions partly by making majority members try harder
• Doctrinal shifts lowering evidence standards enabled aggressive enforcement producing racial disparities
• Drug arrest declines concentrate among African Americans while White rates shift little.
• Drug arrests of Blacks stem from proactive stops while White arrests are reactive
• Drug arrests reduce employment for Black workers specifically
• Drug distribution rates also show the racial crossover effect
• Drug enforcement intensity, not offending, drives racial disparities in sanctions
• Drug prices fell despite intensified drug-war enforcement
• Drug sanction racial disparities have now reverted to mid-1980s levels
• Drug war enforcement has coincided with falling rather than rising drug prices
• Drug-offense racial disparities resist explanation by actual offending differences
• Early racism claims required proving countywide exclusion patterns
• Effective EJ organizing must reject savior-lawyer dynamics while still using legal help
• Effective legal communication requires balancing clarity against condescension
• Egalitarian intentions do not prevent biased behavior without interventions
• Egalitarian self-presentation conceals the true relationship between prejudice and bias
• Eighty percent of Pennsylvania’s juvenile life-without-parole prisoners are subordinates
• Elected judges face re-election pressure that can push decisions away from law
• Elected judges may bend rulings toward voter preferences to secure re-election.
• Elected judges may rationally impose biased sentences to match voters’ crime concerns
• Elected prosecutors increasingly campaign on racial justice even while most districts remain uncontested
• Electoral incentives can push judges toward racially disparate sentencing
• Emergency call bias can transfer across agents in a multi-stage chain
• Emergency call transcripts could reveal where racial bias enters community policing
• Emotional shocks can activate latent racial bias in sentencing
• Even African Americans hold negative bias toward other African Americans
• Even after declines, drug arrest rates remain nearly double their 1980 baseline.
• Even outcome-free routine stops erode citizens’ trust in police legitimacy
• Even people with consciously egalitarian values act in ways that worsen racial disparities.
• Even trivial group distinctions produce in-group favoritism in juror verdicts
• Even unbiased prosecutors can perpetuate racial disparity through inherited policies
• Even unbiased prosecutors perpetuate racial disparity through longstanding policies and practices
• Even well-intentioned defense attorneys harbor an unacknowledged bias blind spot
• Even well-intentioned technology can embed hidden functional biases
• Everyday racist calls may stem from ambiguity rather than explicit bias
• Evidence is mixed on whether anti-bias trainings change behavior at all
• Evidence of existing disparities can reinforce rather than challenge the stereotypes behind them.
• Excluding a community segment from juries removes unknowable ranges of human experience
• Facial affect perception skews so Black faces read as angrier, skewing credibility judgments
• Facial recognition fails because training data overrepresents White faces.
• Facial recognition fails on darker faces because of biased training data
• Facial stereotype association, not race alone, drives harsher sentencing and death penalties
• Failing to code mock-juror race can mask racial bias findings
• Failing to control for population density produces spurious no-disparity findings in waste-siting studies.
• Fairness in legal AI is defined by stakeholders not technicians
• Fatal police shootings of unarmed Blacks measurably harm community mental health
• Fear of being labeled racist deters prosecutors from adopting data-driven accountability reforms.
• Fear of blame makes police agencies avoid discussing racial disparities
• Fear, stress, fatigue, and ambiguity erode officers’ bias-override capacity
• Federal drug-war funding worked by incentivizing state and local adoption of harsh policies
• Federal money leveraged states into adopting punitive drug enforcement
• Fiscal strain plus conservative coalitions drove decarceration of drug offenders
• Fiscal strain, not ideology alone, drove states to decarcerate drug offenders.
• Formal defense investigator training programs only emerged in the late 1990s
• Former police officers as defense investigators retain cop language ill-suited to defendants
• Framing disparities as institutional products can short-circuit stereotype-based rationalization.
• Goal-guided targeted use of risk assessments reduces both detention and bias
• Governmental AI adoption involves three distinct stakeholder groups that must cooperate
• Group fairness requires equal predicted outcomes across protected groups
• Groups punish their own deviant members more harshly to protect group reputation.
• Guilty pleas dominate criminal case disposition overwhelmingly
• Guilty pleas, not trials, now dominate criminal case disposition
• Habit-breaking bias interventions change awareness more than they reduce bias
• High filing rates effectively rubber-stamp biased police arrest practices
• High intraracial homicide among Black males constitutes slow-moving collective self-debilitation.
• Higher dismissal rates for minority defendants may mask prosecutorial screening rather than signal leniency
• Highlighting group disparities can reinforce the stereotypes they stem from
• Hispanic people are vastly underrepresented among California prosecutors.
• Homicide victimization ratio is a policing-bias-free criminality measure
• Homogeneous jury composition was blamed for the Zimmerman acquittal
• Human judgment combined with algorithms outperforms either alone
• Human plus algorithm outperforms either alone on judgment tasks
• Identically proportioned Black men are perceived as larger and more threatening than White men
• Ignoring racial bias in defense work is a self-serving strategy
• Implicit bias alone insufficiently explains racial inequity in criminal justice outcomes
• Implicit bias operates below awareness and resists admission even when perceived
• Implicit bias operates without intention and can contradict conscious egalitarian values
• Implicit bias passes from parents to preschoolers and via TV viewing
• Implicit bias reconciles lived experience with officers’ denial of intent
• Implicit bias skews prosecutors’ pretrial decisions against defendants of color
• Implicit bias surfaces strongest under time pressure but is consciously overridden with adequate time
• Implicit bias training in criminal justice still lacks rigorous outcome evaluation
• Implicit bias trainings rarely produce meaningful behavioral change
• Implicit biases can be consciously overridden through deliberate attention
• Implicit out-group bias triggers measurable physiological threat responses
• Implicit racial bias transmits through media and parenting without conscious awareness
• Imprisonment fails to deter future drug offending across all racial groups
• In-group bias of similar magnitude across groups renders bias undetectable
• In-group bias research is stalled by scarce balanced race-matched data
• In-group judicial favoritism can cancel out in aggregate and hide overall racial bias.
• Incarceration records reduce employment and wages, worst for Black men
• Indigent clients distrust appointed defense counsel for being part of the prosecutorial system
• Individual judges vary widely in how they sentence minority versus white defendants
• Individual juror bias studies may not predict deliberating jury verdicts
• Individual juror decisions do not simply aggregate into deliberating jury outcomes
• Individual juror judgments cannot substitute for deliberating jury decisions
• Individual racial bias scales into systemic bias through high-stakes interactions
• Individual willpower cannot override implicit bias; institutional design must
• Individual-level analysis of police encounters obscures institutional causes of disparity
• Individuals cannot override unconscious bias by willpower; institutions must intervene
• Information exchange models predict heterogeneous groups deliberate better.
• Institutional policies can make even fair-minded individual officers reproduce racial disparities
• Institutional practices can produce racial disparities without any racist intent
• Institutional responses to fatal bias mistakes remain symbolic and disproportionately mild
• Institutions can remedy bias through guard-rail policies without targeting individuals
• Institutions coordinate agent behavior and persist beyond individuals
• Institutions create inequality, so highlighting them counters bias rationalization
• Institutions penalize those who raise race bias rather than the biased conduct
• Institutions standardize agent behavior and persist beyond any individual
• Integrated long-term bias trainings outperform one-time isolated ones
• Intraracial homicide among subordinates functions as slow-motion collective self-debilitation
• Isolating racial bias requires cases identical except for race
• Jail incarceration rates fell for Black and Hispanic people but rose for Whites, 2005-2017
• Judge attributes beyond the law shape judicial outcomes
• Judge behavior can falsely appear biased when prosecutors treat defendants unequally.
• Judge disparities may really reflect prosecutor behavior rather than judicial bias
• Judge fatigue and mood states measurably alter sentencing outcomes
• Judge mood and hunger can cause arbitrary sentencing that awareness might correct
• Judge sentencing shifts with mood, fatigue, hunger, and emotional trauma.
• Judge-defendant race matching produces measurable in-group sentencing bias.
• Judges can project racial bias from earlier actors in the case pipeline
• Judges favor plaintiffs of their own ethnicity, amplified by local terrorism
• Judges show negative ingroup bias toward defendants of their own race.
• Judges show positive in-group bias favoring litigants of their own race
• Judges vary widely in racial bias, not conviction severity itself
• Judges’ case outcomes can show racial bias even when judges themselves are unbiased
• Judges’ personal attributes and even trivial bodily states can shape decisions
• Judges’ politics, gender, and family shape their rulings empirically
• Judges’ race, gender, and politics measurably shift their case outcomes
• Judicial bench cards reduce implicit bias through script-flipping self-checks
• Judicial race and gender affect case outcomes, making bench diversity a remedy for bias
• Jurors of all races show symmetric out-group bias in punishment
• Jurors punish defendants more when crimes fit racial stereotypes
• Jurors punish in-group members harsher than outsiders when evidence is strong
• Jury composition measurably changes racial bias in death sentencing.
• Jury composition of White men specifically drives racially biased death sentencing
• Jury diversity research has barely examined non-Black minority groups or inter-minority deliberation
• Jury verdicts can convict the innocent when racial bias dominates
• Labeling race-neutral disparities racist is itself harmful and counterproductive
• Laboratory mock-jury out-group bias is inconsistent while archival data show strong juror bias
• Law enforcement fears data transparency because bias claims weaponize statistics against them
• Laws penalizing biased 911 calls may backfire via threat reactions.
• Legal AI can be deployed to identify and correct existing bias in law
• Legal realists hold judges’ attributes shape judicial outcomes
• Legal tech can be designed to suggest race-neutral outcomes rather than predict them
• Legal technology can identify and correct bias rather than only amplify it
• Limited attorney quality inflates conviction rates and sentences for minorities
• Line prosecutors attribute racial disparities to poverty and minority culture, not bias
• Long-term integrated bias training changes behavior; short standalone workshops only shift knowledge.
• Low perceived-threat thresholds combine with anonymity to trigger bias accusations
• Low threat thresholds plus anonymity drive racially discriminatory emergency calls.
• Low-prejudice people experience compunction when their behavior contradicts their egalitarian standards
• Machine decisions in disability cases can leave affected groups without appeal
• Machine decisions in disability determinations lack transparency and appeal mechanisms.
• Machine learning has aided legal prediction for over fifty years
• Machine learning models encode the historical base rates present in their training data
• Machine predictions can cut crime by a quarter without growing jail populations
• Machine-based risk predictions can cut crime or jail populations dramatically
• Making race salient in a trial reduces jurors’ racial bias
• Mandatory gun-questions prime callers to report weapons against Black subjects.
• Marketing an AI anti-bias vision requires framing bias as existential threat
• Mathematically distinct fairness definitions can directly conflict with one another
• Measured bias appears at prosecutors’ charging stage, not just sentencing.
• Measured case outcomes show race and gender bias across multiple court settings.
• Measured drug-use data contradict both the offending-explanation and the equal-offending counterclaim
• Measured sentencing gaps do not reveal which group gets the unfair treatment
• Measuring judicial race bias requires cases identical except for defendant race.
• Measuring racial bias in stops requires choosing an appropriate benchmark
• Memory of a crime’s details varies based on the suspect’s name
• Mere awareness of jury diversity reduces White jurors’ bias before deliberation
• Merely believing a jury is diverse makes White jurors self-regulate more and speak less
• Merely believing juries are racially diverse changes how jurors deliberate
• Meta-analytic findings on juror racial bias are mixed and depend on participant coding
• Meta-analytic results hold across study, sample, and test levels
• Mind science bias interventions only work where actors already share egalitarian values
• Minority defendants are convicted and sentenced more harshly than Whites
• Minority distrust of the legal system may shape their verdicts
• Minority juror underrepresentation shapes jury decisions beyond mere biased outcomes
• Minority jurors may use verdicts to correct institutional inequalities
• Minority jurors show in-group favoritism in mock juror decisions
• Minority jurors’ low system trust may shape verdict decisions
• Minority-group presumptions demand explicit respect from authority figures for institutional trust to form
• Mislabelling race-driven disparity as racism could itself be harmful
• Missing minority jurors changes jury deliberation beyond just verdict discrimination
• Mock jurors sentence in-group members harsher when guilt evidence is strong.
• Mock jury experiments use written or videotaped trials to test racial bias
• Mock-juror racial bias findings diverge partly because participant race was unmeasured
• Modern prejudice surfaces only when people can act without appearing discriminatory
• More punitive drug sanctions fail to deter subsequent drug offending
• More racially prototypical faces are more likely to be shot
• Most racial disparity stems from well-meaning people’s unconscious bias, not explicit hatred
• Multivariate statistical controls define whether remaining disparities count as racial discrimination
• Mutually canceling in-group biases can erase all detectable signals of racial discrimination
• Nearly all criminal convictions arise from plea bargains, giving prosecutors disproportionate power.
• Neutral policies can produce racial disparities through disproportionate group exposure
• New lawyers can pass the bar without ever practicing client interviewing skills
• New lawyers may try serious cases untrained in client communication
• No judicial-bias study achieves perfect counterfactual identification of race effects.
• No rigorous outcome evaluation of criminal justice bias training exists yet.
• No single statistical definition of fairness can satisfy all reasonable criteria simultaneously
• Nonlethal police force is used on Blacks far out of proportion to population
• Normalizing implicit bias as universal avoids defensive stalemates in training
• Oakland officers who never restrained white people still searched Black people
• Oakland police show racial disparities across every interaction stage
• Oakland stop data shows racial disparities across most officers, not a few
• Observational judicial bias findings can be confounded by case sorting upstream of the judge
• Observed race disparities in sentencing do not fully capture judicial bias.
• Observed sentencing disparities do not necessarily prove judge discrimination
• Off-duty Black officers are disproportionately shot by fellow police officers mistaken about identity
• Offense-specific disparity data enables targeted policy intervention
• Offices deploying legal AI should publish quarterly bias metrics like financial reports
• Official 911 guidance legitimizes nearly any call as an emergency
• Omitted variables correlated with race bias the discrimination estimate
• Omitted variables correlated with race make regression estimates of judicial bias unreliable
• Omitted-variable critiques of bias studies have weakened with richer data
• Open hatred explains only a small share of racial disparity; unconscious bias drives most
• Optimal predictive legal models deliberately include race as a feature
• Oregon enacted the first law allowing lawsuits against racially biased 911 callers
• Oregon passed the first anti-racial-profiling-call statute in 2019
• Out-group bias is strongest for minority jurors judging White defendants
• Overall racial bias estimates can be decomposed from in-group bias regression coefficients
• Overt racism has been replaced by subtle implicit bias
• Parole decisions worsen as judges tire and recover after food breaks
• People easily generate race-neutral justifications for decisions actually driven by race
• People high in social dominance orientation show more racial bias in jury decisions
• People interpret ambiguous acts as more aggressive when the actor is Black
• People judge Black targets as bigger, stronger, and more threatening even from faces alone
• People punish innocent group members for one member’s offense
• People suppress prejudice only when aware their actions may look discriminatory
• People who endorse inequality also endorse the death penalty and torture
• Perceived adultness of Black children tracks their disproportionate adult charging
• Perceived racial distance from counsel suppresses client disclosure of case facts
• Peremptory challenges remain the main mechanism excluding minority jurors
• Peremptory strikes reveal bias only through explicit material like prosecutors’ notes
• Physical fatigue shifts officers’ shooting patterns toward distant targets
• Police decisions as gatekeepers shape every downstream phase of criminal justice
• Police discretion determines whether anyone enters the criminal justice system
• Police officers exceed and public defenders fall below average social-dominance orientation
• Police officers suppress race-linked shooting errors but still react faster against Black suspects
• Police resist bias data collection out of fear of blame and accusation.
• Police score higher and defenders lower on social dominance orientation.
• Police shootings kill roughly 1 in 1,000 Black men and boys over a lifetime
• Police show reduced shooter bias errors but faster shooting for Black armed suspects
• Police stops increase later delinquency through psychological distress among minority youth
• Police stops themselves increase later delinquency, counterproductively raising crime
• Police training itself instills a ‘prejudice norm’ in recruits
• Police training reduces shooter misidentification but not racial speed bias
• Police treat field data as evidence, which blocks its use for systemic self-audit
• Police treatment disparities span stops, abuse, force, and response speed
• Policing practices, not drug laws, drive most racial disparities in sanctions
• Policymakers, courts, and sentencing reforms together produced racialized drug enforcement.
• Post-2000s drug incarceration racial disparities contracted as Black arrests fell
• Post-Reconstruction jury integration was reversed within a decade by unofficial exclusion practices.
• Power-holders’ anxiety displays read as bias and reproduce the very inequality they fear
• Pre-Batson courts required countywide exclusion records to prove peremptory-strike bias
• Predictive algorithms cannot cleanly label predictors legitimate or biased proxies
• Predictive models can suggest case outcomes under counterfactual race-neutral processing.
• Predictive policing software can lend false objectivity to existing police bias
• Prejudice awareness makes jurors evaluate evidence more thoroughly
• Pretrial detention drives a large share of Black-White sentence disparity
• Pretrial detention drives up to 47 percent of Black-White conviction and sentence disparities
• Prior violent-intelligence information compounds appearance-based bias for officers
• Prison sentences for drug offenses do not deter future drug offending
• Proponents argue careful algorithm design can curtail implicit racial bias in court
• Proponents argued punitive drug sanctions deter offending and raise drug prices.
• Prosecution offices with egalitarian values can still produce racially disparate outcomes
• Prosecution-stage disparity research remains far less developed than policing and sentencing research
• Prosecutor charging decisions retroactively endorse biased police conduct.
• Prosecutor charging decisions, not judges, drive a 10 percent racial sentence gap.
• Prosecutor race predicts more punitive marijuana sentence offers
• Prosecutor race shapes plea offers only in some narrow case types
• Prosecutorial charging practices implicitly endorse biased police conduct
• Prosecutorial data is systematically unavailable for assessing racial disparities
• Prosecutorial data systems cannot measure defendants’ race or ethnicity reliably
• Prosecutorial decisions can be anchored to race-neutral predicted outcomes
• Prosecutorial discretion exercised under uncertainty and time pressure falls back on racial stereotypes.
• Prosecutorial discretion radiates through the whole justice system but stays undocumented
• Prosecutorial discretion, not just police arrests, drove the prison boom and its racial disparity.
• Prosecutorial dismissals show the widest racial-ethnic divergence across groups
• Prosecutorial diversity is advocated as a route to equitable case outcomes.
• Prosecutorial secrecy blocks both research and accountability for discretion
• Prosecutors are least welcomed in communities with racial compositions unlike their own
• Prosecutors attribute racial disparities to poverty and crime patterns rather than bias
• Prosecutors can audit their own charging decisions using race-disaggregated data
• Prosecutors can benchmark race-neutral decisions using models built from closed case files.
• Prosecutors can reduce disparities by decriminalizing offenses with biased enforcement records
• Prosecutors have only recently become willing to discuss racial disparities
• Prosecutors questioned jurors differently by race to justify strikes
• Prosecutors rarely collect the data needed to detect racial bias
• Prosecutors recently treated racial disparity research as institutionally unacceptable
• Prosecutors systematically strike Black jurors whom they perceive as distrusting police.
• Prosecutors who file most police cases implicitly endorse biased arrest practices
• Prosecutors wrote circled ‘B’ notes next to all Black potential jurors
• Prosecutors’ charging discretion functioned as hidden sentencing power in the drug war.
• Prosecutors’ own handwritten notes were the only way to expose racial jury strikes
• Public defenders have verbally abused clients of color during first meetings
• Public defenders, presumed liberal, display stereotyping and prejudice themselves
• Publication bias distorts what we know about race and prosecutorial decisions
• Publication bias skews the research record on prosecution disparities
• Publicized disparities are seized as evidence that confirms the stereotypes that produced them
• Punishing minority defendants can function to preserve existing social hierarchies
• Punitive attitudes express a desire for social hierarchy rather than crime control
• Punitive drug policies failed in purpose yet harmed Black life chances
• Punitive drug policy is attuned to the race of the user population
• Punitive drug sanctions may backfire by destroying desistance-promoting social opportunities
• Punitive sanctions backfire by destroying the social bonds that enable desistance
• Punitive sanctions can backfire by destroying opportunities that enable desistance.
• Quick shoot decisions bias toward shooting unarmed Black individuals
• Race and class background affect every stage of criminal justice processing
• Race and gender of defendants causally affect judicial outcomes
• Race gap in imprisonment is four times larger for men than women
• Race matching alone does not grant defense attorneys license to raise bias
• Race pairing predicts death sentences even after controlling for legal factors
• Race shapes imprisonment far more for men than for women.
• Race still sways sentencing even after legal factors dominate outcomes
• Race-aggresion associations speed up 911 calls against Black suspects.
• Race-based misperception of Black men as threatening reaches into charging and sentencing
• Race-blind prediction models can encode bias; active debiasing improves fairness.
• Race-blind procedures fail because correlated proxy variables still leak bias
• Race-neutral justifications easily mask biased peremptory jury strikes
• Race-neutral justifications for peremptory strikes shield implicit bias from Batson review
• Race-neutral rhetoric coexists with licensed racial profiling for security and policing
• Racial anxiety is a measurable stress response to intergroup contact
• Racial anxiety produces measurable physiological, behavioral, and cognitive effects in interracial encounters.
• Racial bias and out-group bias are distinct juror phenomena
• Racial bias debates in policing are starved by uncollected, unstandardized data
• Racial bias enters criminal justice at every decision stage from legislation to verdict.
• Racial bias in 911 calls spreads through multi-agent information channels into the justice system
• Racial bias in criminal case outcomes varies by decision point and offense type
• Racial bias infects prosecutors’ pretrial bail and plea decisions
• Racial bias operates at every stage from lawmaking through conviction
• Racial bias operates pervasively across all life domains
• Racial bias permeates every stage of the criminal justice system
• Racial bias pervades every domain rather than rare individuals
• Racial crime-rate differences cannot fully explain police contact disparities
• Racial discrimination concentrates at early criminal justice stages where discretion is widest
• Racial discrimination in criminal justice declines as discretion narrows at later stages
• Racial discrimination in criminal justice hits subordinate men hardest
• Racial discrimination in criminal justice is greatest where officials hold discretion.
• Racial discrimination in imprisonment is far stronger against men than women
• Racial disparities compound at every stage of the criminal justice process
• Racial disparities in consent searches can stem from citizens’ fear rather than officer targeting
• Racial disparities in consent searches may stem from community responses, not officer targeting.
• Racial disparities in drug enforcement do not track actual racial differences in drug behavior
• Racial disparities in imprisonment double for men versus women
• Racial disparities in incarceration can stem from arrest-rate differences alone, without any judicial bias.
• Racial disparities in judicial outcomes do not necessarily prove unequal treatment
• Racial disparities in policing arise from systemic forces, not individual racism
• Racial disparities in prosecution do not uniformly favor White defendants
• Racial disparities in prosecution vary by decision point and group
• Racial disparities in sentencing can arise without any judicial bias or discrimination
• Racial disparities in traffic stops increase when officers can see a driver’s race.
• Racial disparities persist even in precincts where people of color are a small minority.
• Racial disparities research on prosecution remains sparse, unlike policing and sentencing
• Racial disparities vary across decision points, offenses, and jurisdictions
• Racial disparity counts as warranted only when legally relevant factors explain it
• Racial disparity in plea bargaining is nearly unstudied despite its centrality
• Racial disparity in police stops widens when officers can visually ascertain the driver’s race
• Racial disparity in policing is distinct from proof of racism
• Racial drug-enforcement disparity traces directly to a deliberate mid-1980s policy pivot
• Racial equity in criminal justice requires institutional redesign, not just attitude change
• Racial framing decides which killers get mythologized versus executed by police
• Racial gaps in drug prison admissions predate the War on Drugs.
• Racial gaps in jail incarceration narrowed between 2005 and 2017
• Racial groups hold genuine divergent goals of liking versus respect in cross-racial contact
• Racial incarceration disparities persist even after controlling for arrest rates
• Racial inequality persists chiefly through unconscious bias in people who believe themselves fair
• Racial inequity in criminal justice involves three distinct covert mechanisms, not bias alone
• Racial name cues cause witnesses to recall fewer aggressive details from memory
• Racial sentencing disparities can arise from case sorting even with unbiased judges
• Racial sentencing disparities do not reveal which group is treated wrongly
• Racial stereotypes distort split-second shoot or don’t shoot decisions
• Racial stereotypes serve the dominant group’s identity needs
• Racial stop disparities hold even in precincts with few minority residents
• Racialized leniency toward the opioid crisis nonetheless reduced Black drug arrests
• Racially diverse juries both reduce bias and improve decision quality
• Racially motivated emergency calls constitute a recognized legal harm requiring remedy
• Raising awareness of bias risk can either reduce discrimination or sharpen scrutiny of evidence
• Random emotional shocks to judges produce arbitrary punishment decisions.
• Real-world data show juror bias but lab studies often fail to replicate it
• Recent drug sanction declines come almost entirely from reduced Black arrest rates
• Reducing racial disparities in prosecution requires offense-specific, decision-point-level data
• Reducing racial inequity in courts requires institutional systems, not just attitude change.
• Reform-minded elected prosecutors cannot guarantee their directives reach actual charging decisions
• Reframing bias as universal brain function defuses defensiveness in anti-racism training.
• Refusing to acknowledge racial bias does not prevent it from shaping behavior
• Regional organizations succeed in inverse proportion to civilizational diversity
• Repeated police stops erode citizens’ sense of equal standing and legitimacy of law
• Requiring explicit consent mandates reduced consent searches without raising crime in Oakland
• Requiring narrative justifications for stops reduces ambiguity in police encounters.
• Research on jury diversity has focused almost exclusively on Black diversity
• Risk-scoring error types depend on which label counts as positive
• Routine police dispatches to prejudice-driven civilian calls harm intergroup relations even without violence
• Routine police stops erode community trust more than fatal incidents do.
• Routine traffic stops erode police-community trust more than famous shootings
• Same-race judges punish in-group juveniles more harshly out of tough love
• Same-race judicial assignment can worsen outcomes for juvenile defendants
• Scripted emergency calls transmit racial bias into police response despite formal neutrality
• Searched minority drivers yield contraband less often than searched Whites.
• Self-presentation concerns mask the link between prejudice and juror bias
• Self-reported drug distribution differences between races are too small to explain arrest disparities
• Self-reported receptivity to anti-discrimination training fails to measure real behavioral change.
• Self-reported training satisfaction cannot measure behavioral change.
• Sentencing varies with stereotypical facial features even absent overall racial disparity.
• Shoot bias targets Black men but not Black women
• Shoot decisions vary with officer fatigue, distance, and target race prototypicality
• Sincere egalitarian conviction coexists with unconscious bias that surfaces only in ambiguous decisions
• Sincere fairness goals do not prevent race-based misjudgment in prosecutors
• Skin tone predicts police stops for Latinos as well as Blacks
• Social dominance orientation predicts belief in justice-system fairness independent of racial affect
• Social media scrutiny has driven recent prosecutorial race-reform campaigns
• Social media scrutiny pushed prosecutors toward racial justice reforms
• Speaking up about race in an all-male white-dominated room triggers retaliation
• Stage-by-stage siting analysis can reveal racism when discrete decisions look neutral.
• Standard legal predictive models deliberately include race to maximize accuracy.
• Standardized dispatch protocols traded discretionary filtering for reduced individual bias
• Standardized emergency dispatch protocols remove one bias while blocking filter checks on caller validity
• Standardized emergency dispatch protocols trade individual discretion for difficulty screening out non-emergencies.
• State criminal-justice violence functions to maintain group-based hierarchy.
• Stereotype threat can corrupt majority actors’ judgment, producing inflated feedback and unearned promotions.
• Stereotype threat causes minority patients to dis-identify from health-promoting behaviors
• Stereotype threat distorts feedback and promotion judgments across races
• Stereotype threat distorts workplace feedback and promotion decisions, harming both targets and providers
• Stereotype threat in court mirrors healthcare nonadherence effects across all legal roles
• Stereotype threat physically manifests as elevated heart rate and blood pressure during challenging tasks.
• Stereotype threat produces measurable physiological stress responses in body and brain
• Stereotype threat suppresses crime reporting and witness cooperation
• Strict dispatch scripts prevent dispatchers from screening out racially motivated false reports.
• Strong race-aggression associations and rapid racial categorization are well documented.
• Structured checklists reduce judicial bias in foster-care decisions for minority youth.
• Studies find racial bias throughout justice-system actors including officers
• Subliminal Black faces speed recognition of crime-related objects
• Subliminal racial primes alter speed of crime-object detection
• Subordinate-on-dominant lethal violence is punished more harshly than other race pairings
• Supreme Court rulings lowering the stop standard enabled aggressive drug enforcement
• Sustained positive peer interactions with out-group members reduce implicit bias over a lifetime.
• Systematic assignment to judges alone can produce racial disparities without any bias.
• Targeted apprehension-boosting intervention cut drug possession offenses by 55 percent.
• Targeted use of pretrial risk assessments reduces detention and racial disparities
• Technical adoption depends on translators bridging engineers and professional users
• Technology produces racial bias without any intending racist
• The 1980s drug war shifted enforcement from borders to domestic streets
• The 1980s War on Drugs created street-level policing as a new enforcement front
• The 1986 Anti-Drug Abuse Act codified punitive drug-war philosophy into law
• The 1986 Anti-Drug Abuse Act created mandatory minimum sentences for drug crimes
• The attorney education and jargon gap intimidates clients into silence
• The bad-apples framing blocks institutional responses to racial bias
• The Black-crime stereotype physically alters visual perception, attention, and memory
• The few bad apples narrative blocks institutional reform against racial bias
• The IAT measures group patterns, not individual bias levels
• The neutral liberal subject is defined against the racialized communitarian other
• The opioid crisis prompted treatment-focused policy, unlike prior drug epidemics
• The punitive drug-control turn after 1980 failed its goals while worsening Black life chances
• The research evidence on anti-bias training effectiveness is genuinely conflicting.
• The social norms defining a legitimate emergency call remain undefined
• The Stroop effect demonstrates how ingrained habits override conscious instructions.
• The Stroop test convinces training audiences that unconscious processes override conscious intent
• The War on Drugs amplified but did not create drug imprisonment disparities.
• The War on Drugs did not create but vastly amplified existing racial disparities
• Time pressure activates implicit racial bias while conscious awareness enables cognitive override
• Total drug arrest trends mirror possession arrests, not sales arrests
• Translators are needed to shepherd legal AI adoption through repeated iterations
• Treatment equality across racial groups requires differential model predictions per group
• Twenty-eight studies confirm the race-death penalty connection.
• Unacknowledged unconscious bias still influences attorney behavior
• Unarmed Black and Latino victims are vastly overrepresented among police killings.
• Unarmed racial subordinates are killed by police at rates far exceeding their population share
• Unbiased judges can appear biased because other actors sort cases beforehand
• Unconscious bias affects even well-intentioned criminal defense attorneys.
• Unconscious bias affects even well-intentioned opposing-counsel defenders
• Underrepresentation of subgroups creates proportional bias beyond direct discrimination
• US Black imprisonment overrepresentation is lower than OECD migrant overrepresentation averages.
• US Black over-incarceration is unremarkable by global subordinate standards
• US drug arrest rates peaked in 2006 and remain nearly double 1980 levels
• US law requires governmental decisions be explainable, constraining opaque AI use
• Using prior arrests instead of sentences drives racial plea disparities
• Voice-based cues can bias emergency dispatchers before first responders arrive
• Voter-roll sourcing skews jury pools toward higher socioeconomic status
• War-on-Drugs doctrine preferred punishment over treatment by design
• What counts as a legitimate 911 emergency is undefined
• Whether bias stems from animus or statistics matters little to defendants
• Whether race-based disparities stem from animus or statistics may not matter to defendants
• White male jurors increase death sentences only for Black defendants
• White male jury composition disproportionately harms Black defendants while leaving White defendants unaffected
• White observers read Black male faces as angrier and more hostile
• Whites’ biased emergency calls underwrite a largely unstudied racial-bias channel.
• Wise feedback restores trust through asset-framed, candid, aspiration-aligned communication.
• Within-race skin tone predicts sentence length independent of legal factors
• Young Black and Latino males are stopped at over eight times their population share