Racial and Ethnic Disparities in the Criminal Justice System

Throughout the nation, people of color are far more likely to enter the nation’s justice system than the general population. State and federal governments are aware of this disparity, and researchers and policymakers are studying the drivers behind the statistics and what strategies might be employed to address the disparities, ensuring evenhanded processes at all points in the criminal justice system. This primer highlights data, reports, state laws, innovations, commissions, approaches and other resources addressing racial and ethnic disparities within our country’s justice systems, to provide information for the nation’s decision-makers, state legislators.

For states to have a clear understanding of the extent of racial and ethnic disparities in the states, they need to have data from all stages of the criminal justice system.

I. Law Enforcement

Disparities Within Traffic Stops

Contact with law enforcement, particularly at traffic stops, is often the most common interaction people have with the criminal legal system.

According to a large-scale analysis of racial disparities in police stops across the United States, “police stop and search decisions suffer from persistent racial bias.” The study, the largest to date, analyzed data on approximately 95 million stops from 21 state patrol agencies and 35 municipal police departments across the country. The authors found Black drivers were less likely to be stopped after sunset, when it is more difficult to determine a driver’s race, suggesting bias in stop decisions. Furthermore, by examining the rate at which stopped drivers were searched and turned up contraband, the study found that the bar for searching Black and Hispanic drivers was lower than that for searching white drivers.

The study also investigated the effects of legalization of recreational cannabis on racial disparities in stop outcomes—specifically examining Colorado and Washington, two of the first states to legalize the substance. It found that following the legalization of cannabis, the number of total searches fell substantially. The authors theorized this may have been due to legalization removing a common reason officers cite for conducting searches. Nevertheless, Black and Hispanic drivers were still more likely to be searched than white drivers were post-legalization.

Data Collection Requirements in Statute

At least 23 states and the District of Columbia have laws related to or requiring collection of data when an individual is stopped by law enforcement. Some of these laws specifically prohibit racial profiling or require departments to adopt a policy to the same effect. Collection of demographic data can serve as a means of ensuring compliance with those provisions or informing officials on current practices so they can respond accordingly.

States have employed many reporting or other requirements for evaluation of the data collected under these laws. For example, Montana requires agencies to adopt a policy that provides for periodic reviews to “determine whether any peace officers of the law enforcement agency have a pattern of stopping members of minority groups for violations of vehicle laws in a number disproportionate to the population of minority groups residing or traveling within the jurisdiction…”

Maryland’s law requires local agencies to report their data to the Maryland Statistical Analysis Center. The center is then tasked with analyzing the annual reports from local agencies and posting the data in an online display that is filtered by jurisdiction and by each data point collected by officers.

The amount and kind of data collected also varies state by state. Some states leave the specifics to local jurisdictions or require the creation of a form based on statutory guidance, but most require the collection of demographic data including race, ethnicity, color, age, gender, minority group or state of residence. Notably, Missouri’s law requires collection of the following 10 data points:

  1. The age, gender and race or minority group of the individual stopped.
  2. The reasons for the stop.
  3. Whether a search was conducted because of the stop.
  4. If a search was conducted, whether the individual consented to the search, the probable cause for the search, whether the person was searched, whether the person’s property was searched, and the duration of the search.
  5. Whether any contraband was discovered during the search and the type of any contraband discovered.
  6. Whether any warning or citation was issued because of the stop.
  7. If a warning or citation was issued, the violation charged or warning provided.
  8. Whether an arrest was made because of either the stop or the search.
  9. If an arrest was made, the crime charged.
  10. The location of the stop.

State laws differ as to what kind of stop triggers a data reporting requirement. For example, Florida’s law applies to stops where citations are issued for violations of the state’s safety belt law. While Virginia’s law is broader, requiring all law enforcement to collect data pertaining to all investigatory motor vehicle stops, all stop-and-frisks of a person and all other investigatory detentions that do not result in arrest or the issuance of a summons.

Cultural Competency and Bias Reduction Training for Law Enforcement

At least 48 states and the District of Columbia have statutory training requirements for law enforcement. These laws require law enforcement personnel statewide to be trained on specific topics during their initial training and/or at recurring intervals, such as in-service training or continuing education.

States That Statutorily Require Law Enforcement Bias Training

In most states, the law simply requires training on a subject, leaving the specifics to be determined by state training boards or other local authorities designated by law. However, some states, such as Iowa and West Virginia, have very detailed requirements and even specify how many hours are required, the subject of the training, required content, whether the training must be received in person and who is approved to provide the training.

Overall, at least 26 states mandate some form of bias reduction training. Find out more about these laws on NCSL’s Law Enforcement Training webpage.

Law Enforcement Employment and Labor Policies

States have also addressed equity and accountability in policing through certification and accountability measures and hiring practices.

For example, a 2020 California law (AB 846) changed state certification requirements by expanding current officer evaluations to screen for various kinds of bias in addition to physical, emotional or mental conditions that might adversely affect an officer’s exercise of peace officer powers. The law also requires the Commission on Peace Officer Standards and Training to study, review and update regulations and screening materials to identify explicit and implicit bias against race or ethnicity, gender, nationality, religion, disability or sexual orientation related to emotional and mental condition evaluations.

In addition to screening, the California law requires every department or agency that employs peace officers to review the job descriptions used in recruitment and hiring and to make changes that deemphasize the paramilitary aspects of the job. The intent is to place more emphasis on community interaction and collaborative problem-solving.

Nevada (AB 409), in 2021, added to statutory certification requirements mandating evaluation of officer recruits to identify implicit bias on the basis of race, color, religion, national origin, physical or mental disability, sexual orientation or gender identity expression. That same year, Nevada also enacted legislation (SB 236) that requires law enforcement agencies to establish early warning systems to identify officers who display bias indicators or demonstrate other problematic behavior. It also requires increased supervision, training and, if appropriate, counseling to officers identified by the system. If an officer is repeatedly identified by the early warning system, the law requires the employing agency to consider consequences, including transfer from high-profile assignments
or other means of discipline.

Another area of interest for states has been hiring a more diverse workforce in law enforcement and support agencies. For example, New Jersey SB 2767 (2020) required the state Civil Service Commission to conduct a statewide diversity analysis of the ethnic and racial makeup of all law enforcement agencies in the state.

Finally, at least one state addressed bias in policing through a state civil rights act. Massachusetts (SB 2963) established a state right to bias-free professional policing. Conduct against an aggrieved person resulting in decertification by the Police Office Standards and Training Commission constitute a prima-facie violation of the right to bias-free professional policing. The law also specifies that no officer is immune from civil liability for violating a person’s right to bias-free professional policing if the conduct results in officer decertification.

II. Pretrial Release and Prosecution

Risk Assessments

Recently, state laws have authorized or required courts to use pretrial risk assessment tools. There are about two dozen pretrial risk assessment tools in use across the states. 

Laws in Alaska, Delaware, Hawaii, Indiana, Kentucky, New Jersey and Vermont require courts to adopt or consider risk assessments in at least some, if not all, cases on a statewide basis. While laws in Colorado, Illinois, Montana, New York, Pennsylvania, Rhode Island, Virginia and West Virginia authorize or encourage, but do not require, adopting a risk assessment tool on a statewide basis.

This broad state adoption of risk assessment tools raises concern that systemic bias may impact their use. In 2014, U.S. Attorney General Eric Holder said pretrial risk assessment tools “may exacerbate unwarranted and unjust disparities that are already far too common in our criminal justice system and our society.”

More than 100 civil rights organizations expressed similar concerns in a statement following a 2017 convening. The dependance of pretrial risk assessment tools on data that reflect systemic bias is the crux of the issue. The statement highlights that police officers disproportionately arrest people of color, which impacts risk assessment tools that rely on arrest data. The statement then set out key principles mitigating harm that may be caused by risk assessments, recognizing their broad use across the country.

The conversation about bias in pretrial risk assessments is ongoing. In 2021, the Urban Institute published the report “Racial Equity and Criminal Justice Risk Assessment.” In the report, the authors discuss and make recommendations for policymakers to balance the use of risk assessment as a component of evidence-based practice with pursuing goals of reducing racial and ethnic disparities. The authors state that “carefully constructed and properly used risk assessment instruments that account for fairness can help limit racial bias in criminal justice decision-making.”

Academic studies show varied results related to the use of risk assessments and their effect on racial and ethnic disparities in the justice system. One study, “Racist Algorithms or Systemic Problems,” concludes “there is currently no valid evidence that instruments in general are biased against individuals of color,” and, “Where bias has been found, it appears to have more to do with the specific risk instrument.” In another study, “Employing Standardized Risk Assessment in Pretrial Release Decisions,” the authors, without making causal conclusions, find that “despite comparable risk scores, African American participants were detained significantly longer than Caucasian participants … and were less likely to receive diversion opportunity.”

In a recent report titled “Civil Rights and Pretrial Risk Assessment Instruments,” the authors recommend steps to protect civil rights when risk assessment tools are used. The report underscores the importance of expansive transparency throughout design and implementation of these tools. It also suggests more community oversight and governance that promotes reduced incarceration and racially equitable outcomes. Finally, the report suggests decisions made by judges to detain should be rare, deliberate and not dependent solely on pretrial risk assessment instruments.

States are starting to regulate the use of risk assessments and promote best practices by requiring the tool to be validated on a regular basis, be free from racial or gender bias and that documents, data and records related to the tool be publicly available.

For example, California (2019 SB 36) requires a pretrial services agency validate pretrial risk assessment tools on a regular basis and to make specified information regarding the tool, including validation studies, publicly available. The law also requires the judicial council to maintain a list of pretrial services agencies that have satisfied the validation requirements and complied with the transparency requirements. California published its most recent validation report in June 2021.

Similarly, Idaho (2019 HB 118) now requires all documents, data, records and information used to build and validate a risk assessment tool to be publicly available for inspection, auditing and testing. The law requires public availability of ongoing documents, data, records and written policies on usage and validation of a tool. It also authorizes defendants to have access to calculations and data related to their own risk score and prohibits the use of proprietary tools.

Pretrial Release

A recent report from the U.S. Commission on Civil Rights evaluates the civil rights implications of pretrial release systems across the country.

Notable findings from the report include stark racial and gender disparities in pretrial populations with higher detention rates and financial conditions of release imposed on minority populations. The report also finds that more than 60% of defendants are detained pretrial because of an inability to pay financial conditions of release. 

States have recently enacted legislation to address defendants’ ability to pay financial conditions of release, with at least 11 states requiring courts to conduct ability-to-pay considerations when setting release conditions. NCSL’s Statutory Framework of Pretrial Release report has additional information about state approaches to pretrial release.

Prosecutorial Discretion

Prosecutorial discretion is a term used to describe the power of prosecutors to decide whether to charge a person for a crime, which criminal charges to file and whether to enter into a plea agreement. Some argue this discretion can be a source of disparities within the criminal justice system.

The Prosecutorial Performance Indicators (PPI), developed by Florida International University and Loyola University Chicago, is an example of an effort to address this. PPI provides prosecutors’ offices with a method to measure their performance through several indicators, including racial and ethnic disparities. As part of their work to bring accountability and oversight to prosecutorial discretion, PPI has created six measures specifically related to racial and ethnic disparities in the criminal justice system. The PPI measures include the following:

  1. Victimization of Racial/Ethnic Minorities.
  2. Case Dismissal Differences by Victim Race/Ethnicity.
  3. Case Filing Differences by Defendant Race/Ethnicity.
  4. Pretrial Detention Differences by Defendant Race/Ethnicity.
  5. Diversion Differences by Defendant Race/Ethnicity.
  6. Charging and Plea Offer Differences by Defendant Race/Ethnicity.

Below is a table highlighting disparity data discovered through the use of PPI measures, gathered from specific jurisdictions.

Point of Discretion/Jurisdiction



Case Dismissal in the 13th Judicial Circuit, Tampa, Fla.
  • In December 2019, a 6% difference existed in dismissal rates between cases involving Black and white victims, with cases involving Black victims more likely to be dismissed.
Case Filing in Milwaukee County, Wis.    
  • In March 2020, an 11% difference existed in misdemeanor drug case filing rates between Black/Hispanic and white defendants, with Black/Hispanic defendants more likely to have their cases filed.
Pretrial Detention in the 4th Judicial District, Jacksonville, Fla.
  • In December 2019, a 3% difference existed in pretrial detention rates between Black and white defendants, with Black defendants less likely to be detained.
  • In December 2019, a 15% difference existed in pretrial detention rates between Hispanic and white defendants, with Hispanic defendants less likely to be detained.
Diversion in Cook County, Ill.
  • In March 2020, a 6% difference existed in diversion rates between Black and white felony defendants, with Black defendants less likely to be diverted.
  • In March 2020, an 8% difference existed in diversion rates between Hispanic and white defendants, with Hispanic defendants less likely to be diverted.
  • Over the three-year period, less than 8% of Black and Hispanic defendants are diverted, compared to 15% of white defendants.
Plea Offer in Cook County, Ill. 
  • In March 2020, a 14% difference existed in rates of pleading guilty to a lesser charge between Black and white defendants, with Black defendants more likely to plead guilty to a lesser charge.
  • In March 2020, a 12% difference existed in rates of pleading guilty to a lesser charge between Hispanic and white defendants, with Hispanic defendants more likely to plead guilty to a lesser charge.
Charging in Cook County, Ill. 
  • In March 2020, the difference in the average number of charges dropped after filing between Black and white defendants was 1.2, with Black defendants having more charges dropped.
  • In March 2020, the difference in the average number of charges dropped after filing between Hispanic and white defendants was 1.3, with Hispanic defendants having more charges dropped.

III. Incarceration

Incarceration statistics help paint a picture of the disparities in the criminal justice system. Significant racial and ethnic disparities can be seen in both jails and prisons. According to the MacArthur Foundation’s Safety and Justice Challenge website, “While Black and Latinx people make up 30% of the U.S. population, they account for 51% of the jail population.”

An October 2021 report from The Sentencing Project, an organization advocating for criminal justice reform, found that “Black Americans are incarcerated in state prisons across the country at nearly five times the rate of whites, and Latinx people are 1.3 times as likely to be incarcerated than non-Latinx whites.” At the time of the report, there were 12 states where more than half of the prison population is Black and seven states with a disparity between the Black and white imprisonment rate of more than 9 to 1.

To have a clearer sense of the racial makeup of who is incarcerated at any given time, some systems developed data dashboards to provide information on their jail populations. In Allegheny County, Pa., for instance, the jail data dashboard is publicly available and provides a range of information on who is incarcerated in the jail. The dashboard provides an up-to-the-day look at the race, gender and age of the jail population. According to the dashboard, on average from Jan. 1, 2019, to mid-November 2021, 65% of individuals in the jail were Black.

Dashboards may also be established by the individual state, though these generally look back over a specified time, rather than providing a close-to-live look at the jail population. Colorado passed a law in 2019 (HB 1297) requiring county jails to collect certain data and report it to the state Division of Criminal Justice on a quarterly basis. That data is compiled in a publicly available Jail Data Dashboard. The dashboard includes information on the racial and ethnic makeup of jail populations in the state. In the second quarter of 2021, 88% of people incarcerated in jails in the state were white, 16% were Black, 2% were Native American and 1% were classified as “other race.” In the same quarter, ethnicity data for incarcerated people showed 67% were non-Hispanic, 33% were Hispanic and 9% were classified with “unknown ethnicity.”

Pennsylvania’s Department of Corrections has an online dashboard providing similar information for the state prison population. The dashboard shows Black people make up 12% of the state’s overall population but 44% of the population in state correctional institutions, while white people make up 74% of the state population and 45% of the state prison population. While dashboards themselves don’t reduce disparities, they help create a clearer understanding of them.

Racial Disparities by Population of Race/Ethnicity in Pennsylvania

IV. Sentencing

Racial and ethnic disparities can also be seen in the sentencing of individuals following a criminal conviction. The use of sentencing enhancements and federal drug sentencing both provide examples of the disparities in sentencing.

Sentencing enhancements in California have been found to be applied disproportionately to people of color and individuals with mental illness according to the state’s Committee on Revision of the Penal Code. More than 92% of the people sentenced for a gang enhancement in the state, for instance, are Black or Hispanic. The state has more than 150 different sentence enhancements and more than 80% of people incarcerated in the state are subject to a sentence enhancement.

In response to recommendations from the committee, AB 333 was enacted in 2021 to modify the state’s gang enhancement statutes by reducing the list of crimes under which use of the current charge alone creates proof of a “pattern” of criminal gang activity and separates gang allegations from underlying charges at trial.

Legislatures are currently taking many steps to increase their understanding of racial and ethnic disparities in the justice system. In some states, this has taken the form of racial and ethnic impact statements or corrections impact statements.

At least 18 states require corrections impact statements for legislation that would make changes to criminal offenses and penalties. These look at the fiscal impact of policy changes on correctional populations and criminal justice resources. A few states have required the inclusion of information on the impacts of policy changes on certain racial and ethnic groups.

Colorado has taken this approach. The state enacted legislation in 2013 (SB 229) requiring corrections fiscal notes to include information on gender and minority data. In 2019, the state passed legislation (HB 1184) requiring the staff of the legislative council to prepare demographic notes for certain bills. These notes use “available data to outline the potential effects of a legislative measure on disparities within the state, including a statement of whether the measure is likely to increase or decrease disparities to the extent the data is available.”

Other states with laws requiring racial and ethnic impact statements include ConnecticutIowa, Maine, New Hampshire, New Jersey, Oregon and Virginia. Additionally, Florida announced a partnership in July 2019 “between the Florida Senate and Florida State University’s College of Criminology & Criminal Justice to analyze racial and ethnic impacts of proposed legislation.”  Minnesota’s Sentencing Commission has compiled racial impact statements for the legislature since 2006, though this is not required in law.

States are also taking a closer look at racial disparities within criminal justice systems by creating legislative studies or judicial task forces. These bodies examined disproportionalities in the criminal justice system, investigated possible causes and recommended solutions.

In 2018, Vermont legislatively established the state’s Racial Disparities in the Criminal and Juvenile Justice System Advisory Panel. The panel submitted its report to the General Assembly in 2019. Part of the report recommended instituting a public complaint process with the state’s Human Rights Commission to address perceived implicit bias across all state government systems. It also recommended training first responders to identify mental health needs, educating all law enforcement officers on bias and racial disparities and adopting a community policing paradigm. Finally, the panel agreed that increased and improved data collection was important to combat racial and ethnic disparities in the justice system. The panel recommended “developing laws and rules that will require data collection that captures high-impact, high-discretion decision points that occur during the judicial processes.”

State lawmakers are well positioned to make policy changes to address the racial and ethnic disparities that research has shown are present throughout the criminal justice system. As they continue to develop a greater understanding of these disparities, state legislatures have an opportunity to make their systems fairer for all individuals who encounter the justice system, with the goal of  reducing or eliminating racial and ethnic disparities.