What the data shows about police use of force by race

From 2018 to 2020, the share of Black Americans who experienced threats and nonfatal use of physical force during police contact increased, according to a report from the Bureau of Justice Statistics (BJS).^[1] The same is true for people in the multiracial, Native, or Pacific Islander American category.^[2]

Rates increased from 3.8% to 4.3% for Black Americans and from 2.5% to 2.7% for multiracial, Native, or Pacific Islander Americans.

Meanwhile, Hispanic Americans and Asian Americans saw a decrease from 3.4% to 2.4% and from 1.3% to 1%, respectively. Threats and nonfatal use of force rates for white Americans remained at 1.5%.

In 2018, Black Americans had the highest rates of police handcuffing (4.4%), threats of force (2%), cursing (1.7%), and pushing, grabbing, hitting, or kicking (1.6%) in cases when police initiated contact or the interaction was related to a car accident.^[3] Additionally, at 0.9%, Black Americans were slightly more likely than Hispanic Americans (0.8%) to have a weapon used on them by police. Hispanic Americans had the highest rate of being shouted at by police.

There is often a lag between government data collection and release. The Police-Public Contact Survey collected data in 2018 and 2020, but Bureau of Justice statisticians did not release the report analyzing the data until November 2022.

For a fuller picture of crime in the US, read about racial discrepancies in police enforcement and get more USAFacts data in your inbox by subscribing to our weekly newsletter.

---

[1] Because the 2020 survey was conducted between January 1 2020, and June 30 2020, and respondents were asked about their interactions with police during the prior 12 months, some, but not all police interaction captured by the data would have taken place during the initial COVID-19 lockdowns and the George Floyd protests.

[2] The Bureau of Justice Statistics grouped multiracial, Native, and Pacific Islander Americans into the same category for this report.

[3] 2018 data was used due to the exclusion of key 2020 data points caused by sample size issues.