UW News

August 28, 2018

New study finds police-related fatalities may occur twice as often as reported

UW News

police lights on car

A study led by the University of Washington and Cornell University uses new data sources to determine the likelihood of dying at the hands of police.


Every day in the United States, an average of nearly three men are killed by police.

This accounts for 8 percent of all homicides with adult male victims — twice as many as identified in official statistics, according to a study by the University of Washington and Cornell University.

The study, published Aug. 8 in the American Journal of Public Health, also shows that the risk of being killed by police, relative to white men, is 3.2 to 3.5 times higher for black men, and between 1.4 and 1.7 times higher for Latino men. Researchers determined these probabilities with six years’ worth of data from Fatal Encounters, a source that collects information from journalists, activists and researchers through public records and media coverage. This method is more reliable than police departments’ own reports, according to the federal Bureau of Justice Statistics.

The study was led by Frank Edwards, a postdoctoral associate with Cornell’s Bronfenbrenner Center for Translational Research and co-authored by Michael Esposito, who did this work as a graduate student in sociology at the University of Washington.

“Official statistics show that deaths attributable to legal intervention by police account for close to 4 percent of all homicides with adult male victims,” Edwards said. “We estimated that over this period, police were responsible for about 8 percent of all U.S. homicides with adult male victims — or 2.8 per day on average.”

Past work on police-involved mortality has been limited by the absence of systematic data, Edwards said. Such data, primarily collected through the Bureau of Justice Statistics’ Arrest-Related Deaths program or the FBI’s Supplementary Homicide Report, are widely acknowledged as unreliable due to limited scope and voluntary data reporting.

“Police departments are not required by law to report deaths that occur due to officer action and may have strong incentives to be sensitive with data due to public affairs and community relations,” he said. “Effectively, we don’t know what’s happening if all we look at is the official data.”

See a related story on KUOW.

Esposito, who is now a postdoctoral researcher at the University of Michigan, points to a 2014 article in ProPublica, which showed how some communities had not reported fatal shootings by police since 1997.

“Our ability to speak about police-involved deaths in the U.S. has really been hampered by shortcomings of official data sources,” he said. “We thought we could make a contribution here by just describing what police-involved homicides look like in our country.”

For this study, the researchers used public records and media reports to identify 6,295 adult male victims of police homicide over a six-year period between Jan. 1, 2012, and Feb. 12, 2018 — averaging about 1,028 deaths per year, or 2.8 deaths per day.

“Using fairly comprehensive data to establish a careful account of where police homicides are happening, and who they’re happening to, was necessary for building foundations around this issue that’s been so central to the country for the past few years,” Esposito said.

Of those 6,295 victims, 2,993 were white, 1,779 were black, 1,145 were Latino, 114 were Asian/Pacific Islander and 94 were American Indian/Alaska Native. During the six-year period identified in the study, black men were killed at the highest rate: at least 2.1 per 100,000 men. Latino and white men had lower rates: 1 and 0.6, respectively.

To look at deaths in specific regions, the researchers used Census designations. In the Pacific region, which includes Alaska, California, Hawaii, Oregon and Washington, black men in large metro areas were killed by a police at a rate of 3.4 per 100,000, Latino men at 1.1 per 100,000, and white men at 0.9 per 100,000.

Specifically in Washington, an average of a little more than three black men per 100,000 people are killed by police each year, compared to an average of about one per 100,000 among Latino men and fewer than one per 100,000 among white men.

The risk varies dramatically by location.

Although risk is high in large urban areas typically associated with police homicide, the majority of these deaths occur in less-populated regions, according to the data.

That was surprising, Esposito said.

“I think that there’s an idea that these events are constrained to large urban areas, but the data suggest that’s not the case,” he said. “That pattern motivates thinking about the dynamics of suburban and rural spaces, alongside large cities, when trying to consider why police-involved homicides occur.”

These data indicate that police-involved deaths aremore common and reaffirms that structural racism, racialized criminal-legal systems, anti-immigrant mobilizations and racial politics all likely play a role in explaining where police killings are most frequent and who is most likely to be a victim, according to Edwards.

“From a public health perspective, developing targeted interventions for sites with particularly high levels of or inequalities in police-involved mortality may serve as a productive framework for reducing them,” he said.

Hedy Lee, a former faculty member in the UW Department of Sociology who is now at Washington University in St. Louis, also contributed to the study.




Adapted from a Cornell University news release.


For more information, contact Esposito at esposm2@uw.edu or Jeff Tyson in the Cornell University Media Relations Office, jeff.tyson@cornell.edu.