Skip to Main Content

Amid growing scrutiny on the use of force by police, less information exists about the number of cases of people injured but not killed in those encounters — cases like the shooting last week of unarmed behavioral therapist Charles Kinsey.

Now, in a first-of-its kind study, researchers have examined the racial breakdown of individuals injured or killed by officers, and the results surprised them.

“I wanted to know whether minorities were more likely to be injured in interactions with the police,” said lead author Ted Miller, a researcher at Pacific Institute for Research and Evaluation. “I was surprised that they weren’t.”


The study found that, once an individual is arrested or stopped and searched, their likelihood of being hospitalized for injuries or being killed by a police officer does not vary by race.

But African-Americans are more likely than whites to be arrested or stopped and searched. So as a result, an individual African-American is more likely to be hospitalized from an injury by police than a white individual is.


The study comes on the heels of a similar and controversial study by Harvard economist Roland Fryer, a draft of which was published online earlier this month. Fryer’s paper found that blacks were more likely than whites to be on the receiving end of physical force but not more likely to be shot, when adjusting for their frequency of police encounters.

Outside researchers said that Miller’s paper, published Monday in Injury Prevention, is notable for analyzing the injuries inflicted by police, and their relationship to race, in a new way.

“The injury estimates are new, and that’s a significant development,” said Charles Epp, a University of Kansas professor who has researched race and police.

But criminal justice researchers warned against drawing far-reaching conclusions from the study.

“I think the study is a noble but incomplete effort to address the question, in part owing to nothing that the researchers did, but simply because of the limitations of the data sets they use,” said Jeffrey Fagan, a professor at Columbia Law School who has long studied stop-and-frisk in New York City.

In many cases, research has shown, people of color are stopped by police in situations where white people might not be. Epp’s research has found that blacks were not more likely than whites to get pulled over for speeding by 10 miles per hour or more, but they were more likely to get pulled over for speeding only a couple of miles per hour.

An equal rate of injuries per stops and arrests doesn’t mean that those injuries occurred in the same type of encounter, or that they were deserved or proportional to the behavior by the injured person.

“One possible interpretation [of Miller’s study] that is easy to make is that people are treated the same way once they are arrested,” said Joscha Legewie, a sociology professor at Yale University. “That’s not necessarily the case, based on the information we have.”

Data for this study came from a variety of sources, including arrest counts from the FBI, a national-level survey of people’s interactions with the police from the Bureau of Justice Statistics, and newspaper databases tracking police shootings compiled by the Guardian and the Washington Post. Miller combined that with data about hospital admissions from the Department of Health and Human Services.

Specifically he was looking at injuries like burns from tasers, watering eyes from tear gas, or wounds from explosives.

Putting all of that together, he found that the rate of hospital admissions for non-Hispanic whites, blacks, and Hispanics were basically the same, relative to the number of arrests or stops — about 1.5 per 10,000 stops or arrests.

But the data was less than ideal. For one thing, it didn’t capture ER visits, since the emergency rooms weren’t recording patient race for all visits. But many more people hurt by cops go to the ER and then home (about 49,000 in 2012) than are admitted to the hospital (about 2,500 individuals).

Finally, while the study looks at data on a national level, researchers say there may be much local variation that isn’t captured.

There is much follow-up work to be done. Fagan would like to see the analysis validated for cities that have good data on police stops, like his home turf, New York City. Miller said that looking at state-by-state variation would be a good next step — but that he doesn’t have any funding to do so. This analysis, in fact, was conducted mostly on nights and weekends, including some holidays.

“These people need this information,” Miller said. “They really need it. And I’m going to go look at it.”

Some of that information isn’t as much about policing as it is about documenting the reasons behind hospital visits. Currently, while some injuries — like shootings — can be documented as police-related, there’s no standard way to specify that someone was bitten by a police dog, or falls after being pushed by a police officer, Miller said. There isn’t even a code for being shot with a taser — to find those cases, Miller looked for cases of electric shock that also involved law enforcement.

“I think we’re more likely to make good progress dealing with hospitals than we are dealing with police data,” said Frank Baumgartner, a political science professor at the University of North Carolina, Chapel Hill. “The hospitals have no particular reason to try to minimize the apparent occurrence of the event.”

Meanwhile, the study may inform the national conversation about policing. Sean Smoot, director of the Police Benevolent and Protective Association of Illinois, and member of President Obama’s Task Force on 21st Century Policing, said that this study stands “strongly in opposition to the current narrative or assumptions” that minorities are subjected to more violence by virtue of their race.

For his part, Miller’s thoughts on the subject have evolved. He called in the paper for training “at-risk groups about appropriate behaviour during police stops.” But now, he says, he believes it’s primarily on law enforcement to change.

“The police are the ones who are doing this every day, and it’s much more important that they know how to de-escalate and that they practice it, and so on,” Miller said. “Breaking the racial pattern has to do with breaking the pattern of stops and arrests.”

  • So, bottom line, if you don’t find that data support the general idea that blacks are disproportionally affected by police violence, you find any possible caveats to undermine those data.
    That’s what I like about science, there is no just one right answer!

Comments are closed.