back to top

Science Is Searching For Ways To Stop Police Killings

Hidden racial stereotypes, marginalized communities, and decades of distrust put both the police and the people they serve in danger. Can science provide the tools to defuse these tensions?

Posted on

Two black men and five Dallas police officers lie dead. There is fear, rage, and despair on the streets.

As America mourns, experts consulted by BuzzFeed News say that it’s time to rethink police training and procedures, to avoid situations that can easily erupt into conflict. This means recognizing that even good cops can kill, if they are placed in a bad situation, and realizing that everyone is influenced by a deadly racial stereotype about “threatening” black men — even people who are not overt racists.

Indeed, a decade-and-a-half of psychology research shows that most of us — whatever our race — would be dangerously prone to shooting first and asking questions later, when faced with a black male suspect. So we need find ways to keep officers’ fingers away from the trigger, experts say, and ask hard questions about why black communities dispute the legitimacy of law enforcement.

“The most alienated minority communities have the experience of both being over-policed and underprotected,” David Kennedy, director of the National Network for Safe Communities, run by the John Jay College of Criminal Justice in New York, told BuzzFeed News.

The dangers posed by our hidden racial biases have been laid bare by research using a first-person-shooter computer game, in which volunteers are shown images of young men set against backgrounds including city streets and parks. Half of the men are black, half are white; some carry a handgun, while others hold a harmless object like a cell phone or a wallet.

Joshua Correll, University of Colorado, Boulder

Spot the bad guy: In computer simulations, volunteers struggle to recognize that white men are holding guns, and black men are unarmed.

In the game, you must press the “L” key to shoot if the man is armed, and “A” to avoid shooting an unarmed man; hesitate and the task times out, costing you points. Over repeated studies, run by researchers led by Joshua Correll, a psychologist at the University of Colorado, Boulder, a consistent pattern has emerged: Most people are more likely to mistakenly shoot an unarmed man if he is black.

Even when volunteers make the right decision, their reaction times reveal their hidden bias. It takes people a few milliseconds longer to recognize that a black man is holding a harmless object, and similarly longer to recognize that a white man is carrying a gun. The race of the volunteer doesn’t seem to make much difference — deep down, Americans of all races seem to view black men as more threatening than whites.

It’s part of a wider stereotype associating blacks with violence and crime: In other research, volunteers shown an image of a black face became primed to recognize distorted images of knives or guns.

One encouraging result from Correll’s research is that cops perform better than the average American. In the computer simulation, they shoot at fewer unarmed men, and are no more likely to mistakenly kill blacks than whites. But this isn’t because police officers have banished their hidden bias: The cops’ reaction times reveal that they have the same problems recognizing that unarmed black men pose no threat. Rather, it seems that firearms training makes cops less trigger-happy.

“When you train officers, the number of errors decreases,” Phillip Atiba Goff of the John Jay College told BuzzFeed News

Still, a laboratory computer game is a far cry from the situation on the street, where police are often under extreme stress. A cop who is exhausted from a long shift and who fears that a suspect may be about to pull a gun may well behave very differently to an officer sitting at a computer terminal.

“If your system gets overwhelmed, you are likely to rely on stereotypes,” Debbie Ma of California State University, Northridge, told BuzzFeed News.

In experiments involving new police recruits, Ma and Correll have found that those who reported having few hours of sleep before playing the simulation were more likely to mistakenly shoot when shown images of unarmed black men.

What’s more, officers from special units dealing with gang crime are just as biased on the first-person-shooter simulation as untrained members of the public — which suggests that regular encounters with black gang members can reinforce officers’ hidden racial biases.

So fatigue and fear among cops may help explain why black men are disproportionately likely to be shot by police.

Good research into racial bias in actual police work is seriously lacking. But Goff, who is president of the Center for Policing Equity, believes that overcoming the insidious effects of racial stereotypes requires policies that relieve the pressure on officers to make split-second decisions when under extreme stress.

In unpublished research, Goff worked with the Las Vegas Metropolitan Police Department to test a policy on one of the most stressful situations officers face: Chasing a suspect on foot. In late 2011 the Las Vegas police introduced a policy that said that officers who led a foot pursuit chase shouldn’t be the ones who lay hands on the suspect to make the arrest, unless absolutely necessary.

Cops who chase suspects down unfamiliar streets are “hopped up on adrenaline” and scared of being ambushed, said Goff. Pulling them back from making the arrest reduced the department’s use of force by 23%, he said.

“What we really need to focus on in training is how to make the tactical decisions so that the split-second things come up less frequently,” Goff said.

Other research suggests that the tensions between police and black communities can arise from how officers talk to people they stop on the street. Police recruits typically get intensive training on when they are legally allowed to stop and search — but little instruction on how to manage that interaction.

That could be a crucial mistake, according to research led by Tracey Meares and Tom Tyler of Yale Law School. When volunteers from across the country were shown police videos of street encounters, people were less worried by whether the stop was legally justified, and more concerned with what the researchers call “procedural justice” — whether the cops seemed to be respectful, whether they listened to what the suspect had to say, and whether they explained their decisions.

What’s missing, said experts consulted by BuzzFeed News, is good research on whether providing cops with better training on how to handle such situations will ratchet down the tension between police and black communities. “The state of the scientific art is that we don’t know very much about what works,” Kennedy said.

But answers are coming from a project Kennedy heads, backed by the Department of Justice. Under the National Initiative for Building Community Trust and Justice, pilot studies are under way in six cities to find out whether better training in procedural justice will make a difference to how the police are perceived.

‘We need to increase the service and protection from criminal justice in those communities while reducing the unintended impact of over-policing,” Kennedy said. “That’s the struggle.”

  • Want to challenge your own hidden racial biases when placed in a cop’s shoes? Try an online version of Correll’s simulation here.

Here’s What We Know About Race And Killings By Police

Attorney For Officer Who Fatally Shot Philando Castile Says He Reacted To Display Of Gun

Graphic Video Shows Baton Rouge Police Fatally Shooting Man On The Ground


Peter Aldhous is a Science Reporter for BuzzFeed News and is based in San Francisco. His secure PGP fingerprint is 225F B2AF 4B8E 6E3D B1EA 7F9A B96E BF7D 9CB2 9B16

Contact Peter Aldhous at peter.aldhous@buzzfeed.com.

Got a confidential tip? Submit it here.