Minneapolis Police Department car
Credit: Tony Webster/Flickr

Within the array of stereotypes inflicted upon black people in America over many generations, the image of violence stands out. From slavery on, they have been seen as dirty, ugly, stupid, immoral, alien, and dangerous. These fictions become more or less prominent with time and circumstance, but they never quite die away. Even when they are not translated into law or practice, they can lurk as “implicit bias” that contaminates behavior. The label “dangerous” is especially pernicious.

Much of the brutal policing now being protested appears driven by the expectation that black men and women will be violent. That supposed trait appears regularly in surveys and simulations. It is an old prejudice ingrained in American society, readily activated by stress and triggering an officer’s split-second fear, which sometimes leads to a shooting, but more often to warrantless frisks and auto searches, handcuffing, and non-lethal physical force.

The role of racial thinking is difficult to measure precisely. Thoughts and actions do not inevitably coincide, and official statistics record end results, not causes. During traffic stops producing no arrests over a thirteen-month period in 2013-14, for example, police in Oakland, California, handcuffed 1,466 African Americans but only 72 whites, Stanford psychologists reported. While 72 percent of the department’s officers had handcuffed a black person who wasn’t arrested, 74 percent had never done so to someone who was white. Handcuffing black Americans was “a script for what is supposed to happen,” the study concluded, a routine presumably based on the violent stereotype but maintained as standard practice. “Norms are a significant driver of behavior,” the psychologists observed. Other experts have seen that rules issued from on high cannot readily overcome a police department’s culture.

Where the stereotype of violence exists, it can be reinforced by the slice of society the police usually encounter—the criminal suspect, not the upstanding citizen. That daily experience becomes a filter through which a world of hostility and risk passes into perceptions, corrupting context and distorting reality. A 2019 study of 100 million traffic stops nationwide found blacks more likely than whites to be stopped, but less so after dark when an officer couldn’t see a driver’s race. Blacks who were pulled over were more likely than whites to be searched. A 2016 examination of files and mug shots determined that “the Whiter one appears, the more the suspect will be protected from police force.”

Off-duty officers of color trying to stop crimes are more at risk of being shot by fellow officers than their white counterparts—comprising 10 of the 14 killed between 1995 and 2010, according to a nationwide study commissioned by the New York governor’s office. “Inherent or unconscious racial bias plays a role in ‘shoot/don’t-shoot’ decisions made by officers of all races and ethnicities,” the study declared.

The words, “all races and ethnicities” deserve attention. When a group of psychologists examined fatal shootings from 2015, they discovered that “black officers were just as likely to shoot black citizens as white officers were,” said Joseph Cesario of Michigan State University. The report put it another way: “As the proportion of White officers in a fatal officer-involved shooting increased, a person fatally shot was not more likely to be of a racial minority.” But it added a possible reason: White victims were more prevalent because more whites than blacks were mentally ill, and some whites were actually trying to get shot, committing “suicide by cop.” Furthermore, black-on-black shootings rise when more black officers are assigned to black neighborhoods, as they frequently are in big cities.

A question is whether black officers are as quick as some whites to shoot unarmed blacks, and whether they share anti-black stereotypes. “People can have biases against their own demographic groups,” said Lorie Fridell, a criminologist at the University of South Florida. “Women can have biases about women. Blacks can have biases about blacks.” But while minority group members often internalize stereotypes held by the majority, how those figure in the mix of motives for violence by black police is far from clear. Class is probably a factor. Some poor inner city residents resent black officers and let them know it. Some black officers return the contempt.

Others can reduce tensions. I saw this while researching books on civil liberties. When I spent nights reporting on an undercover narcotics unit in Washington, D.C., I watched anger build among white cops as hostile black residents hassled and berated them for arresting drug dealers. As tempers reached a pitch, one black officer intervened skillfully to talk both sides down. The unit called him “the reverend.”

The vast majority of police officers are white males with less than a college degree, the cohort that voted strongly for Donald Trump, who was endorsed by their union, the Fraternal Order of Police. They generally hold views on race typical of the country’s white conservatives, according to a 2016 survey of 8,000 officers by the Pew Research Center.

White cops were nearly unanimous (92 percent) in believing that the country had made the changes needed to guarantee equal rights for blacks; only 6 percent thought that continued change was needed. Their black colleagues said the opposite: only 29 percent saw adequate change and 69 percent  wanted more. A similar divide existed on whether fatal encounters between blacks and police were isolated incidents (72 percent of white officers, 43 of blacks) or part of a broader problem (27 percent of whites, 57 percent of blacks).

Only 11 percent of law enforcement academies overall require trainers to have four-year college degrees, according to President Obama’s Task Force on 21st Century Policing. The qualification is required in just 7 percent of state police academies, 4 percent of academies for municipal police, and none of those for sheriff’s deputies.

The militarization of police, both in training and equipment, has been a major concern of reformers. Officers’ military backgrounds are correlated with the use of force. About 6 percent of the country’s population has served in the military, but 19 percent of the police. Indeed, policing is the third most popular job following active duty, after driving trucks and the general category of managing.

Dallas officers who had been deployed overseas with the military fired their weapons 2.9 times more than cops with no military service, and those who hadn’t been deployed were 1.94 times as likely, the University of Texas School of Public Health reported in 2018. In 2017, the Marshall Project found that use-of-force complaints were more common against Boston and Miami police who’d been in the military. In Albuquerque, one-third of the fatal police shootings from January 2010 to April 2014 were by cops with military backgrounds.

Attempts to quantify racial bias using simulations have produced contradictory results. Joshua Correll, a psychologist at the University of Colorado, has run tests since 2000 with a video game showing images of black and white young men, some holding guns, others cellphones or soda cans. Participants tended to shoot armed blacks more often and sooner than they did armed whites, and to decide more quickly to refrain from shooting unarmed men if they were white. In a later study, police officers reacted most quickly, and correctly, to armed blacks and unarmed whites—that is, they shot the blacks and did not shoot the whites. Further simulations showed more racial bias in the shooting decisions by officers of special units who had contact with minority gang members. Training failed to diminish bias.

Other studies, however, have failed to document a correlation between bias and action. Even when officers hold the stereotype of blacks as violent, they don’t always behave accordingly, at least in the simulations.

Two experiments in police simulators recorded brain waves and physiological signs of a heightened threat perception when black “suspects” were pictured. But neither that study nor a more intricate simulation found that racial bias translated into faster, more frequent shootings of blacks.

The simulation, in 2012-13 by Washington State University, enlisted 80 officers in Spokane, 76 of whom were white, who were paid to participate in uniform and with infrared guns that looked like actual weapons—an attempt to get as close to reality as possible. The cops took written and oral exams designed to document prejudice, including the Harvard Implicit Association Test, which documents participants’ propensity to link pictures of black and white faces with weapons. Ninety-six percent of the officers registered implicit racial bias; 78 percent associated blacks with weapons, none associated whites.

In four sessions with six scenarios each, the officers were shown videos of actors playing people in similar situations and clothing. The cops were more careful with blacks. They were three times less likely to shoot unarmed blacks than unarmed whites. (The ratio was corrected for the smaller number of black scenarios; in the raw numbers, 54 unarmed whites and two unarmed blacks were shot.) The cops also took about two-tenths of a second longer to decide to shoot armed blacks than armed whites, a delay that researchers regarded as significant.

This was called “a counter bias” by Lois James, one of the investigators. The report was ultimately titled “The Reverse Racism Effect.” She postulated that the hesitation to shoot blacks grew out of “people’s concerns about the social and legal consequences of shooting a member of a historically oppressed racial group…paired with the awareness of media backlash that follows an officer shooting a minority suspect.”

She might be right, but a simulation is hardly reality. It contains no danger, no risk of death, no adrenalin flowing, no actual fear.

In the Navy, I attended a survival school that concluded with being “captured,” crammed uncomfortably into a wooden box, and “interrogated” with the threat of “torture.” It was the final lesson of the last day of training, so I was extremely brave and recited only my name, rank and serial number.

The simulation’s results do not coincide with the Washington Post’s database of fatal shootings by police, who have killed unarmed blacks since 2015 at a rate two to three times the African American representation in the country at large. African Americans constitute 13 percent of the population but between 25 and 40 percent of the unarmed citizens shot to death by police.

The fitful start of a tentative downward trend is visible, both in absolute numbers and percentages. But it’s too early for optimism. The table above shows deaths recorded to June 11, 2020, and this year’s numbers are already on track to match or exceed last year’s. Furthermore, they do not include deaths that don’t involve firearms, hence George Floyd is not in the figures. Nor is Freddie Gray, who died in a Baltimore police van in 2015 from injuries to his spinal cord. How many others have escaped the database is unknown.

What is known is the resilience of racial stereotyping, conscious and unconscious, which corrupts society until, as in the civil rights movement, it mobilizes enough Americans to act.

David K. Shipler

David K. Shipler is a Washington Monthly contributing writer; Pulitzer Prize–winning author of seven books, including Russia: Broken Idols, Solemn Dreams; and former Moscow bureau chief for The New York Times. He blogs at The Shipler Report and cohosts the podcast Two Reporters. Follow David on Twitter @DavidShipler.