Can Cops Unlearn Their Unconscious Biases?

“Implicit bias” training is spreading to departments around the country, the theory being it can influence officer behavior on the street. But it’s still not clear that the classes actually work.

A police officer stands with one hand behind his back, with his handcuffs, flashlight, and radio visible.
Mark Makela / Getty

SALT LAKE CITY—On a clear, cold morning this fall, Police Sergeant Scott Stuck stood at the front of a long, high-ceilinged room, visible to a reporter outside, his brow furrowed as he faced more than a dozen officers.

Quietly and methodically, Stuck was trying to convince them of an idea that goes against much of what they’ve been taught it means to be a cop—one that at first blush seems to validate the deepest criticisms activists level against police, and even call into question whether the fundamental equation of policing itself might be flawed.

On that cold morning, Stuck was trying to convince the officers that they are biased.

And although nearly everyone I spoke with about the training seemed to take pains to avoid one word—“racism”—that’s the kind of bias that was on everyone’s mind.

“It’s just something that you don’t admit. It’s the elephant in the room that you just don’t talk about,” said Sergeant Sam Wolf, another trainer in the department. “You don’t talk about discrimination and bias, because then … people might think cops are discriminatory, they’re biased. If we admit that, then what does it mean about how we serve the public?”

The morning’s class was only the third to be offered in Salt Lake, but it puts the department among a growing number adopting similar courses, generally referred to as implicit-bias training. In at least four states, state police and their training academies offer the classes, as do local departments in more than a dozen.

But even as the classes spread, it’s not clear whether they actually work. Few specific guidelines exist for what courses should include, how the material should be taught, or how to measure its effects. Indeed, little data exist about their efficacy over the long term. The Obama administration’s Task Force on 21st Century Policing included implicit-bias training on its list of best practices for law enforcement, but without specifics. That ambiguity leaves each agency to decide what the classes should look like—and whether they’re succeeding.

At their root, the trainings spring from one basic proposition: that unconscious biases—including those linked to factors like economic class and gender, but especially racial biases—are the inevitable product of growing up in a society where stereotypes are woven into the fabric of everyday life. Beneath the surface of the conscious mind, biases influence how people frame and interpret those around them—from whether a smile is shy or sarcastic, to whether a hand is reaching for a wallet or a gun.

Crucially, what makes implicit biases implicit—and important—is the fact that they can exist despite people’s best intentions. Over and over, researchers have found that subjects can consciously embrace ideas of fairness and equality and yet, on tests that measure subconscious tendencies, still show a strong propensity to lean on stereotypes to fill in the blanks about people they don’t know. And while biases can influence just about everything, perhaps nowhere do they have the potential to have such a profound, immediate effect as in policing, where a split-second decision by a stranger—exactly the kind of situation where biases come into effect—can mean life or death.

In Salt Lake, the trainers told me their goals can be boiled down to one hope: that by showing officers evidence of the existence of bias, and by encouraging them to discuss their own, seeds of self-awareness will ultimately grow—and change behaviors on the street.


The ideas behind the trainings aren’t particularly revolutionary. The twin assumptions they hinge on—that osmotically absorbing bias is an inherent feature of growing up in America, and that admitting its presence is the first step toward its eradication—have become well accepted in progressive circles. But in the context of law enforcement, accepting bias as inevitable also raises questions about the fundamental equation of policing, which is built on the notion of the fair cop.

Not a judge, the American police officer is nonetheless given—and expected to wield judiciously—substantial power over which crimes to pursue and, crucially, whose behavior to scrutinize for signs of criminality. In turn, that discretion ideally allows police to ignore minor offenses and spend their time pursuing people who pose real threats to their communities. In other words, impartiality, combined with broad power, should produce impartial enforcement of the law.

But that equation is clearly flawed, a fact manifest not only in the deaths of black men at the hands of police, but in the disproportionate presence of police in the day-to-day lives of people of color. Black Americans, in particular, are a third more likely than whites to be stopped by police, up to a third more likely to be ticketed during a stop, and three times more likely to be searched. They’re also as much as three times more likely to be subject to the use of force, with one study finding that unarmed black people are almost three-and-a-half times more likely to be shot by police than unarmed whites.

Salt Lake is not immune to police violence against African Americans, who make up just 3.5 percent of the city’s population. A coalition of civil-rights groups requested the police department  implement implicit-bias training, along with other reforms, after police shootings in 2014 and 2015. The city drew media attention in 2016 when it passed the year-and-a-half mark since its last fatal police shooting. But that run was broken in August, when an officer shot and killed 50-year-old Patrick Harmon.

While linking any one incident to officer prejudice is difficult, all signs point to implicit bias playing at least some role in the broader disparity between police interactions with whites and people of color. In a 2007 study that received wide notice, a team of researchers led by University of Colorado neuroscience and psychology professor Josh Correll ran officers through a simulator that repeatedly challenged them to identify what objects black and white subjects were holding in their hands, and then decide whether to shoot. In repeated trials, officers were quicker to decide to shoot armed black subjects than armed whites, and took longer deciding to spare unarmed blacks. There were virtually no racial discrepancies in the officers’ mistaken shots—but later studies found bias crept in when officers were subjected to mental stress.

While Correll’s work gained notice, it wasn’t the first to spot bias: Studies dating back to at least 2001 have found police officers and civilians are consistently more likely to associate black faces with criminality, to misidentify common objects as weapons after being shown photos of black faces, and to label photos of black people as threatening.

And “this isn’t something that’s just wrong with police officers,” said Kimberly Kahn, a social-psychology professor at Portland State University who studies bias in policing and has helped formulate implicit-bias courses. “It would actually be pretty weird if police weren’t subject to these problems” the way other Americans are.

But, Kahn added, the power cops wield obligates them to better combat their biases, and that begins with acknowledging one simple, central fact: Everyone is biased, and I am, too.


Back in Salt Lake, that admission seemed to be the hardest for officers to make. I wasn’t allowed inside the classroom. The lieutenant in charge of training at the department told me that the proceedings are so sensitive—and the openness they’re trying to cultivate so profound—that my presence could throw the learning process off its rails. But a few of the officers were willing to talk afterward.

More than one referred to the landmark “doll study” that was cited in Brown v. Board of Education. According to its findings, children, including black children, reliably associated positive characteristics with light-skinned dolls and negative characteristics with dark-skinned ones. Officers watched a video reproduction of the study, which is meant to show just how baked-in biases are. “I looked at them and thought, ‘If they’re having it at that age, [and] I have a lot more life experience than they do, I know I must have it,” said Officer Andrew Sylleloglou, a 10-year veteran. “That was quite the eye-opener.”

Similarly, a second officer, Jeffrey Denning, called the study tragic. But he also told me he still viewed racism as a “buzz topic,” inflated beyond its true importance.

Even after taking the course, Denning said he was still skeptical of the class’s central idea: That despite his best intentions—what he described as a deep desire to be fair—negative biases about certain groups could affect his thinking and behavior, and that bias can play a role in the treatment of minorities by police. Instead, he said, disparate use of force on black people seems to be more likely linked to their “higher degree of crime.”

Denning at first said the course had opened his eyes only to positive biases, toward groups like senior citizens, that might lead him to mistakenly let down his guard.

But when asked point-blank if the course had at all opened his eyes to any negative biases he has, Denning had to pause. Finally, he said his answer would depend at least in part on who was asking, whether he was in uniform, and whether—as he put it—“they were anti-cop people.” Then, with a tone almost of resignation, Denning circled back.

“Just me speaking to you, yeah. Unfortunately I do,” he said. “Just like you, and every other person in the world.”

The crux of surmounting bias, said both researchers and the officers themselves, is setting aside the expectation of what police are supposed to be—“kind of the lady justice,” as Sylleloglou put it—and with it the dissonance between what they’ve learned and the way they see themselves. “The hard part is bringing [bias] out into the open and talking about it. And I‘m talking about just amongst peers,” Sylleloglou said. “Imagine talking about it with community members. … I’d be afraid what they’re going to think of me.

“We want to be looked up to,” he added. “We want people to think that we are 100 percent always trying to do the right thing. Maybe someone who’s not in a position of authority like an officer—it might be easier for them to admit that they have biases.”


Expanding officers’ awareness of their prejudices might seem an intuitive exercise. But a lack of standards for implicit-bias training—namely, what exactly courses should include and how to monitor their impact—means no one really knows how effective they actually are, even as they are adopted by more and more departments.

To make things even more confusing, debate also swirls around which tests to use to measure bias, the degree to which implicit biases are truly unconscious, and even the strength of the link between bias and behavior.

“Some of these trainings are based on wishful thinking and intuition,” said Patricia Devine, a psychology professor at the University of Wisconsin who runs a lab focused on prejudice. At least two national groups offer the trainings, and departments can and do also put together their own.

If agencies skip key steps, Devine said, like arming participants with concrete strategies for monitoring their own biases, they won’t work. “By doing this kind of training and not having standards for effectiveness ... you don’t know if somehow they left an ingredient out,” she said.

“The trainings have great potential, but they need to be tied to assessment,” said Phillip Goff, a City University of New York professor and the head of the Center for Policing Equity, a think tank based at the university that helps departments set up trainings. Assessments need to go well beyond feedback from participants, he said: They should include rigorous testing after classes finish to see if officers’ reactions, behavior, or perceptions were actually changed by the material. While the courses designed by his group include that testing, few others do, he added.

Correll, the Colorado researcher, described the risk of uneven standards another way. A training session could include all the right information, and create a basic awareness that bias exists, but still fall short of actually bringing officers across the threshold of acknowledging their own. And a program might look from the outside like it was succeeding, and perhaps even cause changes in officer behavior—but only in situations where they have time to think their actions through.

“That’s cool,” Correll said. “But when someone jumps out from behind the bushes and pulls something from his waistband, that’s not the way the brain is working.”

There’s an additional risk to the lack of standardization, he said: If government leaders and the public treat this first wave of courses as a referendum on whether bias can be mitigated through training, mediocre results from even some departments could fuel criticism of the courses, and doubts that bias reduction itself is a worthwhile pursuit.

Devine and others did sound hopeful notes, emphasizing that law-enforcement leadership with whom they’ve spoken recognize the urgency of the issue. And despite some initial defensiveness, Devine and others said, many officers seem to warm to the training for perhaps the same reason they were attracted to policing in the first place: because they genuinely value fairness.

That sense of hope was reflected by at least one activist in Salt Lake, where Black Lives Matter Utah organizer Lex Scott described her feeling that the trainings are a good thing.

“You can’t have people saying they don’t see color,” Scott said. “I’m a black woman, and if you’re telling me you don’t see color, you’re telling me you don’t see the injustices that I’ve faced, the struggles that I have, … and you definitely can’t celebrate my culture.”

For her to feel like the classes are succeeding, Scott added, she’d need to see two things: First, that people of color are involved in the teaching (Stuck said he believes none are), and second, that the teachers tell the unvarnished truth about how minorities fare at the hands of police.

“They shouldn’t sugarcoat it at all,” she said. “You have those biases, you have to face them.”

Bias is paradoxical. It can drive people’s actions one way, while their conscious minds strain in the opposite direction. It can be a feeling people hardly notice, but one that can profoundly change how they see the world. And it’s fairly straightforward to explain, but difficult to accept. One way around that difficulty would be to water down the hardest parts, like the idea that racial biases in particular are widespread, and that they affect the everyday application of the law. Framing bias only in positive terms—as a warning, for example, not to mistakenly go easy on senior citizens and soccer moms—would likewise be an easy way to make the material more palatable to an audience that is deeply invested in a vision of itself as impartial. Yet that framing conveniently elides the hardest, most important truth of all: The racial divide in policing is real, it is ugly, and it grows from a thing that has worked its way into all Americans’ minds.