Police killings can be captured in data. The terror police create cannot.

The problem with policing is about more than violence. It’s about living a black life in fear.

The ongoing worldwide protest movement was incited by a police killing. The death of George Floyd, who was killed by a Minneapolis police officer, was the spark. Floyd’s death connected what had been disparate demonstrations over other police killings, like those of Breonna Taylor and Sean Reed, as well as those held demanding justice for Ahmaud Arbery, who was killed by white vigilantes, one a former officer. As these protests spread, so did demands for serious reforms, including calls to defund the police.

But police killings are not the whole story. The protests, and all the policy recommendations that have come with them, are pushing back against other systemic problems too.

Some of those injustices are specific and quantifiable: shootings, unfair traffic stops, arbitrary arrests. Others are vague but no less concerning: feeling you have no recourse for complaints about police, the calculus that can go into the decision to call 911, the sense that an investigation into a reported crime won’t be prioritized, the nervousness and fear that must be tamped down as one works to stay calm and keep an officer calm — all while wondering if you are living your final moments.

Not all of these problems can be measured. The number of police killings per year is a statistic that can be tabulated and broken down into easily digested parts: killings per region, per department, per time of day, per ethnicity. But how police make people feel is not quite as easily captured in data. There are ways to try — surveys asking whether officers make one tense or whether one trusts law enforcement — but such questions offer limited insight into what is causing those results or what effect they have.

Erik McGregor/LightRocket via Getty Images

Meanwhile, the issues behind the answers to those surveys do have a clear effect: They leave many black Americans frustrated with and fearful of police. That frustration and fear sometimes become anger of the sort that is driving the protests. Other times, it takes the form of pain and worry, feelings that animate the push for policing reforms. It can also transform into a kind of terror, pervasive and paralyzing, created by the sense that a negative — perhaps even deadly — encounter with the police can happen at any time, anywhere, for any reason.

That terror comes from observation and from experience. Many read of Botham Jean being killed by an officer while eating ice cream on his couch, or recently, of Rayshard Brooks being killed after falling asleep in his car. And so many have had a negative interaction with police themselves: 55 percent of respondents to Black Futures Lab’s 2019 Black Census — which polled more than 31,000 black adults — said they had.

But even this number does not tell the whole story. What were those negative interactions? Did it begin with an officer cursing at them, a beating, or an arbitrary arrest? And were these captured in police reports or on video by bystanders?

It’s difficult to say. Which makes it important not just to look at the data that exists, but to take a nuanced look at how policing is done — and what complaints black citizens have about how they are treated by law enforcement.

Police abuses — including killings — are a source of terror that erodes trust

Police killings of Americans of all ethnicities have trended slightly up recently — but down numerically — according to data collected by the advocacy group Mapping Police Violence. In 2019, the group found 1,098 people were killed by police, a number that works out to about three people per day. When looking at just black victims, there were 259 black people killed — which, if evenly distributed over time, would be nearly five a week.

While it’s true, according to the Mapping Police Violence data, that the high point in police killings of black people over the past six years came in 2015, when 305 black people were killed, that is cause for limited hopefulness: Sure, the number going down by one could mean that, for whatever reason — be it chokehold reforms or deescalation training — someone who would have been killed is still alive. But it is not cause for celebration; 259 people killed is still 259 lives lost.

Robyn Beck/AFP via Getty Images

The events that have incited international protests certainly make it feel as though there are five black people being killed by police each week. George Floyd was killed by now-former police officer Derek Chauvin in Minneapolis on May 25. Since then, the public has learned of the killings of Javier Ambler, Maurice Gordon, Manuel Ellis, Tony McDade, Momodou Lamin Sisay, and Rayshard Brooks. And there are still calls for justice for those who were killed shortly before Floyd, like Breonna Taylor and Sean Reed.

There is research that works to understand this data of police killings. Among the most prominent are studies by Harvard University economist Roland Fryer Jr., currently suspended from the institution for sexual harassment. Fryer’s work found no bias in police shootings and has been cited frequently of late, including by conservative commentators like Jason L. Riley and a Tulsa, Oklahoma, police officer making inflammatory remarks about the need to shoot more black Americans.

The results of Fryer’s work would seem to fly in the face of observed reality, and his methodology has been called into question by other researchers, who have found much larger disparities. As Vox’s Dylan Scott has explained, work by Frank Edwards, Hedwig Lee, and Michael Esposito found black men have a one in 1,000 chance of being killed by police, a probability just slightly higher than the average American’s chance of drowning.

But while a person may avoid water if they cannot swim, it is not possible to avoid police. Some of Fryer’s research supports that intuition: It did find that black people are three times more likely to experience force from officers than white people. The problems in Fryer’s work complicate this, but the disparity is clear observationally as well by spending time in black neighborhoods and listening to the narratives of black people.

The Portals Policing Project attempts to do that. Led by Johns Hopkins’ Vesla Weaver and Yale University’s Tracey Meares and Gwen Prowse, it uses a videoconferencing system set up in 14 neighborhoods in Chicago, Baltimore, Milwaukee, Newark, Los Angeles, and Mexico City to gather firsthand accounts of police interactions through conversations between individuals in those places.

Meares told CityLab the group embarked on the project amid the uprisings in Ferguson, Missouri, following Michael Brown’s killing by police, saying, “We were frustrated that journalists kept saying, ‘This is what people in Ferguson are saying about the shooting,’ without really knowing or studying people who live in those places in a systematic way.”

The solution, Weaver said in 2017, was “to actually listen to what members of these impacted communities said in their own terms — in interactions not with researchers, not with us, but in conversation with others distant physically, but from similar neighborhoods.”

A number of the more than 800 narratives they have gathered have been published in analyses based on the stories, which paint a rich picture of what policing feels like in those cities. Participants have spoken of the weight and effects of constant existential stress — of how it feels to see a police officer, of spikes of anger and fear and uncertainty that are sometimes punctuated by violence and detention, and other times result in being allowed to go free.

One black man from Baltimore spoke of his frustration at feeling the onus of deescalation is placed on the policed rather than the police, saying he shouldn’t have “to worry about, oh if I get smart with this cop, he might punch me and put false charges on me.” The police, he believes, have forced a mindset on citizens that compels them to “try and stay as calm as possible, I’ma be cooperative so I don’t end up in a bad situation, a.k.a. in a bad situation is them planning something, them putting false charges on you, yo. We shouldn’t have to think like that.”

Another man seemed to suggest police interactions as being sort of like a puzzle forced on black Americans, one that solving successfully — and with one’s life and health intact — involves “just giving them what they want and going on your way.”

The trouble with these “successful” encounters is that they do not do anything to lessen the terror that surrounds the police. They are difficult to shake, as there are constant reminders that just because one encounter that does not end in violence or unfair arrest or death is no guarantee that the next will not. There is no way to know who one is dealing with upon encountering an officer, and that uncertainty is at the core of the problem; it gives birth to an underlying level of mistrust, bitterness, and fear.

The story of another Baltimore Portals Policing Project participant illustrates this well. This black man, who spoke with a black man from Los Angeles in his session, described how his trust in police had been heavily eroded over time due to his own experiences. He spoke of how he’s “had evidence planted on me,” and “had police get on the stand and flat out lie.” One of those lies, he said, sent him to prison for a murder he did not commit — he was eventually cleared, but not before spending almost a decade incarcerated. But he also spoke of how he was not just failed by police, but betrayed by them in a time of need:

I see an incident involving a police officer. And, he pulled a guy over for a traffic stop. Man, listen … I witnessed him call backup, and they beat this man, they beat him. And I—I’m talkin’ to my sister who works for 911, she’s a 911 dispatcher, so she said, “Call the police.” I got off the phone with her, I called 911, and I said I’m at this address, and this is what I’m witnessing. Right? So the lady asked me all kinda questions, the 911 operator, what I’m wearing, this, that and the third. So she says, “Well assistance will be there in a moment.” The assistance came, two officers pulled up in a car—plainclothes officers—and they asked me, “Are you such, such, such?” I said yes. “Are you the one that called about the disturbance?” I said yes. They got out the car with their weapons drawn. “Get the fuck on the ground!” Boom, boom, boom, boom, boom! Locked me up and took me to jail.”

It might seem easy to dismiss this man’s story as uncommon or as being just one particularly troubling narrative. But it often feels that every black American has a similar tale, one that, if it did not happen to them personally, happened to someone they know well.

The Portals Policing Project has many similar stories. Atlanta Mayor Keisha Lance Bottoms recently spoke of how her husband was racially profiled and detained. I have many times seen police behave in the way the man in Baltimore described, both arriving at a tense scene and immediately escalating it with verbal abuse and slurs, and refusing to act in moments requiring their intervention.

And like Lance Bottoms’s children do, I know the wild panic that rises in you when your father does not come home — how it feels to spend hours desperately searching for him, your stomach sinking as you frantically worry he may no longer be alive, and the relief and anger that seize you when you discover he has been detained for a minor infraction he did not commit.

As is likely evident from stories like these, the memory of such experiences does not fade. The damage negative encounters have on trust for police is difficult to repair, particularly given that as much as one may hope their experience was a one-time event, there are constant reminders that similarly painful interactions could happen again.

The fear caused by police violence does not discriminate, and it has far-reaching effects

These stories — of hours or years lost to detention; of beatings and verbal abuse; of wondering if you are living your last moments — are perhaps not as horrifying as what the world saw with Floyd, his life leeching from his body over eight minutes and 46 seconds, or with Brooks, officers milling about as blood rushed from his motionless body. But they compound those horrors, they fortify them to form this omnipresent miasma of terror that spares no black American, no matter their wealth or level of education or power.

“Who do we call to protect us?” a man asked during one Portals Policing Protect session. “The people that’s here to protect us is pumping fear in our hearts.”

House Majority Whip Jim Clyburn, one of the most powerful black men in the country, had much the same complaint recently, saying on CNN, “I didn’t grow up in fear of police, even in a segregated environment. We never feared the police. But, all of a sudden now, I do fear the police.”

And that fear has real effects. It raises tensions with police; it leads to hesitation in calling them. But it also raises tension in the body and on the mind. Researchers have worked to understand the effects of stress and tension on the health of black Americans — which as federal officials have constantly reminded the public amid the coronavirus pandemic, is far worse than that of white Americans — and as a 2017 paper by researchers at Lehigh University, the University of Minnesota, and Brown University notes, research has found links between policing stress and health:

Experiencing or witnessing police brutality, hearing stories of friends who have experienced brutality, and having to worry about becoming a victim are all stressors. When faced with a threat, the body produces hormones and other signals that turn on the systems that are necessary for survival in the short term. These changes include accelerated heart rate and increased respiratory rate. But when the threat becomes reoccurring and persistent—as is the case with police brutality—the survival process becomes dangerous and causes rapid wear and tear on body organs and elevated allostatic load. Deterioration of organs and systems caused by increased allostatic load occurs more frequently in Black populations and can lead to conditions such as diabetes, stroke, ulcers, cognitive impairment, autoimmune disorders, accelerated aging, and death.

The authors in particular cite research from UCLA and Charles R. Drew University, which worked to measure the “allostatic load” — a term that refers to the buildup of physical issues as a result of ongoing stress — in black and white Americans. That research found loads to be greater in black Americans. The researchers’ analysis found genetics seemed to play little role, leading them to suggest societal factors are at fault.

Policing is only one of these factors, but it is fully connected to the others. For instance, housing discrimination has led to segregated neighborhoods with unequal policing standards. That discrimination has led to inequality in education, one characterized not only by disparities in achievement but in the use of school resource officers. The locations of those neighborhoods often lead to a greater reliance on public transportation, offering exposure to officers on the hunt for fare evasion and other minor infractions; in employment, those infractions haunt applicants looking for jobs. The examples could go on, but the point is policing is part of a complex system, and any abuses resonate outward.

Even the most cited research into police violence has serious limitations

This complexity — as well as the wide range of what constitutes abuse and the negative outcomes many black Americans fear — is difficult to fully grasp or even describe. In part, that’s because for many black Americans, this fear is simply a part of daily life: something terrifying that can’t be dwelled on too long lest it become all-consuming, but also something so common as to be undeserving of comment.

This is part of the value of the Portals Policing Project: It prompts people to comment on how policing feels and documents those comments for study and dissemination. But for all the conversations that project has recorded, there are thousands more that have not been shared beyond a small group of family and friends. For every video of police killing someone, there are hundreds of deaths that aren’t filmed. And there are thousands of negative interactions that aren’t recorded by bystanders or by police reports. An arrest generates paperwork; police hurling verbal abuse at people, or following them, or putting their hand on their weapon menacingly do not.

Gabriella Angotti-Jones/Los Angeles Times via Getty Images

And so — despite the terror and trauma — these incidents are not included in most studies, including ones commonly cited.

A paper released in January from two assistant politics professors at Princeton, Dean Knox and Jonathan Mummolo, as well as Hertie School of Governance research scientist Will Lowe, critiques many of these prominent studies. This study found fault with Roland Fryer’s methods, and it offered a sharp appraisal of other policing studies, stating many do not properly account for the “inevitable statistical bias that results from studying racial discrimination using records that are themselves the product of racial discrimination” and rely upon “implicit and highly implausible assumptions.”

The authors argue that events that generate police reports are affected by bias. Beyond this being a problem for those affected by these events, this creates an issue for researchers, one the paper sums up like this:

Suppose officers exhibited racial bias as follows: they detain white civilians if they observe them committing a serious crime (such as assault, potentially warranting the use of force) but detain nonwhite civilians regardless of observed behavior. When this is true, comparing stopped white and nonwhite civilians amounts to comparing fundamentally different groups. The analyst will observe force used against a greater proportion of stopped white civilians because of the differential physical threat they pose to officers. Under the traditional approach, the analyst would naïvely conclude that anti-white bias exists, yielding an erroneous portrait of racial discrimination in the use of force.

Knox, Mummolo, and Lowe use papers by Fryer to help illustrate this point. For instance, Fryer published a study in 2019 that found black and Latinx Americans are 50 percent more likely to experience force at the hands of police than white people. But the three researchers find fault with these results, which used police department data from New York City, Los Angeles County, three Texas cities, six counties in Florida, and the Police-Public Contact Survey (a publicly searchable dataset on police interactions), arguing this number is an undercount.

The Fryer work, for instance, included handcuffing as use of force, but Knox, Mummolo, and Lowe argued:

While a naïve analysis that assumes no race-based selection into the data suggests only 10,000 black and Hispanic civilians were handcuffed because of racial bias in New York City between 2003 and 2013, we estimate that the true number is approximately 56,000. And while analyses ignoring bias in stopping would conclude that 10% of uses of force against black and Hispanic civilians in these data were discriminatory, after bias-correction, we estimate that the true percentage is 39%.

The researchers acknowledge that the results may be an overcount based on what it describes as a reliance “on weaker assumptions that in some cases are difficult to verify” —for example, the belief that violence is more likely to occur in police encounters involving “severe” crimes rather than minor offenses. But they suggest the assumptions Fryer used, such as the variables that affect how officers view compliance with commands, are “implausible.” (Fryer notes that some assumptions were less than ideal, but unavoidable because of a refusal on the part of departments to provide him with data.)

This goes beyond an academic debate. For one, this research attempts to describe a phenomenon that has an incredible — sometimes existential — affect on people’s lives and well-being. But for another, it is being used to shape debate and public perception.

Data on police violence hides as much as it reveals

I and others have cited research — such as the study finding black men have a one in 1,000 chance of being killed by police — to explain what it can feel like to be a black man in America and to suggest the threat of police killings is real and concrete. Perhaps that has affected readers or policymakers.

But some, such as Heather MacDonald in a recent op-ed in the Wall Street Journal, have used Fryer’s research to make the opposite point. In a piece titled “The Myth of Systemic Police Racism,” MacDonald writes, “Research by Harvard economist Roland G. Fryer Jr. also found no evidence of racial discrimination in shootings. Any evidence to the contrary fails to take into account crime rates and civilian behavior before and during interactions with police.”

Putting aside the issue of this first sentence excluding the full span of police violence, the statement about Fryer’s research is true. But it does not get into the assumptions he — and all scientists, engineers, and researchers — are forced to make to do their work.

Like other scientists, he lists those assumptions, as well as limits, in the paper MacDonald cited, writing, for instance, that the fact officers “may be more likely to charge black suspects with crimes such as resisting arrest or attempted assault on a public safety officer rather than misdemeanors, relative to whites, for identical behavior” is “an important limitation” of the dataset he has assembled. He writes, “It is unclear how to estimate the extent of such bias or how to address it statistically.” The Knox, Mummolo, and Lowe work is, in part, an effort to resolve this issue.

This limitation does not undermine Fryer’s work or make him a bad researcher, but it is a reminder that the data can’t capture everything, particularly datasets that, as Fryer himself notes, are reliant on police cooperation.

And this is something important to keep in mind when thinking about the second statement MacDonald makes in that excerpt, about crime rates and behavior. The behavior piece of this comes from earlier in MacDonald’s article, in a section on the “routine violence” one sees in black communities, which contains a fallacious argument that “black-on-black” crime is worse than police violence against black Americans. That is separate from this discussion, but Victoria M. Massie and Khalil Muhammad have explained very well for Vox what the argument misses and how it is racist.

MacDonald’s reference to crime rates, however, further illustrates the limitations of data. She writes, “In 2018, the latest year for which such data have been published, African-Americans made up 53% of known homicide offenders in the U.S. and commit about 60% of robberies, though they are 13% of the population.”

This is taken from Department of Justice data, and it is what the data shows. But this data does not take into account inequities in policing. If, as Fryer suggests, a black man is three times more likely to be stopped by police, then it would seem that police are stopping a disproportionate number of black people and are thus more likely to find a disproportionate number of black criminals.

Other federal data from 2018 that is important to keep in mind is the FBI’s database on clearance rates — that is, solved crimes. In 2018, the FBI states 62.3 percent of “murder and nonnegligent manslaughter” (the category MacDonald categorized as homicide in her piece) were cleared. How many of those 37.7 percent of unsolved murder and manslaughter cases were committed by white offenders? And how would having that data affect the black crime statistics MacDonald cited? It is impossible to know because that data does not exist.

This is not to attack MacDonald, but it is to say data can occlude as much as it can reveal; the same is true of data cited earlier in this piece. It is valuable, particularly to those without firsthand experience with things like police violence, but it has serious limitations and cannot be the sole source of policymaking or an individual’s understanding.

The data might show police killings are declining or that there is no racial bias in policing. But I have seen that bias, as many Americans of all ethnicities now have. And when one’s own observations, and the realities of communities that bear the brunt of that policing, say something that runs counter to data, that has to be given some weight.

It’s important that Americans seek to better understand the issue of policing, but in working to fix it, they must be careful not to oversimplify it, or to discount the familiarity that comes with experience. Because police violence is complex and multifaceted — and its effects are real and terrible and traumatic.

Support Vox’s explanatory journalism

Every day at Vox, we aim to answer your most important questions and provide you, and our audience around the world, with information that has the power to save lives. Our mission has never been more vital than it is in this moment: to empower you through understanding. Vox’s work is reaching more people than ever, but our distinctive brand of explanatory journalism takes resources — particularly during a pandemic and an economic downturn. Your financial contribution will not constitute a donation, but it will enable our staff to continue to offer free articles, videos, and podcasts at the quality and volume that this moment requires. Please consider making a contribution to Vox today.