What the Data Do and Don't Say about Policing and Race

Harvard's Roland Fryer argues both Left and Right misunderstand or misrepresent the empirical evidence


Harvard economics professor Roland Fryer had an interesting Wall Street Journal op-ed summarizing his research on race and policing. It begins:

I have led two starkly different lives—that of a Southern black boy who grew up without a mother and knows what it's like to swallow the bitter pill of police brutality, and that of an economics nerd who believes in the power of data to inform effective policy.

In 2015, after watching Walter Scott get gunned down, on video, by a North Charleston, S.C., police officer, I set out on a mission to quantify racial differences in police use of force. To my dismay, this work has been widely misrepresented and misused by people on both sides of the ideological aisle. It has been wrongly cited as evidence that there is no racism in policing, that football players have no right to kneel during the national anthem, and that the police should shoot black people more often.

As for what his research shows, Fryer claims the following:

  • "There are large racial differences in police use of nonlethal force."
  • "Compliance by civilians doesn't eliminate racial differences in police use of force."
  • "We didn't find racial differences in officer-involved shootings."

Some conservatives like to point to this last finding to rebut claims of racial disparities in policing without noting Fryer's other findings from the same research suggest just the opposite, nor do they note the limitations of Fryer's research (which he himself is quick to acknowledge).

The above is largely based on this study. Fryer also notes research by Phillip Atiba Goff et al. and Ted R. Miller et al. reaching larger similar conclusions. Later in the op-ed he also discusses the paper Eugene posted about here.

This is obviously not the last word on this important subject, but it's an interesting contribution to our understanding of racial disparities in policing.

NEXT: "An Elite Progressive LISTSERV Melts Down Over a Bogus Racism Charge"

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. The reason the right focuses on the last point is because the discussion was started by the left with the framing of “police are killing all black people,” or some other hyperbolic claim… but those claims were always limited to police killings. It was only after the bombastic claims were hurled to get attention did anyone (and then, only some) ever attempt to clarify that the complaint was first and foremost about policing in general. And yet, for many who do try to make that pivot they fail to abandon the falsehood of genocide by cop. And as long as that claim is hoisted, it deserves the most attention as it causes more damage to everyone. It causes damage by drowning out otherwise legitimate claims and by making the messenger of otherwise legitimate claims less trustworthy. And at some point, when taken with the context of some of the other things said about society as a whole and whites in particular, it starts to make the messenger not just less trustworthy but also seem to be a malefactor. You don’t get to say that the side who disputes a falsehood is wrong for being cautious to believe more falsehoods, but say that is what always happens. Perhaps we should demand truth from those who make claims rather than blind trust from cynics.

    1. Near the end should read “…but sadly (not say) that is what always happens.”

    2. Do you think ‘the left’ hasn’t complained about police brutality in general towards black suspects for quite a while? It’s more likely that it takes a filmed killing to bolster a movement.

      1. I am not saying there aren’t killings or problems that need to be addressed. I’m saying that by lying about the nature of the problem (genocide by cop) means that you undercut your credibility. As such, others are right to be cynical. That does not mean the real problems do not deserve to be addressed. It means the messenger is a liar. The messenger has raised the bar for themselves before others will listen. And a killing, as sad as it is, may be what it takes to shock people… but it can not be theater if it is to be believed. When the left says that cops are killing blacks… for being black… and the people look into that and find it to be a myth… then people are going to demand much more proof of other less dramatic claims too. That doesn’t make those other claims false… the boy who cried wolf didn’t lie either at the end of the story. But it sure as hell was his own damn fault no one took him serious.

    3. I probably don’t agree with your ideological priors, but your analysis is more accurate than I would probably like to admit.

      Invoking death is the great talking point. So it is thrown around freely. Remember the incident with the birder and the dog walker? (So so long ago….) Instead of just saying “she shouldn’t have threatened to call the cops”- an absolutely legitimate complaint, people said “she’s trying to get him killed”, as if every or almost every time a white person calls the cops on a black person, the cops come and shoot the black person.

      Or the New York Times reporters complaining about Tom Cotton. They didn’t just say they didn’t like the piece, didn’t think it was well argued, didn’t belong on their pages. They said it would get black reporters killed. Again, there’s zero evidence of that. They just said it. Death is the great talking point.

      We have a huge problem with racism in policing; and we all know that most of it does NOT show up as officer-involved shootings. Rather, most of it is the day-to-day crap blacks have to put up with in being pulled over all the time, stopped on the street, frisked, asked to get out of cars, and otherwise being subjected to stuff that white people aren’t subjected to nearly as often. We also get an occasional George Floyd, which of course is absolutely horrifying. But “they’re gonna kill us all” is the talking point, rather than “every black male has to fear being pulled over or harassed by the police based on nothing”.

    4. And as long as that claim is hoisted, it deserves the most attention as it causes more damage to everyone. It causes damage by drowning out otherwise legitimate claims and by making the messenger of otherwise legitimate claims less trustworthy.

      If Mr Floyd had survived, roughly no one would have paid any attention to his treatment. It was precisely because he did die that – apparently – polls now show 65% support for Black Lives Matter.

      These “otherwise legitimate claims” would languish in obscurity if there were no spectacular and televisable, even if statistically more dubious, claims to which they could be attached for better emotional effect.

      The public, as Prof Somin would advise, is composed largely of low information voters. They (which includes me) are not paying a great deal of attention. To grab attention, you need some fireworks.

      There are loads of very well meaning young people out there who sincerely believe that Israel is the wickedest state in the world – solely because unhappy events there are filmable and make the evening news. North Korean concentration camps and Chinese organ farms are not filmed and do not make the news.

      The lesson of “You can fool all the people some of the time and some of the people all the time, but you cannot fool all the people all the time” is seriously misunderstood. The important point from a political perspective is that most people are fairly gullible most of the time. Which is good enough for government work.

      1. The problem with seeing this as a positive (dubious claims gaining attention thereby granting coattails to more legitimate claims) is that as people go on and find it increasingly difficult to marry what they were told was true to what is true… you breed resentment. That creates disunity over the long run. That is my point. The “cure” of lying may very well be worse than the illness in the long run as it will begin to impact everyone in a myriad of ways. Look how this is spiraling out of control already. The momentum is being perverted into censorship, Soviet style erasure of history, and demands of admissions of guilt by innocent people. Today those problems seem trivial to the life of Mr. Floyd. But what of the lives of the cops killed in response? Those killings were justified by a lie. What of people’s livelihoods destroyed by fire and theft justified by a lie? Eventually the pervasiveness gets to a point that you can’t say the lie is worth preserved… can you?

        1. I’m not applauding mendacity, false advertising and gaslighting. I’m merely pointing out that it usually works.

          The Soviet Union lasted 74 years, which is pretty good. But even now, after all the lies that have been exposed, people have moved on. I betcha if you strolled onto campus anywhere in the US and asked “any idea who Beria was ?” you’d get 99% don’t knows and even from the 1%, you’d be no more likely to be offered the right answer than the speculation that he pitched for the Yankees in the 80s.

          Even with constant reminders of how unpleasant the Third Reich was, I doubt you’d do better than 1% even on Goebbels.

  2. “There are large racial differences in police use of nonlethal force.”
    “Compliance by civilians doesn’t eliminate racial differences in police use of force.”

    This seems highly unlikely given the violent crime disparities between races compared to the use of nonlethal force numbers. Police would have to be 4-8x as likely to use nonviolent force against a Black American committing the same felony for it to be true.

    1. Do you have specific criticisms of the research?

      Have you looked at the data presented?

      Or are you just saying your “gut feel” is that it must be wrong?

      1. Yes, my specific criticism is that if a person pulled over for a DUI is 4x as likely as another person to commit a battery his compliance rates are obviously not correct.

  3. What about the converse — racial breakdown of those who kill cops?

    1. With over 800K LEOs in the US, the number of officers killed by felonious means is under 100/year, closer to 50/year for the whole country. With such low numbers a racial (or any other) breakdown of the killers is likely to be statistically meaningless.

      The vast majority of officer line of duty deaths are the result of traffic accidents.

  4. Maybe we should begin to consider the possibility that the reason minorities get in trouble with the police more is because maybe they cause more trouble or are around more trouble. You know…like the simplest possible answer?

    I would take the systematic bias side more seriously if they could come up with something any trend or significant data was naturally present and wasn’t the result of a fishing expedition proceeding from the foregone conclusion that a vast conspiracy must exist. But nope after decades and millions/billions of dollars thats pretty much still the crux of their arguments.

    Its kinda weird how the white supremacist conspiracy screwed up and targeted women and asians they hate so much even less than they should have even if everything was proportional.

    1. “Adding controls that account for important context and civilian behavior reduces, but cannot fully explain, these disparities.”

      1. These are the same people who think they can capture all the dynamics and psychology of how humans negotiate compensation and prove the pay gap exists by controlling for a couple of variables on a spreadsheet.

    2. There are multiple studies out there that show that for searches of cars during a traffic stop, police are more likely to search a car driven by a minority, but they are actually more likely to find contraband in a car driven by a white person.

      That would be the opposite of minorities getting in trouble more with the police because they cause more trouble.

      1. Yeah. And this has been repeatedly studied.

        Which means, conservatives insistence that all the police racism out there must be about crime control is… a prejudice. A racial prejudice.

      2. Alright lets set it…

  5. Really a shame the Fryer study you linked to is behind a paywall, Professor Adler.

  6. Heckuva libertarian commentariat that’s ride-or-die for police being great so long as we’re talking about blacks being killed under color of law.

    Lets remove race from the issue:
    Violent crime has gone down. Police spending and tactical intensity have not. We don’t live in the 1980s anymore; maybe our police shouldn’t either.

    1. The police need the gear to arrest is normies for going to the beach, the playground, attempting to open our stores, or to stand there looking menacingly while our historical figures get torn down or white conservatives get assaulted by Marxist Pantifa or White Lives Don’t Matter radicals.

      1. From what I can tell from your not-very-coherent post you want the police to be armed and trained primarily for civil unrest.

        Which is a pretty incorrect understanding of the primary purpose of police.

        1. No.

          What you should take from my post is that unlike your lies about police wontonly killing blacks, what they are really doing is arresting normal people for going to the beach, for letting their kids play on the playground, for trying to open their businesses, while simultaneously standing down in all their mean looking military style assault gear when our history is getting torn down by Marxists or when Black Supremacists are assaulting regular normal people.

          1. So you don’t really want to talk about what I brought up, but would rather yell about how the real problem with government force is that’s it’s used against the wrong people.

            How very textbook authoritarian of you.

            1. I’m an authoritarian because I rejected your propaganda?

              Let me guess, my speech is violence, and also my silence is violence therefore you’re justified in assaulting me?

    2. Foxastro Butterfieldo : Violent crime has gone down. Police spending and tactical intensity have not.

      1. I get the reference, but since I’m only arguing correlation I don’t understand how it’s germane.

        1. The spending etc. was largely responsible for lowering the crime and now keeps crime down.

          1. That’s hardly proven. A reflexively pro-police stance to take.

            Police having MRAPS keeping violent crime down?

          2. Also: Police spending has largely remained steady. So even if your causal scenario is correct, why has crime continued to fall even in areas where police budgets have remained steady?

            1. Why does the car continue to slow down as you keep your foot on the break?

          3. No, the spending wasn’t responsible for the crime going down, or at the very least, no one has been able to prove it. There is significant evidence broken windows policing helps (which does not require huge investments) but there is also evidence that increased incarceration, less lead poisoning, increased abortions, air-conditioning, video games, and ubiquitous free porn all have a share in the decline as causes.

            NYC used broken windows policing to lead to major decline in crimes, but slower and steady declines elsewhere during the same time period shows that NYC’s success was concurrent with a generalized decline. My favorite theory is that NYC under Guiliani just locked up the bad guys for minor crimes before they got around to the big ones.

            1. I suspect lead has something to do with the lower crime rates.

              1. It’s all correlation…as lead in environment increased due to increased use of leaded gas and paint, crime went up, with a corresponding decrease over time when lead use was reduced. Excessive lead exposure leads to lower intelligence and reduced self control, a feature of common criminals.

                However, a fair counterpoint, is that lead was everywhere in Japan as well for the same reasons, and we have very good data on both Japan and the US, but there was no increase in crime in Japan. My unprovable theory, is that there was more cases at the margin in the US.

              2. I totally buy the lead hypothesis as well, Dilan. The correlations are just too close and many to not be causal.

                But I ain’t a statistician.

                1. You definitely don’t want tolook at this 🙂

                  (FWIW, I don’t believe or disbelieve the lead hypothesis – the correlations are intriguing, but I’m just not into faith based reasoning. Ditto for the its ‘legalized abortion’ counterpart)

            2. Let us not forget age either. Age and testosterone (which begins to decline from age 20) are magnificent statistical correlates of crime.

              Though the sex structure of the US population doesn’t change, the proportion of the population composed of young men has certainly declined over time.

    3. FWIW, Sarcastro, the commentariat here is divided about the police. Plenty of conservatives and libertarians here, for instance, want to get rid of qualified immunity or favor a strong Fourth Amendment, both of which would be very beneficial to blacks who are abused by the police.

      1. Agreed. And some are consistent. But I’ve noticed even some of the anti-QI folks have suddenly discovered a strange new respect for the po-po.

        Tribalism is a helluva drug.

  7. To my dismay, this work has been widely misrepresented and misused by people on both sides of the ideological aisle. It has been wrongly cited as evidence that there is no racism in policing, that football players have no right to kneel during the national anthem, and that the police should shoot black people more often.

    Hmm – I’m struggling to identify the ideological aisle he has in mind.

  8. Why was a link to the Roland Fryer study supplied that required a payment of $20 when it is available other places for free? For example, here: http://www.nber.org/papers/w22399.pdf

  9. So the question is this: what are the reasons for the disparities in police use of nonlethal force? What percentage of it is (a) mindless racism unconnected to reality in any way, and what percentage is due to (b) the previous experience of the police suggesting a greater likelihood of violent resistance in black communities? This is not to suggest that police have a right to treat people differently based on their race, and in fact rough treatment under (b) is, at least in part, a self-fulfilling prophesy.

    But what I am wondering is this: how much of the disparity has to do with pure racial animus, and how much has to do with police procedures in high crime areas?

    Here’s another question: suppose it is true that a police officer, after some time on the job, concludes that a disproportionate amount of violent resistance he has experienced has come from African Americans and so he unconsciously starts anticipating that by adopting a harsher demeanor with them from the start, which, we will assume, is the correct approach to take only with someone who actually resists or is violent. Is this approach to be judged less harshly than the same approach undertaken not as a reaction to prior experience but purely out of racial animus?

    1. Slate had a good piece today interviewing police officers about Floyd. And one of them made the point that it’s important to rotate everyone around through the various parts of the city, so they don’t spend, say 20 years policing one high crime minority neighborhood and develop a bunch of racial stereotypes.

      1. Ahh yes, the war against noticing.

        1. What is the war against noticing?

          1. I assume Allutz is referring to the belief that racial stereotyping (aka racial profiling), which is a subset of what Fryer calls information-based discrimination (aka statistical discrimination), is wrong.

    2. The Roland Fryer paper identified 125 factors related to “important context and civilian behavior” that are thought to be influential in interactions with the police that involve use of force, and tried to control for them. The raw data showed blacks and Hispanics being 50% more likely to have an interaction with the police that involves any use of force. After controlling for these factors this was reduced to 18% more likely. Curious that when it comes to use of lethal force there was no racial difference in either the raw data or when contextual factors are taken into account.

    3. The paper has something to say about this though its conclusions are preliminary. Search these comments for “paywall” to find a free version of the paper. The relevant distinction is between informational (or statistical) discrimination vs taste-based (by which the author means racial animus-based) discrimination.

Please to post comments