Today saw the release of a very good — but also very limited — study suggesting that white officers are far more likely to use “gun force” against blacks than black officers are.
The study works by leveraging a natural experiment in the way cops in some places get assigned to 911 calls: The nearest available cop is assigned to each call and doesn’t get to choose whether to go. If these calls are more likely to result in force when a white officer gets assigned to a black neighborhood (relative to a black cop working the same beat and shift), it seems reasonable to conclude race is making a difference.
That’s exactly what the study finds, in contrast to several previous studies using other, less precise methods — though those other studies tended to focus on fatalities specifically, and in one case also found racial disparities in non-lethal force. Most shockingly, “while white and black officers use gun force at similar rates in white and racially mixed neighborhoods, white officers are five times as likely to use gun force in predominantly [80 percent or more] black neighborhoods.”
But now let’s talk limits. The study includes only two cities, and the black-vs.-white results for “gun force” come from only one of them. (The analysis of the second city focuses on white-vs.-Hispanic comparisons, and the data don’t even say what type of force was used.) There is a lot of variation across police departments in terms of how much they use lethal force and the degree to which cops stop blacks more than whites, so you can’t just assume a result from one city generalizes to others.
To make matters worse, to get the data, the researchers had to promise not to name the cities — though they are able to tell us a few things about them, such as that the city the “gun force” results come from “has large populations of both blacks and whites, a total population of over 240,000 and has a homicide rate that ranks in the top 20 among the nation’s 100 largest cities.” From this it seems at least possible that the city is not representative of the country in ways that may strongly affect the results. But we really don’t know.
And by the way, what is “gun force” exactly? The study reports that the “type of force is recorded by the police officer, including whether the force was a gun. According to the police department records division, nearly all of the incidents of gun force are the discharging of the gun.” This is not reassuring. Presumably this can’t include every time an officer draws a gun but doesn’t fire, which should be a non-trivial proportion of all gun uses. But then what situations make it “nearly all,” so that we can’t just interpret the statistic as referring to discharges? How confident are we that these data are coded consistently?
A couple other potential issues the authors themselves note: They have no way of knowing whether the force in any given case was excessive, or whether an officer instead used less force than he was trained to, so differences in force levels don’t necessarily prove that the white cops used too much force. The authors also can’t account for the suspects’ behavior, so if black residents respond differently to black and white cops, that could be a factor too. And the data can’t distinguish outright bias from, say, the effect of white cops’ being less able to de-escalate volatile situations in black neighborhoods owing to cultural differences.
Nonetheless, this study makes a good case that there’s a serious problem in the two cities it focuses on, and even if the problem isn’t ubiquitous, there are no doubt many more places like them. I’ve been following the race-and-police-shootings debate for a long time, and this is the very best evidence I’ve seen for bias in lethal force. Hopefully those cities will take it to heart, even if they won’t identify themselves publicly, and hopefully other cities check their data for similar patterns.