In the past couple of weeks I’ve written about the views of the African-American English and Comparative Literature Professor John McWhorter (e.g., here and here) on whether the higher per capita rate of deaths of black people at the hands of police implicates racism. His view was “not necessarily,” and he posed an alternative hypothesis. His hypothesis is based simply on the fact that arrests for violent crime in America is higher for blacks than for whites; the article below, from the Boston Globe, says 3.6 times as often.
Now if presuppose that there is no racism as the null hypothesis, and then posit that there is a probability x, identical for blacks and whites, that an encounter with the police will lead to police murder of the suspect. If this is the case, and blacks encounter police more often, then one can still get a higher per capita death rate of African-Americans than of whites even with identical “x”s. So the higher per capita rate does not in itself implicate racism (n.b., McWhorter does highlight that there are data, like those from traffic stops, that do indicate racism on the part of police).
As I pointed out at the time, nobody seems to have adduced the right data needed to settle the situation:
Of course these are just parallels and don’t answer the question we want to know: are black people more likely to be murdered by cops on a per capita basis, in a given set of encounters, with controls from white suspects in similar situations, with all other things roughly equal? If that’s the case, then racism is implicated. Anecdotes like those above won’t answer that question.
The same point is made by Aubrey Clayton, a mathematician and author, in this Globe article (click on the screenshot below). I think he makes the point unnecessarily complicated because he uses math and many American are innumerate, but the point is pretty much the same as mine:
Here’s Clayton’s analysis, which is further complicated because he doesn’t use equal number of incidents for whites and blacks, so you have to multiply up:
Could the reaction to high-profile killings like those of Breonna Taylor and George Floyd be a matter of confirmation bias? Could the narrative of police racism be disproved with a tweet-sized calculation?
These statistics are consistent with excessive use of deadly force against Black people, due to a mathematical phenomenon called Simpson’s Paradox.
The key point is that not all encounters with police are equally deadly. In any given kind of encounter with the police, a Black person can be likelier to be killed than a white person even if the overall rate of deaths per encounter appears lower for Black people. This would happen because Black people have many more interactions with police in non-deadly situations — a dynamic exacerbated by racism. And all those extra encounters dilute the rate.
Consider two extremes of police encounters: traffic stops and active shooter scenarios. Suppose, hypothetically, that a white suspect is killed by police in one out of 100,000 traffic stops and nine out of 10 shootings. And imagine that Black suspects are killed by police after 20 out of 1,000,000 traffic stops and in 10 out of 10 active shooter incidents. In each kind of incident, Black suspects are killed more often than white suspects. In aggregate, though, the percentage is higher for white people: 10 out of 100,010 white people are killed vs. 30 out of 1,000,010 Black people, because the white people tend to encounter the police in more grave situations.
He could have given the per capita death rate here, which is 0.01% for whites and 0.003% for blacks, despite the higher per capita death rate of black suspects in both traffic stops and shootings. This depends on the particular figures chosen, of course, but the upshot is that if encounters with the police have different death rates for different races, and those encounters have substantially different frequencies among races, then one is not justified in imputing different death rates as a statistical result of number of police encounters. In other words, higher crime rates of blacks than whites doesn’t take racism off the table as a cause of higher per capita death rates of black suspects. In fairness to McWhorter, I don’t think he makes that argument, but broaches it as a possibility that needs to be considered, and I agree.
(Note, though, that the 3.6-fold difference cited by Clayton is “arrests for violent crimes”, not traffic stops.)
Getting the right data, however, involves more than just calculating homicides in traffic stops versus active shootings. If someone stopped in traffic has a gun, or refuses to obey police orders to show their hands, those factors may differ among groups and also lead to differential death rates in very similar kinds of arrests. Getting the right “controls” here may, I’m starting to realize, may be impossible, as each encounter has its unique aspects. Clayton realizes this:
There are, of course, more than two types of police encounters in reality, and whether any of them involves deadly force will depend on many factors, such as whether the suspect is armed and making threats, how many officers are on the scene, and so on. The actual data is far more complex than in this simplified example, and there isn’t consensus over whether clear evidence of encounter-specific racial bias exists. There are just too many variables for the data to be definitive on its own.
That’s why one study, frequently cited as evidence that Black people are killed just as often (or less often) as others in similar situations, has been critiqued by other researchers who noted that “its approach is mathematically incapable of supporting its central claims.”
So give credit to Clayton as well for saying that we just don’t know the role racism plays overall in the disparity between death rates between whites and blacks in police encounters. And perhaps we can’t settle that with data (remember that the race of the police officers must be considered as well, assuming that blacks can’t be racist against blacks nor whites against whites).
And here’s Clayton’s conclusion, which seems reasonable. It’s not really a conclusion but a caveat:
The inflated number of non-lethal encounters Black people experience due to racial profiling could be what shifts the balance, perversely using one kind of discrimination, over-policing, to mask another: the greater use of deadly force against Black suspects. Simpson’s Paradox predicts these counterintuitive results whenever data is averaged over inconsistent group sizes. Here, the inconsistency lies in the types of interactions Black and white people have with police. Since these are distributed differently, the pooled numbers can get the story backwards.
. . . as they do in Clayton’s example above. The lesson for us: while in certain cases racism is clearly involved in the murder of black suspects (police remarks on the scene, differential deaths due to racial profiling in traffic stops, which is a fact, and so on), we cannot say, as so many are doing now, that the disproportionate number of deaths of blacks at the hands of police is prima facie evidence of structural racism in the police. We need controlled data to say that. We may not get that data, but at least we can do things to try to eliminate anything that smacks of racism in police departments, like getting information on traffic stops and other stops for nonviolent crimes like drugs.