First time reading this from the Boston Globe: “The statistical paradox of police killings: In the numbers of fatal encounters with the cops, one kind of discrimination masks another.” Aubrey Clayton may be a good mathematician, but his appeal to Simpson’s paradox only matters if the facts are accurate—and they’re not. Clayton is off the mark in his claim concerning racial disparities in civilian-officer encounters. By a lot. Whites are far more likely to encounter police in both police-initiated (including street stops and arrests) and public-initiated contacts (whites are significantly more likely to call police to report a possible crime, for example) than are blacks.
What Clayton doesn’t talk about is that police shootings are determined not by racial bias (where is the evidence of this?) but by individual and situation factors, and these factors are more likely to be present on a proportional basis when black suspects are involved. Even then, the police kill twice as many whites as blacks on an annual basis. More than that if you count Hispanics as white, the race two-thirds of Hispanics identify as (Hispanic isn’t a racial category but an ethnic one).
Moreover, even if the facts supported Clayton’s claim, it would not follow that racial bias explained the disparity. Any claim that racial disparity is the result of racial bias must come with clear evidence of racial bias. Disparity is not prima facia evidence of bias. Ever. It’s like this claim that we need to do away with standardized testing because blacks don’t do as well as whites as a group and therefore the tests are racially bias. How about this: blacks don’t do as well as whites as a group because they are not as well prepared for the test as whites? That’s what the evidence shows. Is the lack of preparation racist? The evidence suggests that these effects are produced by class and culture, not race. So probably not. (See John McWhorter’s New York Times article “Lower Black and Latino Pass Races Don’t Make a Test Racist.“)
For the record, civilian-police encounters run in the tens of millions annually. The police kill approximately a thousand civilians a year (again, most of whom are white). The proportion of those civilians who are unarmed is around a dozen–and being unarmed does not make a person not dangerous. The fact is that the probability that a violent offender will be killed by a cop is vanishingly small. Given that the United States is the most criminally violet advanced democracy in the world, Cops show remarkable constraint in interactions with civilians.