The best look at this comes from Brad W. Smith, a researcher from Wayne State University in Detroit. In a 2003 paper, he looks at the impact of police diversity on officer-involved homicides in cities of more than 100,000 residents and cities of more than 250,000 residents. Regardless of city size, there wasn’t a relationship between racial representation and police killings—officer diversity didn’t mean much. At most, in smaller cities, female officers were more likely to commit shootings than their male counterparts, a fact—he speculates—that could be tied to sexist pressures on female officers, who might feel the need to act “tough” to prove their bona fides.

What mattered for police shootings wasn’t the makeup of the police department, it was the makeup of the city. In all measured cities, an increase in Black residents brought an increase in police shootings. In smaller cities, a substantial change in the proportion of Black residents resulted in a slight increase in the predicted number of police-caused homicides. And in the larger cities, the same change increased the chance for police-caused homicides by a factor of 10 compared to smaller cities. Put another way, the quickest way to predict the number of police shootings in a city is to see how many Blacks live there