Quantifying Black Lives Matter, by The Economist.

As a teenager, Roland Fryer had “unpleasant” run-ins with police. Officers pointed guns at him six or seven times. Even now, the youngest African-American to get tenure at Harvard wonders why police shout loudly at him as soon as he forgets to indicate when driving. …

Then he collected a lot of data. …

What shocked Mr Fryer was when he looked in detail at reports of police shootings…. To his surprise, he found that blacks were no more likely to be shot before attacking an officer than non-blacks. This was apparent both in the raw data, and once the characteristics of the suspect and the context of the encounter were accounted for.

Mr Fryer dug deeper into the data. He combed through 6,000 incident reports from Houston, including all the shootings, incidents involving Tasers and a sample in which lethal force could have justifiably been used but was not. What he found was even more startling: black suspects appear less likely to be shot than non-black ones, fatally or otherwise. …

Mr Fryer adds that blacks who were reported by cops as being perfectly compliant with police instructions during their interactions were still 21.1% more likely than whites to have some force used against them. This points to racial prejudice. …

In explaining why racial bias is present in all cases except shootings Mr Fryer suggests that it may reflect how officers are rarely punished for relatively minor acts of discrimination. When he shadowed cops on patrol, Mr Fryer was told repeatedly that “firing a weapon is a life-changing event”—and not only for the victim. Although activists argue that too many officers get off lightly when they harm civilians, cops find it hard to escape any scrutiny after discharging their weapon. More transparency and accountability are therefore needed, even when police encounter members of the public.

hat-tip Matthew