I think it would be useful, maybe explosive, if someone did an experiment to see if the image of a black person causes pain to a racist. There have been experiments performed which show that images of a romantic partner or the Virgin Mary cause a reduction in experimental pain in test subjects. It might not be possible to perform a similar experiment with images of black people without the subject guessing how they were expected to respond, in other words, to do it double blind.
But there is a body of data which supports the idea that police are more trigger happy in the presence of black people than white people. The Implicit Association Test measures the degree to which a person is racist against a variety of groups by measuring the time it takes for the subject to evaluate an image or word. The subjects are asked to press one key if the image has to do with white or good, and a different key if the image is black or bad. They do this twice to compare their response with their dominant hand against their non-dominant hand. Then they do the tests again with evaluating the image if black or good vs. white or bad. The subject is responding by pushing the E or I keys on the keyboard with their index fingers. This is a lot like the policeman deciding to record his evaluation of a situation by twitching his index finger on the trigger of his firearm. The people who run the IAT out of Harvard have found that unconscious bias against black people is present in most people, even black people.
I think that data from the IAT support the thesis that most police are more likely to evaluate black people as bad, or shootable, than white people, due to their unconscious bias. Ridding ourselves of such bias is extremely difficult. Can police help it, that they are so biased? This is why racism is built into the system. Reparations are just one possible way of compensating for this bias.