Vox has a post about racial bias and police shootings. The story by Vox writer Jenée Desmond-Harris included quotes from Joshua Correll, who investigated racial bias in police shootings with a shooter game, in his co-authored 2007 study, "Across the Thin Blue Line: Police Officers and Racial Bias in the Decision to Shoot" (gated, ungated).
Desmond-Harris emphasized the Correll et al. 2007 finding about decision time:
When Correll performed his experiment specifically on law enforcement officers, he found that expert training significantly reduced their fatal mistakes overall, but no matter what training they had, most participants were quicker to shoot at a black target.
For readers who only skim the Vox story, this next sentence appears in larger blue font:
No matter what training they had, most participants were quicker to shoot at a black target.
That finding, about the speed of the response, is fairly characterized as racial bias. But maybe you're wondering whether the law enforcement officers in the study were more likely to incorrectly shoot the black targets than the white targets. That's sort of important, right? Well, Desmond-Harris does not tell you that. But you can open the link to the Correll et al. 2007 study and turn to page 1020, where you will find this passage:
For officers (and, temporarily, for trained undergraduates), however, the stereotypic interference ended with reaction times. The bias evident in their latencies did not translate to the decisions they ultimately made.
I wonder why the Vox writer did not mention that research finding.
I doubt that the aggregate level of racial bias in the decision of police officers to shoot is exactly zero, and it is certainly possible that other research has found or will find such a nonzero bias. Let me know if you are aware of any such studies.