Unfortunately the metric it uses is whether they were ever incarcerated for any reason. Someone hostile to those results would probably dismiss them as due to police racism.

One could try to use homicide and victim surveys as a mostly objective proxy for all violent crime (very difficult and impossible, respectively, for racist police practices to affect), compare that to the violent crime rate of each racial group, and use any disparity to estimate how racist the police and justice system are being.

But then one would risk getting a result one doesn't like.