True, but if blue people commit a higher number of crimes, then it's reasonable there will be a great number of witness reports describing blue suspects, and th

True, but if blue people commit a higher number of crimes, then it's reasonable there will be a great number of witness reports describing blue suspects, and therefore you can expect a higher number of encounters between police and blue people. You can examine the link between suspect descriptions and the number of encounters and see if an ethnic group generates more or less encounters per description to check for racism in who police decide to encounter. You can then look at the rate at which police encounters become violent, and if that rate is different than the rate for green people, that would also be clear evidence of racism on the police force.