[mdjeu]

Yes, better data is needed. I feel weird being a skeptic of this study, because I personally believe law enforcement in the US is out of control, but...

In this area, there's so many important factors relating to the nature of the events, that "armed" versus "unarmed" and race are insufficient to get a good estimate of what's going on. If someone is unarmed but posing a threat to an officer (e.g., by reaching for their gun), and that threatening behavior varies by race, it's not bias, but racial differences in criminal behavior.

To be clear, the whole law enforcement system seems abundantly racist to me in all sorts of ways, but to me studies like this only become fodder for criticism. Maybe the problem is that the data isn't detailed enough, though, and this is the best that could be done. If so, the solution is better-quality data.