@Robby:
"we find no racial differences in either the raw data or when contextual factors are taken into account,” means they throw out all the data that shows black people get shot more by calling it "contextual." so they can get to "police are unbiased, honest!" result.
I'm having a lot of trouble parsing this. Does this really mean what you say it means? I thought this means that "the ratio of non-white people getting shot overall seems to be consistent with the ratio of non-white people getting shot when it is justified (whatever the definition of "justified" is)". To inform the discussion further, here is a direct quote from the paper:
Using data from Houston, Texas – where we have both officer-involved shootings and a randomly chosen set of potential interactions with police where lethal force may have been justified – we find, in the raw data, that blacks are 23.8 percent less likely to be shot at by police relative to whites. Hispanics are 8.5 percent less likely. Both coefficients are statistically insignificant. Adding controls for civilian demographics, officer demographics, encounter characteristics, type of weapon civilian was carrying, and year fixed effects, the black (resp. Hispanic) coefficient is 0.924 (0.417) (resp. 1.256 (0.595)). These coefficients are remarkably robust across alternative empirical specifications and subsets of the data. Partitioning the data in myriad ways, we find no evidence of racial discrimination in officer-involved shootings. Investigating the intensive margin – the timing of shootings or how many bullets were discharged in the endeavor – there are no detectable racial differences.
A very obvious caveat to this entire exercise, and all studies of this nature is that the data is very self-selecting. For instance, the datasets used in this study were volunteered by the respective police departments, who might have an agenda (or might have curated their records before giving them to the author). This in itself completely puts any and all conclusions that can be derived from this study under a cloud. A randomised sample from the entire nation, accounting for contextual factors, would be a much better dataset to consider.@Robby:
Wait, it wasn't even based on real data, just simulations?
It was based on real data. It collected data from "from three large cities in Texas (Austin, Dallas, Houston), six large Florida counties, and Los Angeles County" along with other data sets. What the study says is that the outcome of this study is in agreement with some sort of simulation study done previously.