via US Police Shooting Data, Part 2: A Civilian’s Review
There’s more tweaking to do inside the Vice data set of big-city police-shooting incidents, though some tweaks are more concessionary than ameliorative. The abundance of NAs flecking the sheet’s fields is not a matter of slipshod data collection; Vice’s preface to the data recalls the issues the site experienced with respondent cooperativeness and data quality. But missing data are missing, and so the fact that shooting victim ages (SubjectAge, in column H) don’t appear for more than 53% of incident cases remains a fact; and while of course the available ages of the other approximately 2000 victims amounts to a sizeable cohort, one wonders if the absent data inflict a sampling skew on the ages. I suspect they don’t, but the question should be recorded.
