Full disclosure here; I made a slight correction to my previous post. As it made it sound as though black people being subjected to disproportionate police-civilian interaction was purely random chance, which was not my intention.Per interaction, Yes. But whites and blacks don't have the same rates of police interaction.
This study does not contradict the statistics in past studies, what it does is help understand them better. What this study combined with past statistics suggest, is police don’t fatally shoot unarmed black men at disproportionate rates because that is what they set out to do; but rather purely a product of initiating more frequent interactions with members of the community in question (this also explains the marijuana arrest discrepancies).
The disproportionate interactions are likely do to some combination of socioeconomic status, area, and racial bias (given the authors inability to eliminate it as a factor in the case of non-lethal force, it stands to reason that the lesser act of initiating interactions is likely affected as well). But this topic needs more study to figure out how it breaks down. Unless someone knows of one?