Quote:
Originally Posted by Kevin
What I think would be helpful would be to have statistics about race when police make contact with the public. I'd like to see statistically how much more likely it is for a black driver to be pulled over than a white driver. Only with information like that can we really make sense of the statistics above.
|
This retains part of the "chicken-and-egg" conundrum, though, since presumably police patrols will be concentrated in the high-crime parts of a given city. So now, we've added another factor: police presence will invariably lead to more arrests, which will lead to a greater police presence, which will invariably . . .
The controls would have to be tighter, something along the lines of "how much more likely is it for a black driver to be pulled over on a specific stretch of road versus a white driver, proportional to the raw totals of black drivers versus white drivers." This kind of data is nigh impossible to obtain, so most of the research is done in "model" form . . . bringing about its own problems, as the assumptions used to design the model affect the outcome.
The earlier issue, of course, leads to another question: for cities in which police presence is concentrated in ____ part of town (where ____ is black/hispanic/asian/white/poor/rich/stupid/ballpark/whatever), why did that start? Why does it continue?