Bad statistics make a bad situation worse

The shootings of white police officers in Dallas is horrific and unacceptable.  It is also horrific that blacks have been treated with such violence and disrespect that many feel impelled toward violence as the only appropriate response to law enforcement.  I don't know the solution to this appalling set of events.  But I do have some understanding of the kinds of thinking that make this terrible situation worse and make resolution even more difficult.

In the aftermath of the Dallas shootings, lots of articles are popping up that cite statistical differences between blacks and whites in a variety of areas, including socio-economic levels, employment/income, and death rate.  So far every article I have seen indicates a serious lack of understanding about statistical differences.

I just read a "fact" of this kind published today. Marc Ambinder, in The Week today, said "There's overwhelming evidence that, in the heat of the moment, police officers are more likely to shoot black people simply because they are black. (If you're a black teenager, you are 21 times more likely to be the victim of a police shooting than you would be if you were white)". That's a horrific disparity, and undoubtedly, at least to a degree, reflects genuinely biased use of force against blacks.

BUT:  The ONLY way such a statistical fact can be valid is if EVERY other factor besides race were equivalent between groups.  To assume that it is entirely and only because of racial difference is to fall prey to the kinds of exaggerations that promote racial anger and bigotry.

Are we comparing, for instance, white teenagers in Minneapolis with black teenagers in Atlanta?  What about all the other differences?  Are the groups matched for education?  Socio-economic status?  Gang memberships?  Who kills the teenagers, white or black police? Other teenagers?  Are the groups equally engaged in lawful or unlawful behavior prior to the shootings?

Actually they are not matched for ANYTHING except race, which means the person quoting these "statistics" is finding what he was already looking for, racial bias by police.  We don't need to stir up the pot with misleading and misunderstood statistics.  It's bad enough, and responsible reporters and writers of articles should accept an obligation to be careful and accurate in their use of statistics.

I don't anticipate much interest in the above notes, though I think they are in fact important to understand.  But they are not exciting and they reveal that much of the statistical "evidence" cited to account for or explain or justify the shootings in Dallas is primarily emotional and a dramatic interpretation of statistics to exaggerate and justify the shootings.