Updated

Tensions are still high in Ferguson, Mo., after the Aug. 8 police shooting of Michael Brown. On Friday, it was reported forensics showed Brown’s blood was on both the inside of police officer Darren Wilson’s car as well as Wilson’s gun. The gun had also been fired twice within the car. This evidence hardly squares with early witness accounts that Brown was shot with his hands in the air while he was surrendering.

With forensic evidence finally coming in, Officer Wilson’s shooting Brown is looking as if it were justifiable self-defense.  But that hasn’t stopped people from making it a racial issue.

[pullquote]

Over the weekend, the New York Times noted some black leaders, such as Congressman Elijah E. Cummings, D-Maryland, “often invoke voting rights and the death of Michael Brown, the unarmed black man shot by a white police officer in Ferguson, Mo., as a way to mobilize black voters.

About a week ago journalists at ProPublica and Slate further fueled the anger many blacks felt about the shooting. Slate’s headline read: “Black Teens Vastly More Likely to Be Killed by Police Than Whites Even After Adjusting for Crime Rates.”

The incendiary finding got massive uncritical news coverage including the New York Times, NPR, major newspapers around the country such as the Atlanta Journal Constitution and the Dallas Morning News, and the major networks.

If their claim is right, and blacks are unjustly facing “vastly” higher rates of being shot by police, it is a serious indictment of the police. Fortunately, these allegations are wrong.

There are two parts to their assertion. First, that black males ages 15 to 19 died at the hands of police at a rate 21 times the rate of white males of the same age.

Second, that black Americans are no more than two to three times more likely to commit violent crimes than whites. They put these two points together and conclude that after adjusting for violent crimes by race, police still kill young blacks at least 7 to 10.5 times as often as they kill young whites.

But if you’re going to look at the rates that police kill black and white male teenagers, you have to adjust for the rate that black and white male teenagers put themselves at risk – not the rate that both male and female blacks and whites of all ages commit violent crime.

Among blacks, teenage crime is much more of a problem. Black male teenagers were nine times more likely to commit murders than similarly aged white males, not two to three times as likely.

Doing the math, blacks male teenagers are killed by the police at a 2.3 times greater rate than whites after adjusting for these murders. It’s still considerably higher, but by a fraction of the rate asserted by Slate.

Can we still conclude that police singles out black teenagers?  As ProPublica acknowledges, the data on justifiable police killings are “terribly incomplete,” and “vast numbers of the country's 17,000 police departments don't file fatal police shooting reports at all.”

But there is an additional problem.  The numbers aren’t random.  It so happens that urban areas, that are 50 percent more heavily black than the rest of the US, are the ones that report justifiable police killings. That means police probably kill whites more often than these incomplete statistics indicate.

While ProPublica assures us the errors in the data can’t offset the 21-fold difference, it is highly unclear whether the actual, much smaller remaining 2.3-fold difference is merely a result of such incomplete data.

ProPublica justifies its conclusions by quoting David Klinger, a University of Missouri-St. Louis professor and an expert on these data. But Klinger told me in a phone call that he clearly informed ProPublica that he didn’t want to be quoted on the issue because the FBI Uniform Crime Report data on justifiable police homicide is “no good.”  (One of the ProPublica authors, Ryan Gabrielson, denies that Professor Klinger told them this.)

Extreme claims, such as those by ProPublica and Slate, might generate readers.  But let’s not scream “racism!” before carefully studying the data.