At SL&P, Doug Berman has a post titled, "New research suggests race of victim impacts NC death penalty administration."  The post links to and quotes this AP article on one more race-and-death-penalty study by the notorious Pierce and Radelet.

These are the same guys who claimed to find racial "disparity" in California death penalty cases even after controlling for the aggravated nature of the homicides. If you read the actual study, though, you find that their "control" was simply to count the number of "aggravating circumstances" with no attempt to account for the fact that some circumstances are far more weighty than others. If the perp. is a mass murderer, that is one aggravating circumstance. If the murder was committed in the course of a commercial burglary, that is also one circumstance. These two cases are equal in P&R's methodology, so if they are prosecuted differently and the races are different, that is evidence of discrimination.

That measure of aggravation is so crude, it is kind of like walking into the forensic lab and seeing a technician measuring the caliber of a bullet with a balsa wood yardstick from the paint store. Um, shouldn't you be using a micrometer?

So how does the NC study control for the highly relevant, powerful factor of the murderer's prior crimes? It doesn't. Not at all. Zip.

The fact that renders the study's conclusion worthless is revealed near the end of the story. The incendiary charge that, "A convicted killer is three times more likely to get a death sentence in North Carolina if the victim is white rather than black" is the first sentence.