Well, this is depressing to read. Especially that my fellow non-believers were still fairly likely to think that the serial killer was more likely to be a non-believer than a religious person. Given the huge predominance of religious people in prisons vs. non-believers, this seems odd to me. However, it does appear that the study was well-done and that the sample size was large enough to draw statistically significant conclusions.
Guess we still have a long way to go.