It still jumps to a lot of conclusions. First, it assumes that all religious people are created equal, in that it lumps e.g. all Christians and Muslims together without really attempting to look at the diversity within that group. They also failed to address any potential disparity between parents' reports of religiosity and what the kids' were. The authors make a big deal out of the disparity between parents' prediction of their kids' generosity and the actual results, but they don't bother to look at whether that same disparity exists between what parents think of their kids' religiosity and what the kids actually think. Their methods also don't explain why they took a greater condemnation of violent behavior from others, and then somehow twisted that to mean "less empathic." Seems like they decided the answer they wanted first, and then filtered the results through that in order to get where they wanted to go.