I’ve been thinking about this a little more recently, and I’ve been trying to think of what would make a more compelling argument. Don’t get me wrong, I’m extremely anti-religion, but I’m also anti-crappy-argument. (Although some people I’ve discussed this with online have taken my latter stance to mean that I’m actually a religious moderate. Words fail me. Them too, I guess)
Data would make “religion causes people to do evil things” a lot stronger, rather than the typical post-hoc religious arguments that people make. So here’s a sociological observation that would bear that out.
Take, say, the population of the United States, and document as many of the churches as possible (of any and all religions). Attend sermons for six months to a year. Create a transcript of all the sermons, and (if they have them) everything said in the discussion groups. The names of all speakers can be anonymized from the outset, as this is not immediately relevant (and this would allay some privacy concerns). All of this should be coded for violence, calls for retribution, calls for peace/passivity, etc.
Correlate the catchment area of those churches with home addresses of people charged with crimes within that period. Control for all socioeconomic variables, and compare like with like.
Generally speaking, criminals tend to be produced by lower socioeconomic areas (due to lack of opportunity, role models, etc). If this were controlled for, and it could be shown that there were a significant difference in crime rates between two socioeconomically similar areas where the only difference was that the Religious Speaker in Area A focused more on the retributive/violent aspects of their religion vs the Religious Speaker in Area B, then I would start to lean more towards the notion that “religion causes people to do evil things”.