There are two problems with that: 1. I don't think the data actually exists. The commonly cited figure of 0.2% atheists is extremely dubious. 2. There are too

There are two problems with that: 1. I don't think the data actually exists. The commonly cited figure of 0.2% atheists is extremely dubious. 2. There are too many confounding factors and hidden variables. A plurality of Hindus speak Hindi, but believing in Hinduism neither causes, nor is caused by speaking Hindi. Likewise, if atheists are underrepresented in prisons, there are just too many simple demographic explanations to draw any conclusions about morality.