The only data point I've seen that empirically supports religious prevalence is that very christian countries have a statistically significantly lower level of dishonesty. (see: China for a counterexample)