N = 885
I suspect that atheists separate themselves from the herd not only on the question of God, but in other areas as well. These are folks who ask more questions, and are more likely to buck convention.
Well, respectable society says illicit drugs are bad, and I can easily see the atheist saying, says who?
In 2003, the General Social Survey asked if people if they had have ever used illegal drugs since 1990. The top graph shows the percent saying yes by one's belief in God. On the left are atheists, and as we move right, we have those who say: there is no way to know; there is some higher power; they believe sometimes; they believe but have doubts; and they know there is a God.
We can see that atheists are much more likely to have used drugs in the past than believers. Now, I thought older people might be more likely to believe in God and less likely to use drugs, so I limited the analysis to a narrow range of people old enough to have used something in the past decade plus. Atheists again were much more likely to be drug users, but the numbers are so low, it's hard to have much confidence.
Another way to look at the relationship is logistic regression. I regressed drug use onto belief in God (treating the measure as increasing confidence in God's existence) and found that atheists are significantly more likely to use (coffiecient of -.390, a T-statistic of 3.56, p-value .000, and a pseudo R-squared of .05, N = 885).
I am assuming that the higher rate of drugs among non-believers is simply due to their non-conventionality, but are they more likely to abuse them as well? Respondents were also asked if they have an alcohol problem. The second graph above summarizes the results. The same pattern is observed.
Once again, the samples size of problem users is very low. Logistic regression shows a significant increase in alcoholism among non-believers (coefficient -.417, T-statistic 2.89, p-value .004, pseudo R-squared .05, N = 885).
Samples are larger if we look at church attendance. The bottom two graphs show that illegal drug use and alcoholism are more prevalent among those who never attend religious services. (From left to right, religious attendance moves from "never" to "more than weekly".)
Overall, we can see that skepticism toward God and staying away from church are associated not only with recreational drug use, but with abuse of alcohol. Now, is it that non-conventionality leads to more use, which in turn raises the risk of getting hooked on the stuff, or is it that some atheists find life without God unpleasant, and so they search in other places for the "spirits" they can't find at church?
Well, I calculated the mean happiness of atheists and believers, and they are basically the same. So GSS data support the conclusion that non-believers use more, and in so doing are more likely to become addicted.