Atheism - cause or consequence?

Fellow Torontonian, and far more successful blogger, Larry Moran (of Sandwalk fame) dug up an interesting study just published in Evolutionary Psychology.
To the atheists out there this paper is pretty much old news - nations which are more secular/atheistic are less violent, less crime ridden, and more socially equitable than are more religious nations.

Many take this as proof that religion is evil.  While that is a correct position, I wonder if that's really what this paper shows.

Luckily the PDF of this paper is free.  Unfortunately, the paper is only available in PDF so I cannot readily post the graphs.  So instead I'll provide one example (see image to the right) - in this case we're looking at murder (on the 'Y')  verses Religion-Secularism  Scale (low = religious, high = secular) on the 'X'.  The letters correspond to various nations U = USA, I = Italy, C = Canada, D = Denmark, etc.

 Long story made short - this paper quantified the religiousness of a handful of western democracies, and then compared that to a variety of other measures like murder rate, suicide rate, social inequity, poverty rates, STI rates, divorce rates, and so forth.  In total religiousness is compared to ~30 different societal measures.
In nearly every case, the more secular societies were better off - lower crime rates, lower degrees of social inequity, longer lasting marriages, and longer & happier lives.  Seems cut-and-dried, but a close look at the graphs (including the one above) revealed something interesting - the USA is an outlier in virtually every field.  Statistically speaking, this is a bit of a pickle - does this mean the USA is an abnormality that is incomparable to the rest of the countries analysed, or does it mean that the "outlierness" in the religious axis is representative of the "outlierness" of the USA in the other measures?

This is a double-edged sword.  If you remove the USA from all the graphs most of the advantages of secularism go away (although some remain - childhood mortality, life expectancy, abortion rate, poverty rate and a few more).  But at the same time, removing the USA from the graph also removes most of the "diversity" in terms of religiousness; meaning that we'd be looking for differences over a very small scale.

And, as a final caveat, we have to keep in mind that these are simply correlations - and as every stats teacher out there will tell you, correlation does not equal causation.  As such this study doesn't tell us religion = violent, less equitable societies.  Its equally possible that violent, less equitable societies = religion.
But my moneys riding on the former, not the latter.

Paul, G.S. (2009) The Chronic Dependence of Popular Religiosity upon Dysfunctional Psychosociological Conditions. Evolutionary Psychology 7: 398-441. [PDF]