[btilly]

First thing I found in Google is http://www.nationalreview.com/corner/382901/little-religion-....

If you identify evangelical Christians by self-identification, my statement is clearly true. Digging into the difference between self-identification and religious practice is not something that I saw before, and is a fascinating data point.

[philh]

Like I warned about in my sibling comment, the map is comparing divorce rates per capita, not per marriage. It does not say what you claim. (Also, eyeballing it I couldn't even tell you whether the south has an unusually high divorce rate per capita. It's an awful presentation of the wrong data.)

The bar chart is comparing divorce rates per marriage, and does not find that religious affiliation affects divorce rate; but churchgoing does, at least in some religious groups.