Religion and Parties

Yesterday, Noah raised the question about state-level partisanship and I mentioned the role of religion.  Chris Cillizza writes today at The Washington Post:

How religious you consider yourself is a remarkably accurate predictor of which party's presidential candidate you will vote for.

That fact was affirmed this week when Gallup released a 50-state study of the most and least religious states in the country. We took the data -- which GovBeat wrote about here -- and overlaid it with the 2012 presidential election results. Here's what we found.

The 19 most religious states -- ranked by Gallup as those who identify as "very religious" -- all went for former Massachusetts governor Mitt Romney in 2012. (Romney won 24 total states.) In those 19 states, President Obama averaged 39 percent of the vote. ...

 On the other end of the (religious) spectrum, the opposite is true. President Obama won the 14 least religious states in the country. He averaged 61 percent of the vote in those places; if you take out the District of Columbia, which Obama won with 91(!) percent, the president averaged 59 percent in the remaining 13 states.

Needless to say, other forces were at work in 2012.  And as those of you who have studied stats should remember, you should always beware of the "ecological fallacy" -- assuming individual behavior from aggregate data. In this case, however, survey data do confirm the relationship between religiosity and the vote.  See exit poll data: