Religion and the 2004 Election

[YoMartin]

This is quite interesting. I would only say that you need to add other elements in the picture. By only considering religion you could end up explaining the differences by the wrong reason. If Northeast catholics vote Democrat but Southern catholics vote Republican, then religion is not the cause for their voting behaviour. Also, the plurality of voters in certain county may be catholic, and candidate X may win that county, but maybe all the catholics voted for candidate Y while all the other people voted for the winning X. That´s the problem with aggregate data, which can only be solved with rather complex statistical models, or by just using polls.