One of the commenters to my post below suggested that Mike Huckabee was unlikely to do well among Catholics. Philip Klinkner (who is really blogging interesting stuff on the races) has some county-level data from Iowa suggesting that this is true.
GOP caucus results (counties won by Huckabee in blue; by Romney in red)
Distribution of Catholics in Iowa (the redder the county, the more Catholics)
An eyeballing of the graph suggests that the parts of the state where Huckabee had most trouble were indeed, more often than not, those places where there were more Catholics. Klinkner runs a regression testing how percentage of population religious, percentage of population Catholic, percentage of population evangelical, and percentage of population rural affected voting for Huckabee, and finds that the coefficient for Catholicism is negative, high, and statistically significant.
Update see here for Klinkner’s response to some of the methodological criticisms.