I think we have to seriously consider whether there is some sort of a Bradley Effect in the polling on gay rights issues, although one of the pollsters (PPP, which had a very bad night in NY-23) got it exactly right. As for the model, I think I’ll need to look whether the urban-rural divide is a significant factor in a state in addition to its religiosity: Maine is secular, but rural. At the end of the day, it may have been too much to ask of a state to vote to approve gay marriage in an election where gay marriage itself was the headline issue on the ballot. Although the enthusiasm gap is very probably narrowing, feelings about gay marriage have traditionally been much stronger on the right than the left, and that’s what gets people up off the couch in off-year elections.