New Hampshire exit polls show a huge gender effect. Women voted for Clinton over Obama and Edwards at 46%-29%-15%. (Contrary to some reports Edwards doesn't appear to have held onto women much worse than Obama.)

Independents voted for Obama at 41%, to 31% for Clinton and 18% for Edwards. As for the youth vote -- Obama got the 18-24's, lost the 25 to 29, won the 30's and lost in the 40 and above areas (hat tip BooMan). The tears seen around the world, combined with the ironing stunt and a fair chunk of misogyny in the media may have something to do with it.

The most interesting numbers to me were the percentage of favorable and strongly favorables who voted the candidates they favored. Amongst the strongly favorables Clinton got 81%, Obama 66% and Edwards 48% (meaning Edwards got less than half the people who love him, while Clinton got most of them). Even Obama was having trouble sealing the deal with people who like him. Even more interestingly people who just had "somewhat favorable" views of Obama and Edwards voted at 8% for Obama and 40% for Edwards while Clinton picked up more of their favorables than either of them did - 51% of Obama's and 43% of Edwards'.

In terms of Edwards dropping out (which, as I've said, I don't see him doing yet, or probably at all), the split is 29% to 19% among highly favorables who are already crossing over. If the split continued along those lines then Obama would get a significant boost from Edwards dropping out, and perhaps an even larger one if Edwards gave him a strong endorsement and tried to deliver support. If I were Obama I'd be on the phone, pronto, but the problem is I don't really see much that he can offer Edwards and from Edwards point of view, as long as his numbers aren't dropping, he probably figures he can always deliver delegates at the convention if it comes to that.

One wonders if this was related to electability concerns, with Clinton being viewed as more electable that Obama and Edwards.

Clinton got more union votes than anyone else, Edwards did slightly better with union voters than his overall numbers.

In general those with more education and more money went Obama/Edwards, while those with less went for Clinton.

There is some evidence for racism in Obama's inability to break out of the younger age groups (which are less racist overall) and in his striking inability to convert favorable numbers into actual votes.

However the split in gender is enough to explain the differences between the polling and the final results, and pollsters are justified in saying that a lot of the events that made the gender split occur happened after their polls, so odds are the Bradley Effect was not a major factor (though we'll keep an eye out on future polls and results).

The race is far from over. Unless events keep happening to reinforce it, I expect the sympathy effect to subside amongst female voters and even with it, Clinton's edge wasn't that large. No one has this wrapped up yet.

Nonetheless Obama clearly needs to figure out how to reach beyond his base to those who think somewhat well of him, but for some reason aren't breaking to him.

Clinton.has done the best at turning potential support into real votes. Obama and Edwards aren't. They need to figure out why.

Photo credit New York Magazine