Why Were the NH Polls So Wrong? What Will Happen Next?

Delegate count now Obama: 25, Clinton: 24, Edwards: 18.

Next up: Nevada. Obama gets the state SEIU endorsement (which looks like it was a bit of a chaotic affair); the union has 17,500 members. He also picked up the endorsement of the 60,000 member culinary workers. Since the Nevada Democratic party is anticipating that 40,000 people will take part in the caucus (though Harry Reid says it could be as high as 100,000), it certainly bodes well for Obama in the state.

As Mike Lux notes, the momentum is still with Obama but Edwards becomes a kingmaker:

Edwards is dead as a Presidential prospect for 2008, but if this becomes a long, drawn out, fairly evenly-divided fight, and Edwards stays in and keeps drawing around 20% everywhere he goes, he could be the guy who decides the nominee. I could easily see this playing out where Obama wins a bunch of states and Clinton wins a bunch, and they both end up with about 40% of the delegates, and Edwards keeps playing to the end of the primary process and then delivers his 20% to one of them. Could make things interesting right up to convention day.

The reports of life in the Clinton campaign may be premature. If in fact a late display of emotion and the media pile-on it triggered helped Clinton over the finish line in New Hampshire, it isn't a repeatable formula. She needs to retool her messaging. And in the name of all that is holy, get those sleezebag surrogates off the televison set. Just when you think you may feel some sympathy for her again, one after another slimes across your set looking like they've either just come from selling smack on 6th and Bonnie Brae (Jay Carson) or they want to know how they can get you in a Le Sabre today (Terry McAuliffe).

I know Mark Penn has a penchant for manufacturing polling to meet what he already believes, but they got their asses handed to them that way in Iowa. Run a Q Score on these guys before you cut them loose again. It has to be sub-arctic.

Sullivan speculates that the reason the polling was so off yesterday is because of the "Bradley Effect," where white voters say one thing to pollsters and then do another in the voting booth. He posits that since the Iowa caucus votes were cast in public and the New Hampshire ballots were private, it may have had something to do with the ability to accurately poll the race. But Pollster had a standard estimate for Obama of 36.7% going into last night, and he wound up with 37%. So the Obama polling was accurate...it was everything else that was off.

****

Why were all the New Hampshire polls so utterly wrong? How did Clinton overcome such a huge deficit in the polls in one day?

Chris Bowers peculates:

This almost never happens. Here are some possible reasons: