I know that other pollsters don't trust Zogby results, and if this were one of their Internet polls, or one of the many polls corporations and third-party groups commission from them, I would understand. However, I have never really seen any problem with their live-interview telephone polls commissioned for media outlets, so I take their tracking poll seriously. Here are the numbers from today (yesterday in parenthesis):
I'm starting to think that Obama might take this. He now holds an extremely narrow lead in the five-poll average, and the "sensitive" estimate from Pollster.com shows him ahead by 3%. In other words, the trend is with him at the very end. Only two polls show him in trouble right now, the always pro-Clinton American Research Group and the questionable Insider Advantage. There could be more polls tonight and tomorrow morning, so while I am not ready to predict that Obama will be ahead in the entrance poll, I am getting close. I know this must be incredibly frustrating to his supporters, but I am starting to think that, just like 2004, Edwards needed one more week. Clinton, by contrast, may have needed one fewer week.
As with the Democratic side, the three means of averaging current Iowa polling show different leaders. Romney leads in the five-poll average, Huckabee leads in the standard regression line average, and Romney holds a very narrow lead in the sensitive regression line average. I think Romney is going to win though, because most polls now show him ahead in the state and all but one poll gives him a favorable trend. Then again, the one outlying poll still showing Huckabee in a position to win is the Des Moines Register poll, which is supposedly the most-trustworthy of all Iowa polls.
I am also pulling for Ron Paul to somehow sneak into third place. If he finishes ahead of McCain, then I think McCain's momentum in New Hampshire is stopped. The perfect anti-McCain New Hampshire storm would probably also include both Romney and Obama taking Iowa.