Second look at dog-eating?

That’s the biggest lead he’s had against Romney dating back to last August. Follow the link and scroll down and you’ll see that it’s independents who are driving it, flipping from a 45/39 split in favor of Mitt 10 days ago to a 45/43 in favor of O. Gallup reasons that a dip in gas prices must be behind it, but the dip we’re talking about here is roughly five cents in the national average. Hard to believe you’d see a poll effect this dramatic because of something as meager as that. I think what we’re looking at here is a very noisy poll.

Not sure about this one, though:

The WMUR Granite State Poll shows Obama leading Romney 51 to 42 percent [in New Hampshire]. That’s close to the 10-point lead Obama had over Romney in December…

The results are a dramatic flip from October, when Romney led Obama 50 to 42 percent. Smith said that shows the Democratic base is mobilizing…

The poll shows a mixed picture in terms of the president’s approval rating in the Granite State. Fifty percent said they approved of the job Obama is doing, while 47 percent said they disapproved. The gap has shrunk a bit since February, when 51 percent said they approved and 43 percent disapproved.

I don’t understand the “Democratic base is mobilizing” logic when it comes to polls. Their base would have answered this question the same way in October as they’re answering now. It’s more likely that Romney’s numbers have shrunk because, thanks to six months of a bitter GOP primary, indies know him better now than they did then and some don’t like what they see. That’s not a disaster — conservative Super PACs will begin the attack ad carpet-bombing campaign against O soon enough — but surely no one’s thrilled to see Mitt start the general election this far back in his own backyard. Nominating a “Massachusetts moderate” was supposed to help Republicans put northeastern states in play, states like New Hampshire and … er, realistically I guess that’s it.

I have a powerful eeyorish urge to go curl up in the fetal position so I’m going to go do that now. Here’s something to entertain yourself with, though — a new election model from WaPo that purports to predict the likelihood of an Obama victory or defeat based on two simple indicators, economic growth over the first three quarters of this year and his approval rating in June. In order to knock his chances of reelection below 50 percent, you need zero percent growth and a job approval of 45 percent. Possible, but not likely.