New PA poll

The American Research Group has a new poll this morning, showing Hillary 20 points up on Obama in Pennsylvania. Glenn Reynolds doesn't see "how it could possibly be good news for Obama."

But I do.

OK, the poll itself isn't great news. But lets face it - she's going to win Pennsylvania. And it has never really been in doubt. The relevant question right now isn't "who wins Pennsylvania," it's "what's the margin?" And, even more importantly, "what's the perception of the margin afterwards?" If everyone knows that you're going to lose, you can win even by losing, if you "cover the spread." What Obama needs is for one of two things to happen in Pennsylvania. The first would be an outright win. That would effectively end the race. But it's not going to happen. The other thing that would work for him is to lose close, or at least closer than people predict. To that end, the worse the polls look for the next week, the more opportunity there is for him to beat expectations, to argue that his "bitter" comments and Reverend Wright haven't hurt him ("after all, she was always going to win") and head into North Carolina denying that she's taking any momentum out of a state where she won by less than everyone expected.

That's the way it works. Sometimes, winning the numbers is less important than winning the perception, and if they go into next Tuesday with people expecting her to win by 20% and she wins by 12%, then he wins and she doesn't. And the Democratic race is entirely about perception now, as neither candidate is going to finish with a majority of the delegates. The entire race, at this point, is about building a persuasive argument for the super-delegates and party "wise men," because that's who will end up making the decision. When the rank-and-file Democrats have had their say, they won't have chosen a nominee yet. To use a sports analogy, the Republican primary campaign was a "stop-watch" campaign, decided by objective reality - McCain won the delegate count. The Democratic primary campaign is going to be a "judging" campaign, where the party elders end up choosing from two candidates who failed to reach the finish line on their own.