Allegheny (Pittsburgh): Clinton 50%, Obama 44%
Philadelphia: Obama 54%, Clinton 41%
Northeast (Scranton): Clinton 69%, Obama 29%
Southeast (Philly Burbs): Obama 50%, Clinton 46%
Northwest (Erie): Clinton 52%, Obama 43%
Southwest: Clinton 68%, Obama 26%
Central (State College): Obama 50%, Clinton 44%

I have to say that these numbers do not look or feel right to me. If I were to guess the numbers in Philadelphia, I think Obama will easily clear 60% of the vote, and probably push 65%. In the Philly suburbs, I think it's safe to say that Obama will reach the high 50's at a minimum. Clinton might win Allegheny County, but I'll be surprised. And if you told me that Obama was carrying the central part of the state and holding Clinton to 52% in the Northwest, I'd tell you that he was going to win in a walk.

But I don't really have a feel for the Southwest and the Northeast. If Clinton is really carrying 70% there then there is a level of racial resistance to his campaign that I had feared but hoped would not materialize.

Here's the upside to these internals. I cannot see any way that Obama could possibly do worse in the Southwest and the Northeast than he is doing in this poll. He can do worse in the Central part of the state and the Northwest, but he is almost assuredly going to do better than this poll in Philly, the Philly suburbs, and in Pittsburgh.

Not only that, but I think turnout in Philly and the Philly suburbs is going to be extraordinarily, disproportionately high.

In other words, I do not see Clinton winning by this margin and still see her victory to be at risk.

Update [2008-4-21 10:56:59 by BooMan]: We now have more internals from the Public Policy Polling (PPP) poll that is out this morning (Obama 49%, Clinton 46%).

Southwest: Clinton 51%, Obama 38%
Northeast: Clinton 65%, Obama 27%
Southeast: Obama 58%, Clinton 32%
West Central: Clinton 57%, Obama 31%
South Central: Obama 52%, Clinton 39%

Before you get too excited about these numbers, the internals also show that this poll is predicting that the Southeast will make up 45% of the total electorate. I don't have numbers at my fingertips to assess the plausibility of that assumption but as I said yesterday, these differential regional turnout projections are the real key to predicting the outcome of the race. I do expect crazy ass turnout in the Southeast, I just don't know if 45% of the electorate is a reasonable expectation.

However, these regional breakouts look a lot more in keeping with my sense for the state than the Quinnipiac numbers. In particular, the high fifties numbers for the Southeast look about right.