As of this moment, about 9:25 PM Eastern Time, there are less than 15% of precincts reporting, a Clinton margin of 6%, and most networks have already called Pennsylvania for Clinton. But I took a quick look at some critical numbers, and this thing could end up really, really, really close.
SurveyUSA has generally been the most reliable pollster so far this primary season, and their final Pennsylvania poll had Clinton at 50% and Obama at 44%. I also generally trust CNN's exit polls, and their demographic data samples match up in a lot of ways, and I was curious to see the degree to which the actual numbers themselves matched.
Of course, keep in mind that actual votes may not match up - however, even as Kos points out on the front page that Obama has a 10 point margin in Pennsylvania, less than he needs, it has since expanded to 16 points. This is a fluid situation, and I could be completely right or completely wrong.
So I've selected a few critical demographic groups for a selective breakdown (as in, I didn't go through each PA region, especially since SUSA subdivided into 6 and CNN into 5): race, age, gender, education and region. I would have posted all of the results directly to this diary, but I'll link to a screen capture of my results instead. Everything that is highlighted in blue is a "good for Obama" number, as in that either the turnout of the demographic in question was favorable (or neutral) to Obama or the percentage of voters within that demographic was favorable. The "really good for Obama" numbers are highlighted in purple, and the "good for Clinton" numbers are highlighted in green.
The bottom line here is that Obama probably didn't lose ground on any of the selected demographics, and gained a significant amount of ground in most of them. The only favorable data for Clinton are that she outperformed her SUSA numbers within gender, and got higher female turnout, and that there is better turnout among voters 65 and older. Other than that, Obama generally kicked ass across the board and way outperformed, especially on a regional basis. Basically, the last SUSA poll had southeastern PA at 43% of the vote, with Obama winning the region 55-41. If SUSA sampled correctly and it is closer to 43% than 29%, and CNN got the exit percentages right, Clinton is truly in serious trouble. This is especially true given Obama's vastly improved performance among female voters.
So before anyone starts worrying about Clinton racking up serious margins as she goes to victory, take another look at this data. Obama might not win, but it is going to be damn close (if I'm right), and that's victory enough for me.