New PA polls out from Quinnipiac, SUSA, and Rasmussen.

Latest Pennsylvania primary polls:

If you believe last week’s SUSA poll was valid, Obama has improved 4 points since then. Now kick yourself for believing last week’s C+18 SUSA poll was valid, you fool! Yes, even SUSA produces bad data every now and then (*cough*Missouri*cough). No, this week’s outlier is yesterday’s ARG poll; ahhhh, that feels better. The idea of ARG looking more accurate than SUSA was giving me the heebie-jeebies, but you know what they say about broken clocks. In this case, sometimes a bad pollster produces a better result than a good one, but usually not. And once again, ARG’s notorious for changing their turnout model for their final poll of the race so that it matches better with the other pollster’s results.

So what’s the sitch? Quinnipiac shows no change from the week prior, while Rasmussen gives the most rational looking result, with a four point improvement for Clinton. Rasmussen also produces a good median (C+9) between Quinnipiac and SUSA (FORGET ARG!). So, I’m prepared to update my prediction for the week. Where last week it was “Clinton by less than 10%”, as things stand today I have gone back to my prediction from two weeks ago, Clinton would win Pennsylvania by 5-15% if the numbers remained as they are. This fits in with Elliot’s prediction of C+9%, and a delegate split of C+14. In other words, not the blowout Clinton desperately needs, but enough to carry her limping campaign to North Carolina and Indiana. A stay of execution, if you will.

In other news, Rasmussen daily tracking has already put Obama’s lead over Clinton back to 50%-41%, after yesterday’s 6% lead and Sunday’s 1% deficit. Looks like “bittergate” has little traction nationwide. In addition, looks like Gallup is showing Obama picking up 1 point, he’s now 11-points ahead of Clinton.

(Note from Elliot): I should also point out that if you aren’t counting Susquehana or ARG (the former has an unusually high number of undecideds and the latter is, well, ARG), it would appear that Obama has a floor at 40%, that means that it is almost certain that the absolute best that Clinton can hope to do in Pennsylvania is to beat Obama by 20% (which I believe to be unlikely). According to the turnout model by Daily Kos diarist PocketNines, this means that Clinton is unlikely to get enough votes in Pennsylvania to put her within striking distance of Obama on that front (even if you include Florida, which even though I don’t believe it should be factored, is more reasonable than trying to make the intellectual argument that Obama has zero support in Michigan).