The problem with those PA polls showing Romney within the margin of error is that apart from a few local papers and PPP (who had Obama +6 with a 3.5% MoE) the only people polling PA were Rasmussen, Gravis and other firms with a long documented and now proven GOP bias. And, just for the record, Rasmussen (and Gallup) were NOT using 2008 turnout models, which is why they got most calls so badly wrong.

I hate to play the "pop your head out of the echo chamber" game again since we do it so often, but "yard signs prove Romney will win OH" is a Peggy Noonan quote; you can do better. As a coda on all those "skewed" and "biased" polls, one of the first stories Politico posted on election night was about the Romney campaign admitting their internal polls showed him down by 5 in OH (which was outside the MoE) just before the election.