Just a quick update on PA polls. Good news for Toomey in the new polls, but some of the samples just seem way too Republican for Pennsylvania
Franklin & Marshall poll has Toomey up 43-36, 48-40 with leaners. Sample is GOP +2 Poll date 10/18- 10/24
http://images.lancasteronline.com/local_old/305240/F-MPollOctober2010.pdf (broken link)
Taken along the same time was a Reuters/Ipsos poll taken 10/20-10/24 showing a 46-46 tie among likely voters, Sestak up 46-42 among registered voters. That sample was 46-45 GOP among likely voters
http://images.lancasteronline.com/local_old/305240/F-MPollOctober2010.pdf (broken link)
CNN Time Poll 10/20-10/26 Toomey up 49-45 with likely voters, among registered voters Sestak is up 47-43 among registered voters. It doesn't offer the % of each party, but does show how they do among Republicans, Democrats and Independents. Doing the math it appears that the likely voter sample is GOP +2 to GOP +4.
http://i2.cdn.turner.com/cnn/2010/im.../topstate8.pdf
In the Morning Call/Muhlenberg Tracker Toomey is back up to 8 48-40, poll date 10/24-10/27. No cross tabs released yet (will post it once its out), the sample has generally ranged from GOP +3 to GOP +5
Now, one reason why I am questioning some of these Party Registration and ID's in the likely voter samples is because the Dems have a rather strong registration and Party ID advantage in Pennsylvania. Now its obviously not going to be at the typical level, GOP turnout will be stronger. However, the Democrats have always had a very strong ground game and GOTV effort in Pennsylvania.
Some of these differences between registered voters and likely voters will materialize in some parts of the country. The GOP will likely have the advantage across the country, but in certain areas I don't think the advantage will be as strong as the gaps suggest. Areas where the Dems have always had a strong ground game such as Pennsylvania is one of those areas I see the GOP maintaining that LV advantage compared to RV's, but not at the levels suggested.
I also think you will see a narrowed gap in areas where the Dems rely on the well educated liberal vote, Colorado (Denver & Boulder) and Wisconsin (Madison) are examples of that.
That isn't to say the GOP isn't going to have a turnout advantage on election day compared to registered voters in areas such as Colorado, Pennsylvania and Wisconsin, they still will, I just don't see the gap as large as some of these polls are suggesting.