Strategic Vision

Swann 41%
Rendell 51%
Diamond 2%

Santorum 41%
Casey 47%
Romanelli 4%

Santorum 41%
Casey 48%

Diamond of course has dropped out of the race, and his voters are not Rendell voters. If you add the 2% to Swann, then this poll gives Rendell an 8 point lead.

Rasmussen

Swann 38%
Rendell 50%
Other 5%

Santorum 40%
Casey 48%
Romanelli 5%

The crosstabs in the Rasmussen Poll indicate that the "other" category is probably Diamond voters, so I would say that most of the "other" voters are in the Swann column. Thus this poll could actually have the race as close as 7 points.

Keystone

Swann 34%
Rendell 53%

Santorum 39%
Casey 44%
Romanelli 4%

The Keystone Poll has a rather interesting crosstabs. For example, this poll contained a much more unfavorable sample for Swann than usual for the Keystone Poll. The most recent poll sampled 2% more Democrats than it did in the previous poll and 4% fewer Republicans. It should also be pointed out that the last

The final Keystone Poll in 2002 was not distinguished for its predictive quality. It had Ed Rendell leading Mike Fisher 52% to 33% among registered voters. It had Rendell leading Fisher 54% to 36% among registered voters forced to make a decision. Finally, it had Rendell leading 55% to 35% among likely voters. The result on election day? Rendell 53% and Fisher 45%. Apparently every single undecided voter went for Fisher.

In all fairness though, this seems to be a bi-partisan problem for the Keystone Poll. In 1998, the final poll had Tom Ridge leading Ivan Itkin 56% to 14%, with a third party candidate winning 8%. Once again, it seems that every single undecided voter ended up going to the challenger. Why Keystone bothered to publish a poll showing the Democrat winning 14% of the vote, and why anyone even bothers to take it seriously bothers me.

Quinnipiac

Swann 38%
Rendell 57%

Swann 34%
Rendell 54%

Santorum 40%
Casey 47%

Santorum 44%
Casey 51%

Santorum 39%
Casey 45%
Romanelli 5%

Santorum 42%
Casey 48%
Romanelli 5%

So Quinnipiac is very consistent on the Senate race, having Casey with a six point lead over Rick Santorum. However it shows Rendell leading Swann by roughly twenty points. That said, the Quinnipiac Poll is nothing that Republicans should take very seriously when it comes to a Gubernatorial election. In 2002, the final Quinnipiac Poll reported that Ed Rendell led Mike Fisher 54% to 35% among likely voters. Once again, all the undecided voters seem to have broken for the challenger?

Zogby

Santorum 42%
Casey 51%

Swann 44%
Rendell 48%

Gallup

Casey over Santorum by 18 points
Rendell over Swann by 22 points

The Gallup Poll is an outlier. Almost every other poll has the Senate race at 5 to 6 points. It seems that it is off by at least 12 points. If you subject 12 points from Rendell's lead, the spendaholic Governor only leads by 10.

Conclusion: Unfortunately, we have no really good poll on where this race stands. We have four polls that have an untested history in Pennsylvania Gubernatorial elections, three of which show Swann down between 4 to 7 points. We have two polls that have quite an undistinguished record that show Swann down around 20 points. Will Spendell win a second term? Although there is no way of being sure of how large an advantage, he does have a lead right now. However, it is a lead that Swann can easily eliminate and overturn by focusing on the completely reckless spending of the present administration.