[Tennessean4Bush]

Congressional election years without a Presidential election on the ballot should compare with each other. Presidential election years should compare with each other. You can infer some things by looking two years back, but it doesn’t always translate the same. I think Gallup does the best job, and Rasmussen, at capturing the partisan makeup and adjusting against that in the same year. At some point in Fall Gallup begins to crystallize what it believes is going to be the LV model and you see an adjustment. I think Rasmussen makes that adjustment much more regularly. I could be wrong, it has been a few years since I really studied it.

[MuttTheHoople]

Then they should make an aggregate of the 2004 election as well. I have a feeling Republicans will crawl over broken glass to vote in this election, while Democrats will have to be plied with bottles of wine and cartons of Kools in the MB Church buses to get to the polls.

I agree with Rush- the 2012 election will be a bloodbath, combining the 1984 Presidential election with the 1994 Congressional elections.