Rasmussen has a new generic ballot out, and that means it's time to see how the Swingometer projects the election to go based on that result.
Since last week the Democrats have lost two points in the poll, but the Republicans held steady, resulting in a R 44-D 35 reading. If we subtract out the independents, third parties, and undecideds to get a two party vote, that turns out to be R 56-D 44 (R+12). In 2008 the two party vote was R 45-D 55 (D+10). Net that's a R+22 swing from 2008. Plug 22 into the swingometer and we get a 60 seat Republican gain (40 need for majority).
That would be one of the largest changes in the partisan makeup of the House ever, if this poll is to be believed. Even 1994 saw Republicans gain only 54 seats in Bill Clinton's first midterm. To find a bigger change we have to go back to 1948 when Democrats won 75 seats on Harry Truman's coattails. The largest midterm I find going back would be Franklin Roosevelt's second midterm, losing Democrats 72 seats and gaining Republicans 81 in 1938.
2010 will be Barack Obama's first midterm. Will he really lose 60 seats? If the situation doesn't change, and Rasmussen is accurate, he probably will. But if nothing ever changed between June and November then we wouldn't need to take many polls, would we?