Gallup's final pre-election poll has a Republican lead of 15 points among likely voters, 55 percent to 40 percent. That is not only larger than the Republican lead in any other national poll and about twice the size of the average Republican lead in recent polls, it also implies a GOP victory on Tuesday of unprecedented proportions. Republicans have never had anything close to a 15-point popular vote margin in the past 80 years. In the postwar era, the largest GOP margin was nine points in 1946, followed by a seven-point margin in 1994. A 15-point popular vote margin would be considerably larger than the Democratic margin in 2008 (around 11 points) that produced 258 Democratic seats in the House. Such a margin would probably result in a gain of close to 80 seats and between 250 and 260 GOP seats in the new House, more Republicans than in any Congress since the 1920s. A House landslide of this magnitude would almost certainly also be accompanied by a Republican takeover of the Senate -- something that state polls seem to indicate remains unlikely.
Gallup's final likely voter estimate has an excellent track record. But this year they are, for the first time in recent history, a true outlier. Either Gallup is onto something that no other polling organization with the possible exception of Fox News is seeing, or they're going to have a lot of explaining to do on Wednesday.
http://www.commondreams.org/headline/2010/10/19-3
Democrats outnumber Republicans in voter turn-out in states with early voting.
My advise to voters - follow the money to see who owns your favorite candidate.
Now that is change I can believe in.
Gallup simply is providing a range based on various turnout scenarios. The key distinction they are making is that they believe there will be higher turnout than other pollsters are indicating, among the highest in recent memory.
One can argue whether turnout will rise to the levels Gallup predicts, but their underlying RV sample shows the GOP with only a +4 margin, which is absolutely consistent with what other polling has showed.
To sum it up, this is not a sampling issue. Rather it is now into the dark arts of predicting turnout based on sample data.
Given the historically high partisan gap in primary voting between the parties, even adjusting for incumbency, this forecast is not totally insane.
We have seen early voting exceed even 2008 levels in some places, so do not be surprised if Gallup is more right than wrong.