Even more fun with numbers

Soon after I wrote my post of yesterday regarding the Republican advantage in generic polling for Congress, Gallup released its final numbers. They put the Republican advantage at 10 to 15 percent, depending on turnout assumptions. And Gallup’s analysis of turnout indicators suggests that turnout will be on the high side. The high turnout model yields the 15 percent advantage.
The survey period is Oct. 28-31. The sample size is 1,539 likely voters, much larger than the samples used in the three polls I wrote about.
A 15 percent advantage in the actual voting would probably mean a net pick-up of more than 80 seats, which was my top-line number [note: one model says 76 seats, but I don’t think I buy that]. And I agree with Bill Oits who says “If we get 70, the Dems will be shell shocked in a way not seen since 1984.”
A wave of this magnitude would likely help produce Repubican wins in nearly all of the close Senate races, wouldn’t it?
UPDATE: There is some confusion as to whether it is Gallup’s high turnout or low turnout model that produces the 15 point Republican lead. In saying that it’s the high turn-out model, I relied on this statement from Gallup:

Gallup’s analysis of several indicators of voter turnout from the weekend poll suggests turnout will be slightly higher than in recent years, at 45%. This would give the Republicans a 55% to 40% lead on the generic ballot, with 5% undecided.

However, Nate Silver of the New York Times, a professional at analyzing polls, says that Gallup’s traditional turnout model has the Republicans up by 15 points, while the higher-turnout model has Republicans up 10 instead. One of our readers told us the same thing.
As a general matter, I think low turnout in an off-year election favors Repubilcans, since Democrats do better with registered voters than with likely voters. But maybe in a “Tea Party” year, high turnout among likely voters favors the Republicans.