Gallup generic congressional narrows slightly.

[FBaggins]

Though the registered voter gap expands to five points. Remember that Gallup's poll combines two weeks of polling, so some of this reported data was also included in last week's release. This is the second straight week of slight tightening. Obviously that's better than the alternative, but not fast enough if Gallup's LV model is close to accurate.

PRINCETON, NJ -- Gallup's tracking of the generic ballot for Congress finds Republicans leading Democrats by 5 percentage points among registered voters, 48% to 43%, and by 11- and 17-point margins among likely voters, depending on turnout. This is the third consecutive week the Republicans have led on the measure among registered voters, after two weeks in September when the parties were about tied.

...snip...

If the elections were held today and roughly 40% of voters turned out -- a rate typical in recent years -- Gallup's Oct. 7-17 polling suggests Republicans would win 56% of the vote -- 8 points greater than their support from registered voters, and 17 points ahead of Democrats, at 39%. If turnout is significantly higher, Republicans would receive 53% of the vote (a 5-point improvement over their registered-voter figure), and the Democrats, 42%.



http://www.gallup.com/poll/143777/GOP-Holds-Solid-Leads-Voter-Preferences-Week.aspx

[WI_DEM]

1. I think these "likely voter" models are for the birds. I don't see the GOP winning by 17-points

if they did they would win 100 or more house seats and that's not going to happen.

[FBaggins]

2. I'm certainly hoping that Gallup's is.

But there's too much of a track record for LV polling in general to just dismiss the entire set. A far smaller percentage of registered voters vote than say they will vote when polled. Pollsters have to come up with some way to identify the "likely" voters from the people they call.

The standard questions have a pretty good track record. The odd thing about Gallup's number is that they try to assign a sliding likelihood and provide multiple estimates based on turnout... and it's a little odd how they do it. Rather than weight the pool of people with a 45% likelihood of voting, they count them all at turnout levels above that line and none of them at levels below that line.

I think that oversimplifies reality more than is necessary.

[flpoljunkie]

3. Bottom Line: Not clear how likely these patterns are to continue through the month

Bottom Line

U.S. voters' preferences for the party they will support in this year's House elections have been quite stable over the past three weeks, with Republicans leading by low single digits among registered voters. They lead by substantially more than that among likely voters, including both high-turnout and typical-turnout scenarios.

It is not clear from the historical record how likely these patterns are to continue through the end of the month. Gallup's pre-election polling in prior midterm years documents that there were some years when the structure of voter preferences by this point in October was generally maintained through the elections (1998 and, to a lesser extent, 1994), but others when it changed substantially (2002 and 2006).

http://www.gallup.com/poll/143777/GOP-Holds-Solid-Leads-Voter-Preferences-Week.aspx

[Imajika]

4. These are catastrophic numbers...

I have my doubts about Gallup's LV screen, but if this scenario plays out it would be a tsunami and we'd lose seats we haven't even been talking about.