Galloping: Back Up to 52-45 in Gallup, Now 49-46, Romney +3, Among Registered Voters

Although no other poll has the race as this not close, I sure feel better that one poll shows Romney way out in front.

Gallup's Frank Newport appeared on Fox News Sunday and defended his poll -- now maligned by most of the left -- as "extremely solid." Although liberals gripe that Gallup's likely voters screen is just too damn tough (it asks you where you local polling place is, for example: Likely Voters should know that), I would like to offer the possibility that other polls' screens are too soft.

Plus, Gallup now has Romney ahead with registered voters. Actually, they've had him ahead there for about a week. What's up with that?

An important thing to watch is directionality. The WSJ/MSNBC poll has generally been very favorable to Obama (the last poll had Obama out front 49-46), but now shows it as a 47-47 tie.

Rasmussen continues showing it 49-47. That's big. Rasmussen has tended to bounce around, sometimes showing a small Obama lead, sometimes a small Romney lead. Lately it's been steady for a small Romney lead (or a tie). Even nicer is Romney's four point lead, 50-46, in a battleground poll of eleven swing states (taken as a whole).

Holy Schneikies: I don't know what to make of the IBD/TIPP daily tracker poll, which has generally been somewhat favorable to Obama. It now gives him a 5.7% lead.

Oddly -- implausibly -- it claims that "moderates" favor Obama by twenty two points.

Self-described “moderates” now prefer Obama over Romney by 22 points, the second largest margin since we began polling.

Obama holds a hefty 35-point lead among urban dwellers, and a comfortable 5-point lead in the suburbs.

Yeah I don't know what the funk is going on there.

One indication that this poll might be a little goofy is that they provide breakdowns for four broad regions of the US (NE, South, Midwest, West). Obama is leading big in all areas except the South.

In the South, Romney's ahead... by a single point, 46-45.

That doesn't seem credible to me.

It also has what seems to be a weak Likely Voter screen:

Sample Size: 913 likely voters (identified from 1044 registered voters with party affiliation of 37% Dem, 30% GOP, 32% Ind.) Methodology: Traditional telephone methodology using Random Digit Dial (RDD) landline and cell phone samples

That means that more than 90% of registered voters are likely voters.

Which is not the case.

The partisan split is D+7 among the registered pool, I think. Actually they're unclear if this is the split among likelies or just registered. Even if they mean this is the split in the pool of registered voters (so we don't know the split in the Likely Voter pool), with 91% or more of registered voters being deemed "Likely Voters" the split wouldn't end up being too different.

Via @drawandstrike, they also say that 7% of Democrats will vote for Romney, but 9% of Republicans will vote for Obama.

Again, I find that unlikely.

Real Clear Politics: Has Obama with a 0.2% lead!

There are now four polls giving Obama an edge and four polls giving Romney an edge.

The two polls showing big leads are Gallup and IBD/TIPP. Gallup +7 Romney, TIPP +5.7% Obama.

Ehh. I guess overall it's a straight up tie. I tend to think tied polls are good for Republicans, though, and somewhat good for challengers generally.

One More Thing About IBD/TIPP: It says that 22% of conservative-identifying voters are voting for Obama.

Now, there is a segment of "conservative" voters who do vote Democratic. They are what Obama elegantly termed the "bitter clingers." While they are culturally conservative, they are generally in favor of Big Government largesse (though they won't call it that-- they'll put it in populist terms like "givin' the little guy a fair shake.")

That said, it seems unlikely to me that only 69% of conservative-leaning voters support Romney. 22% support Obama, supposedly, and 7% are... undecided?

posted by Ace at 02:09 PM

| Access Comments