1. Generic Ballot. Conflicting numbers last night from Gallup and Rasmussen on the generic ballot. Rasmussen finds a tighter race than earlier, with the GOP holding a three-point lead. Gallup, meanwhile, says that if the election were held today, the Democrats might be on track for a 1920-style debacle, when the Wilsonian version of the party won just 38 percent of the two-party vote.
The Gallup numbers are eyebrow raising, no doubt, and while I think a GOP victory of 10 points is a possibility, the latest Gallup numbers seem to me to be unrealistic. In fact, I get the impression that they’re still rubbing their eyes over at Gallup:
Gallup's historical election trends suggest that the race often tightens in the final month of the campaign. In September and October 1994, 2002, and 2006, Gallup's likely voter estimates showed larger margins for the leading party than what the final estimate showed (with the final poll in 2002 moving from a slight Democratic advantage to a Republican lead in the final poll). At this point, four weeks remain until Election Day, and given the already-high levels of Republican enthusiasm, it is possible that Democrats could have relatively greater gains among likely voters over the next month. This history suggests that the likely voter model results at this point should be viewed as describing the current state of affairs, but not as predictive of the final party vote shares on Nov. 2.
Gallup has been suggesting for some time that we shouldn’t put too much stock in their generic ballot numbers – last week they were indicating that the tie among the Democrats was ephemeral. Now, they’re hinting that the Republican lead is too big.
What to make of this?
I’ll put it this way, you’d have to give me some killer odds to get me to bet on a GOP +18 result, or even a GOP +13 result, which is what their expanded likely voter model has (more on these dueling likely voter models in another installment).
But beyond that, this is what I’d suggest. Put out of your mind the topline numbers, and you see something similar in both Rasmussen and Gallup: Republicans are running away with the independent vote. The differences in their final results are due to how many undecideds are left, how well both sides are sorted, and how many Democrats and Republicans are in the sample. My feeling, however, is that the two sides are ultimately going to be very well sorted (95 percent or so of Republicans voting Republican; 95 percent of Democrats voting Democratic), and the Democrats and Republicans should once again reach rough parity, as has happened in each of the last four midterms. The big question for now is how the independents break, and in both of these polls they are breaking heavily toward the GOP.
When we get closer to the election, we can start asking more fine-grained questions about whether the two sides will be evenly split in Election Day party identification (as was the case in 2004) or whether the GOP has a modest lead in party identification (the inverse of 2006). For now, the voters to watch are independents, and both Gallup and Rasmussen are in agreement on them.
2. Dems Bouncing Back? For the last couple days, mainstream media analysts have been promoting the Democratic line that the party is rebounding (see here, here, here, and here). Evidence I have seen in favor of this hypothesis includes: candidate-sponsored polls in House districts, the fact that a few statewide candidates in the Midwest have closed their gaps, and even that the DNC raised a record sum last month. Democrats have suggested that this means they’ve hit bottom, and the “Category 5” hurricane is now just a “Category 4.”
Maybe, but this theory overlooks key data points – for instance, that Barbara Boxer and Patty Murray cannot stay over 50 percent in Democratic states on the Pacific Coast. It also ignores the fact that the GOP is pulling away in the Colorado Senate race. In the last week, Republican challenger Ken Buck has increased his lead over incumbent Democrat Michael Bennet by about four points, from 2.5 percent to 6.5 percent.
Here we have a Democratic incumbent in a state that gave President Obama a nine-point victory, and it’s slipping away. Likewise, statewide Democratic incumbents are stuck in roughly the same position in Illinois, Iowa, Ohio, Nevada, and Wisconsin. These are all states that voted for President Obama - and, in all of them, incumbent Democrats are mired in the low-to-mid 40s (challengers are not faring much better in Pennsylvania or Michigan, either). The only difference in any of these contests is where the Republican candidate stands. A few of the Democratic incumbents have managed to pull their Republican opponents down a bit, but they are all still in terrible shape.
Also, Erick Erickson of RedState rightly notes that much the same argument was made by Democrats back in 1994.
3. NRCC Ad Buys. The NRCC is going up with a bunch of ad buys – totaling $4.4 million – in 45 swing districts this week. The ads follow the same basic script – linking the local incumbent to Nancy Pelosi, usually highlighting the percentage that each of the Democrats has voted with the speaker. Here’s a full list of the districts where the GOP is playing, organized by region (Democratic incumbent in parentheses). You’ll see the overwhelming majority of the buys are in the Midwest and the South – with Pennsylvania topping the list for the most districts in play. I've linked to the NRCC ads as I have found them.
East (8 Seats)
PA 03 (Dahlkemper)
PA 07 (Open)
PA 08 (Murphy)
South (16 Seats)
AL 02 (Bright)
AR 01 (Open)
SC 05 (Spratt)
TN 08 (Open)
TX 17 (Edwards)