There have suddenly been a rash of polls over the last week or so in Virginia. The list is as follows (all scores McDonnell-Deeds):
Clarus Research: 42-37
DKos/R2000: 50-43
Rasmussen: 48-46
Washington Post: 51-47
Insider Advantage: 48-44
Before going into depth on the bits and pieces of each one, let me give the general gist that I'm getting on the direction of the race:
- The thesis story has made a dent but not a crater. It will probably continue to do a bit of damage as events unfold, but it doesn't seem to be doing much to Independents or Republicans (especially the latter, many of whom like the ideas in said thesis). Thus its impact was in solidifying Deeds' base, and just doing this isn't going to be enough to win.
- McDonnell is doing slightly better among Democrats than Deeds is among Republicans. Surveys vary, but he seems to have a slight edge in this regard.
- McDonnell's edge comes mainly from a large edge among independents. Though the edges vary from 7-22 points; I tend towards about 12-15 once you squeeze undecideds out. This is ultimately what Deeds needs to rectify in order to win.
Poll-by-Poll:
The Clarus poll has the highest undecided count (21%), which suggests that they didn't lean on the undecideds much at all; nobody else even yields half of that. The flipside is that those undecideds probably trend slightly McDonnell (based on his lead among Independents and the fact that Independents yield 1/3 undecided in the Governor's race). As with most polls, down the ballot the picture is much the same though the undecideds keep rising (Independent undecideds are close to 50% in both of the other races).
One thing of interest was that over half of voters hadn't heard about the thesis story (though said story doesn't seem to be generating a lot of waves), though I'd also echo something noted in the polling analysis: Democrats tend to be outraged and close ranks around Deeds, but nobody else seems to be caring much.
Final note on Clarus: The sample broke slightly Republican on party ID (29-27 R/D; 37-34 R/D with leaners added).
Next up is Kos. This sample trended Democratic (39-34), yet in spite of this McDonnell posted his biggest lead in the latest polls. Part of this is that McDonnell got a lot of Democratic votes in the sample (15%), but part is also that he posted a 13-point lead among Independents.
As usual, I'll skip Rasmussen. No crosstabs means no meaningful analysis is really possible, and I'll throw in my usual mention of Rasmussen's methodology issues as of late.
The Washington Post poll did a pretty good job of squeezing out the undecideds. The sample here was also more Democratic than before (the last poll they did, back in August, showed a 2 point Democratic edge among registered voters and a 7 poing GOP edge among likely voters; this one showed a 7-point Democratic edge among RVs and a 3 point Democratic edge among LVs). The 4-point lead was also close to the consensus result.
Unfortunately, WaPo's poll also declined to offer any crosstabs, but at the very least I can say that part of the large swing between this poll and the last (which showed a 15-point McDonnell lead) was the sample issues in the summer polls which I spent most of the summer grumbling about.
Finally, the Insider Advantage poll gave some good crosstabs. It had a slightly Republican-leaning sample (about 35-33), and gave McDonnell a 4-point lead. This poll was curious in that Deeds did better among crossover voters than did McDonnell (12% vs. 8%), but McDonnell thumped Deeds among Independents (55-33). Those two factors make me a little bit leery here. Both are probably just sampling noise, but I'll be keeping my ear to the ground to see if another poll comes out.
McDonnell's lead has shrunk over the last week or two from the thesis, but Deeds probably needs to pull another issue out of his hat to be able to win this. McDonnell's campaign is increasingly focusing on tying Deeds to Obama and Pelosi, as the drilling issue is fading as gas prices fall. Deeds needs one more issue he can hit McDonnell on, and he needs one that will resonate with Independents and/or whites (the key demos he seems to be having trouble with) if he's going to win.
Overall, I think the polls have a good handle on the race, in spite of the inconsistent sampling. I'm wondering if wording had something to do with the wild swings among Indies in these polls, but for now I can't be sure.
Finally, on a topic of personal interest, Virginia residents seem to slightly favor a public option (48-42), with 20% of Republicans supporting it in the Kos poll. I'll file that under "pleasant surprises" for this week.