Another week, another set of polls. The latest SUSA poll suggests a drop in support for Deeds and a boost for McDonnell; the latest Rasmussen one also shows a similar, if less pronounced, pattern. PPP shows no such change. The basic numbers are as follows:
PPP: McDonnell 48, Deeds 43 (McDonnell +5)
SurveyUSA: McDonnell 55, Deeds 41 (McDonnell +14)
Rasmussen: McDonnell 51, Deeds 42 (McDonnell +9)
The long and short of these numbers is that the Thesis Bounce seems to have worn off for Deeds, but it did at least seem to tinker with who was supporting each candidate, with the bases solidifying somewhat.
Rasmussen goes to the side, as usual, though I will talk about that poll in brief in a bit. There are some interesting dynamics at work here, though, particularly in the PPP sample, and as usual, I'll take each poll on its own.
First, PPP
In the case of PPP's sample, I compared the stats from their last poll (carried out at the end of August) with their current one. The samples were as follows:
Result, 2009 Governor's Race:
August: McDonnell 49, Deeds 42
September: McDonnell 48, Deeds 43
2008 Vote
August: McCain 49, Obama 45
September: McCain 45, Obama 48
Party ID (D-R-I):
August: 38-31-31
September: 37-29-34
Clearly the samples weren't identical. There was a slight shift Democrat-wards in the samples (about 3-4 points), mainly in the sample favoring independents. Now, in all fairness, there were some slight shifts within these groups...unfortunately, they were not the kind that I suspect we want to see:
McCain Voters, August: 86% McDonnell, 7% Deeds
McCain Voters, September: 88% McDonnell, 7% Deeds
Obama Voters, August: 9% McDonnell, 82% Deeds
Obama Voters, September: 8% McDonnell, 80% Deeds
The shift, based on 2008 vote, is slightly away from Deeds. The same thing shows up with a party ID breakdown:
Republicans, August: 89% McDonnell, 4% Deeds
Republicans, September: 96% McDonnell, 2% Deeds
Democrats, August: 7% McDonnell, 85% Deeds
Democrats, September: 5% McDonnell, 82% Deeds
Independents, August: 60% McDonnell, 29% Deeds
Independents, September: 53% McDonnell, 37% Deeds
In short, Deeds has made up ground among independents but lost ground not only among the GOP (which is honestly not as big a surprise as one might think if one expects voters to, in general, "return home" to their party for Election Day), but also among Democrats (albeit to "undecided").
It's also worth noting that they did ask a question on the thesis. The result was basically that McCain voters got a null result (11% more likely, 11% less likely, 78% no impact) while Obama voters were less likely to vote for McDonnell (3%/56%/41%). Given the statistics about Obama voters noted above, this suggests a limited impact.
This also shows up in ideology and party ID questions (GOPers and Conservatives unaffected, Dems/Liberals less likely to vote McDonnell but not likely planning on doing so in the first place), though moderates and independents both showed signs of eroding support for McDonnell (note the shift among independents above). In the end, though, any changes between August and now seem to be down to noise in the makeup of the sample.
On to SUSA
SUSA's poll was very similar to PPP's in two regards:
- The last poll was pre-thesis.
- The lead was basically unchanged from the last poll.
The sample and the results were remarkably close to the last poll, back in August. I continue to say that I feel that SUSA includes a few too many McCain voters in their sample, but at least with some stability in the sample (remember, they were bouncing around a lot before), comparisons can be drawn.
Curiously, SUSA shows McDonnell losing Democrats but picking up:
Democrats, August: 19% McDonnell, 80% Deeds
Democrats, September: 13% McDonnell, 85% Deeds
Independents, August: 52% McDonnell, 41% Deeds
Independents, September: 59% McDonnell, 35% Deeds
I have absolutely no answer for why this happened.
Finally, Rasmussen
Those of you who have been following me over the last few months know the disdain I have for Rasmussen. However, his polls aren't wholly useless: They should at least be comparable to each other as far as methodology goes, and thus while they're probably off the mark in some regards, they're at least consistently off the mark in the same general way. I used the term "Reliable Invalidity" to describe this once before.
Because of this, I'll show the last few Rasmussen polls for Virginia (all scores McDonnell-Deeds):
6/10: 41-47 (D+6)
7/14: 44-41 (R+3)
8/10: 49-41 (R+8)
9/1: 51-42 (R+9)
9/16: 48-46 (R+2)
9/29: 51-42 (R+9)
The sum of this echoes what's shown up elsewhere: McDonnell has held a decent lead for most of the race. It has faded on two occasions: First, in the wake of Deeds' upset primary victory; and second, following the thesis breaking. After both incidents, the lead slid back to where it was before.
In Sum
The thesis seems to have done some short-term damage to McDonnell and hurt him among moderates and independents...but McDonnell made up for this through a solidification of his own base. The "damage" among liberals/Democrats was mostly illusory, as he didn't have a snowball's chance of winning most of those voters barring a Deeds meltdown. The race goes on, and Deeds continues to need another punch to have a good shot at this race. The thesis just didn't stick well enough to do what Deeds needs done.
A final note is that for the most part, the down-ballot races are moving in tandem with the up-ballot ones. However, SUSA may show some tightening in the AG's race. PPP also shows it. Some of this is Shannon's cash edge (let's face it, a 2:1 cash on hand edge will leave a mark) and some of it is the candidates. However, at least for the moment, Shannon outperforming Deeds by 2% or so doesn't look like it's going to cut it, while the fates of Wagner and Bolling are almost inextricably tied to Deeds and McDonnell, respectively.