Over the weekend, the Washington Post turned some heads when they showed Ken Cuccinelli with a 10-point lead among “certain” voters (and a 5-point lead among all registered voters). Today, NBC/Marist shows Cuccinelli ahead of Terry McAuliffe 45%-42% among likely voters.
Before we get into the differences between the two surveys, some other good news from NBC/Marist:
— Bob McDonnell’s approval rating is high, 61%/24%.
— Ken Cuccinelli is viewed favorably, 42%/27%, while Terry McAuliffe’s image is more modest, 32%/24%.
— More Virginian voters believe Cuccinelli’s ideology is about right (39%) than McAuliffe’s (32%).
— More Virginian voters think McAuliffe is too liberal (28%) than Cuccinelli is too conservative (27%)
Besides the ballot, image, and ideology, Cuccinelli also leads McAuliffe on most of these statements:
“Which candidate…
…better undestands the problems of people like yourself?” (34%-30 Cuccinelli)
…do you trust more to do what’s best for Virginia?” (39%-33% Cuccinelli)
…is closer to your position on social issues such as abortion?” (33%-31% Cuccinelli)
…cares more about the middle class?” (30%-31% McAuliffe)
…shares your values?” (35%-29% Cuccinelli)
Now for the difference between the surveys.
(For the record, if I had to choose between NBC/Marist and Washington Post, I’d go with the Post. A few days before the election, NBC/Marist released their last survey of the Presidential race in Virginia. Like most other public pollsters, they had Virginia as a virtual tie, a 48%-47% Obama lead. Two WEEKS before the election, the Washington Post showed Obama with a 51%-47% lead, which ended up being the exact result on Election Day.)
So what would account for the 7-point difference between two live interview surveys conducted at roughly the same time? I spot three major differences:
1) NBC/Marist uses a more traditional likely voter screen, while the Post showed results by those “absolutely certain to vote”. A stricter screen will give more weight to the most reliable voters, who tend to be older and whiter.
2) Whether due to item #1 alone or if other factors were involved, NBC/Marist’s party breakdown is 27% GOP/31% DEM/40% IND, compared to the Post’s more balanced 26% GOP/28% DEM/33% IND.
3) Unlike all the other live interview surveys released thus far in Virginia, more voters in the NBC/Marist survey say they have heard of McAuliffe and Cuccinelli and have opinions about them.
These differences account for the changes; however, there are more significant problems with the NBC/Marist survey that give me pause. They haven’t released their crosstabs (unlike the Post), but they make several references, and I want to highlight three demographics that I think are unrealistic, or at the very least less likely, compared to the Post’s survey:
Women: We’re all familiar with the gender gap. Women tend to vote more for Democrats and men tend to vote more for Republicans. In 2012, Obama won women in Virginia 54%-45% (-9), while Romney won men 51%-47% (+4), creating a gender gap of 13 points. The Post survey (looking at registered voters) showed Cuccinelli leading men by 10-points (50%-40%) and losing women by 1-point (40%-41%), for a gender gap of 11-points. This is roughly in line with the 2012 exit polls, but a bit more favorable to the GOP, which is understandable given that the GOP candidate is leading.
In the NBC/Marist survey, Cuccinelli is losing women huge, 36%-50%, but he’s winning men by a similar margin, 49%-36%. That creates a gender gap of 31-points, which is simply massive. The Post survey is much more in line with what a reasonable observer would expect.
Seniors: I’ll note up front that NBC/Marist has an age group for 6o+, while the Post and the exit polls have the group at 65+. However, I don’t think those five years would account for the differences. In the NBC/Marist survey, seniors are “virtually tied” (without giving numbers), while Cuccinelli leads seniors by 12-points (47%-35%). This is much more in line with conventional wisdom as well as the 2012 exit polls, where Mitt Romney won seniors in Virginia by 8-points (54%-46%).
Independents: There are multiple ways to determine an “Independent”, and I think that may account for the differences here: The Post asks R, D, or I straight up. NBC/Marist has different categories for “Strong Democrat”, “Not-so-strong Democrat”, “Democrat leaning Independent”, and “Just Independent”. The D-leaning and R-leaning Independents are all lumped into the broader “Independent” category, which is why it’s a huge 40%, even though self-ID’d Independents only made up 29% of the vote in 2012. Giving these more nuanced options to voters is generally a good thing, even if it inflates the # of Independents. However, D-leaning Independents and R-leaning Independents were both an identical 14%, so this method didn’t tip the balance one way or another on the surface.
However, in the NBC/Marist survey, Independents are tied at 36% each. In the Post survey, Independents favor Cuccinelli by 5-points (42%-37%). In 2012, Romney won Independents by a full 11-points (54%-43%).
NBC/Marist is a decent polling outfit, and their numbers are generally reliable. However, when compared to the Washington Post, I trust the Post’s numbers more, due to their recent track record in Virginia, the fact that they fully released their crosstabs, and because their results among women, seniors, and Independents are much more in line with what we saw in 2012. Either way, both surveys confirm what should be clear to everyone by now: 1) Voters still have a lot to learn about both candidates, and 2) despite what Democrats think, most Virginians like and trust Ken Cuccinelli and don’t think he’s extreme. Based on what we’ve seen from the other side thus far, they’re following the same blueprints that led to their 2009 disaster.
