The Nielson Brothers are out with a new poll with some… interesting results.
Here’s their headline figures for South Dakota registered voters (not likely voters):
Mitt Romney 49, Barack Obama 43
Kristi Noem 47, Matt Varilek 46
Kristie Fiegen 43, Matt McGovern 41
Chris Nelson 54, Nick Nemec 30
How accurate are these numbers? They’re very favorable to the Democrats. Set aside the absolute numbers, and the relative numbers seem about right with conventional wisdom: Varilek outperforming Obama (which he may or may not do, but has to if he’s going to win), McGovern outperforming Nemec.
Rejecting a poll just because its numbers don’t look right is always risky. Sometimes a poll that looks wrong is wrong. Sometimes preconceptions are wrong. The best way to judge is by comparing to other polls. Unfortunately, we don’t have any comparable matchup polls for South Dakota. What we do have is the Gallup poll from a little bit ago, showing 38 percent of South Dakotans approve of Barack Obama.
That asked about approval, not who people would vote for. How closely do approval numbers track vote preference?
Pretty closely, it seems. Here’s a few national polls with Obama’s approval and support:
Rasmussen, 44% approve, 43% support
Pew, 51% approve, 51% support
Democracy Corps, 50% approve, 50% support
So let’s say, in the absence of more numbers from the Nielson Brothers, that approval and support are equivalent. That means their poll is friendlier to Obama by five points than Gallup. (We shouldn’t necessarily presume that Gallup is the more accurate of the two, but for the sake of this exercise I will.)
If that lean is true across the board, then Varilek’s at 41 percent to Noem’s 52, McGovern’s at 36 percent to Fiegen’s 48 percent, and Nemec is at 25 percent to Nelson’s 59 percent.
This is just back-of-the-envelope stuff, so don’t take this too far.
Regardless, being at 41 percent and around 10 points down on Noem at this point in the campaign is actually a pretty decent result for Varilek. That doesn’t mean he can close, but it means he’s within striking distance.
Who knows. We’ve got one data point. I hope Rasmussen or one of the other polling agencies releases some numbers to give us more context.
If you’ve got a lot of time on your hands and want to read more about Nielson Brothers Polling, I wrote a very long blog post on them at Mount Blogmore some time ago.
UPDATE: A little more context for people who don’t have enough time to read that linked post. NBP has a relatively short track record as a public pollster. Some of their polls have been pretty close to actual results (such as the 2010 U.S. House race in South Dakota), while others have been way off, including a wacky poll in the 2010 governors’ race. Some Republicans point to Paul Nielson’s history as a Democratic candidate as proof the firm is biased. The Nielson brothers insist they provide accurate results and point out that the other Nielson brother, Mark, is an economist who “leans to the right.”