Unlikely Voter

I’m not the only one who’s been skeptical of Nate Silver’s dogged attacks on Rasmussen Reports, demanding that they open up their secret sauce just because the firm’s results are less favorable to his preferred political party than other pollsters’ results are.

Mark Blumenthal is on the case, esposing what is fundamentally a hypocritical position by Silver and his site, fivethirtyeight.com.

Daily Kos has dropped Research 2000 as the site pollster going forward, citing Nate Silver’s poor marks for the firm’s results. I’m glad of this.

Today we have high profile races in South Carolina, Nevada, California, and Arkansas to look back on and see how well the pollsters’ numbers matched the actual results.

As I’m just one person who lacks the luxury of doing this site full time, I’m not ashamed to work off of Real Clear Politics to make these assessments.

Today’s day off is not free. It was paid for in advance.

Having attended a speech of Nikki Haley’s in Atlanta last year, her run for Governor of South Carolina is one I’ve followed. Living here, I’ve also watched the California Senate race. Both situations seemed to be stable: Haley was stuck in fourth, while in California Tom Campbell was staying ahead of second place Carly Fiorina.

Then Sarah Palin intervened, and both Haley and Fiorina shot into first place in new polls. Now we have confirmation of both events. Sarah Palin carries respect in the Republican Party.

Research 2000 very much wants you to know that they do polls for newspapers in St. Louis, Lexington, Fort Wayne, South Bend, and Reno. They don’t want you to think of them primarily as doing polls for left-wing activist outlets Daily Kos and now Democracy for America, the group founded to continue the work of former Presidential candidate Howard Dean.

Someone should tell Daily Kos and Nate Silver this though, as they beat the wardrums against R2k competitor Rasmussen Reports, that R2k is trying to look less partisan, not more.

Per Real Clear Politics there were six pollsters who took on the Pennsylvania primary, which Joe Sestak won handily by 8, 54-46 over Arlen Specter.

Let’s see who got it right, and also give credit to a particular pollster where it is due.

Many pollsters rely on random telephone calls in order to get their interviews done. It’s cheaper and easier than having to get people to sit down and do face-to-face interviews, because over the phone a broader range of people can be accessed. The result is a better sample than can be gotten from, say, a shopping mall.

But that ease may be on the way out if people stop owning landlines.

Some researchers led by Noah Smith at Carnegie Mellon tried an experiment: Could they predict the results of traditional polling, which has as its core feature a genuine random sample of people, with careful monitoring of Twitter?

Tommy Thompson was going to give Russ Feingold a tough race, but he decided not to run. How are Republicans doing now to challenge the three term incumbent? Rasmussen has a look.

One race: North Carolina Senate general between Republican Richard Burr and prospective Democrats Cal Cunningham and Elaine Marshall. Two polls: Public Policy Polling and Rasmussen Reports. Two markedly different results: Rasmussen shows Burr nearly 10 points higher than PPP does.

What’s going on?

When I see people, and particularly conservatives, discussing why one poll is better than another, I see the sample pool frequently cited as a reason for favoring one or another. Specifically, some poll watchers insist that any poll not filtered for likely voters, instead of just registered voters or even adults, is not useful in a political context.

The dirty secret is that not all likely voters are created equal. Every pollster has his own secret sauce, and we have to be careful when trusting that kind of filtering. It might not be what we expect.

In looking at recent polling, I said I needed more data before I could write off Quinnipiac’s results as meaningless. Real Clear Politics has much more data to look at than I have, thanks to having been doing this for much longer, and they see a trend in Quinnipiac’s Ohio polling that leans toward the Democrats in every race in that state.

Food for thought, to be sure.