Skip to main content
ABC News
Wisconsin Senate: Feingold 46%, Smithee (R) 44%

This morning, I posted an article that questioned whether polls — and in particular, the polls released by Rasmussen Reports — were disproportionately representative of a very small segment of very high-information voters: people whom we might think of as political junkies, or even political activists.

This afternoon, we got an interesting example of that. Rasmussen released a poll suggesting that Ron Johnson, a businessman from Oshkosh, is running just 2 points behind Democratic incumbent Russ Feingold in Wisconsin’s Senate race.

Have you ever heard of Ron Johnson? I hadn’t until about three days ago. And it seems unlikely that many voters in the Badger State would have heard about him either. Johnson didn’t announce his candidacy until two weeks ago. He’s never held political office before. He runs a small business called Pacur Inc that has only 73 employees. He doesn’t have a campaign website up. He doesn’t have a Wikipedia page. Prior to this morning, he’d been mentioned in the Milwaukee Journal-Sentinel, the state’s largest paper, just three times over the past month.

And yet, according to Rasmussen, 68 percent of Wisconsin’s “likely voters” have already formulated an opinion about Ron Johnson! And 94 percent of the state’s voters know how they’d vote if he were to run against Russ Feingold!

Now, it’s possible that this is an artifact of Rasmussen’s methodology rather than something that should be regarded as literally being true. For example, if their question wording discourages people from pressing the “don’t know” button, people might be formulating an “opinion” about Johnson based solely on the fact that he’s a Republican. Or perhaps they just like pressing buttons more or less at random. If you used Rasmussen’s methodology and identified a fictitious candidate named Alan Smithee (but identified him as a Republican who was running for Senate), perhaps he’d wind up with 68 percent “name recognition” as well.

But if we indeed take this result literally, it is pretty ridiculous. Johnson is good on the stump and has started to generate a lot of buzz among the conservative and tea-party intelligentsia. He has some upside potential — and may well defeat Feingold in the end. But common sense would dictate the number of Wisconsinites (even “likely voters”) who know who Ron Johnson is right now is closer to 6.8 percent than 68.

I’ve had a lot of criticisms for Rasmussen Reports over the past couple of weeks. Many of the problems their polls exhibit are common to other pollsters throughout the industry, if they seem to be a bit exacerbated in Rasmussen’s case.

But the fact is that Rasmussen is the 500-pound gorilla in the room. They drive a lot of traffic and narrative. They account for something like 30 percent of all horse-race polls that have been released thus far this election cycle.

And their polls really stick out, with results that are as often as not vastly different than what other pollsters show, almost always being more favorably disposed toward conservative candidates and causes. A lot of what I do — especially when we’re still pretty far removed from November and when nobody should sweat any individual poll all that much — is to highlight polls that seem weird or misguided. If Rasmussen accounts for 30 percent of all polls, they probably account for an outright majority of weird polls. Even if Rasmussen is right and this winds up being a disastrous cycle for Democrats (indeed, I remain fairly bullish on the Republicans’ prospects right now), this is not a particularly healthy state of affairs for the industry.

Nate Silver founded and was the editor in chief of FiveThirtyEight.