Xoff, a contributor at Uppity Wisconsin, recently voiced some beef he had with the results of a Badger Poll which put Russ Feingold six points ahead of Ron Johnson, 25-19.
What's missing here is any kind of reality check. How could someone release a survey which is so far off from everything else that's been done? Are people supposed to believe this is right, and all of the others are wrong? The odds against that being true are a lot higher than any margin of error in this poll.
The difference is the methodology. In an email, Prof. Kathy Cramer-Walsh offered an explanation for why Badger Poll yields different results than other pollsters.
Badger Poll spends a lot of resources on getting a high response rate. The UW Survey Center puts its interviewers through intensive training so that we can do our best to convince people to participate (what we call "refusal conversions"), and we call each number back many times (10 times) to try to reach people who don't answer. We send advance letters to people, too, as an additional way to try to increase response rates.
All of this is to say that the Badger Poll likely gets a more representative sample than does a poll like Rasmussen that has a much shorter field period. The people you don't catch in a three day poll are quite likely different than the ones you do in ways that matter for political opinions.
In addition, Badger Poll (see page 2) does not ask the people it surveys a follow-up question that forces respondents to choose between the candidates.
Are the majority of voters so undecided that they'll abstain when they go to the voting booth? Of course not. However, a casual party preference or vague admiration for the incumbent could realistically put a respondent in the "not sure" column four months before the election.
In this sense, Badger Poll may not offer a very useful prediction of the election results, but it does offer a more realistic glimpse at actual public opinion than the bigger polls.
As for the Feingold race...I continue to have mixed opinions on how competitive the race will actually be. It's hard for me to believe that Tea Party-infused anger over big government and bailout will be the doom of a senator who easily won re-election as an anti-Patriot Act dove at the height (ok, maybe a little after the height) of the War on Terror. He seems to have played his cards right politically.
Nevertheless, I get mixed reports from politicos on the race. Some are incredulous that I would ask if Johnson has a chance, and others say this will be the closest race of Feingold's career.