Our friends at Republican Reports --oops, I meant Rasmussen Reports -- have made some news in Wisconsin this week with two polls showing Republicans leading in the races for governor and US Senate.
You'll have to excuse me if I'm a little skeptical.
Rasmussen Reports is one of those firms that does polling without actually speaking to any voters. It's all automated. That methodology is questionable enough that a number of media outlets won't even report it. The Wisconsin Associated Press is among them.
The firm tries to establish its bona fides by citing past Wisconsin polls where it correctly predicted the winner. As Brew City Brawler pointed out, even when getting the result right Rasmussen consistently underestinated the democratic candidate's vote. It also has a history of showing Republicans doing surprisingly well early, then producing some more realistic numbers closer to election day, perhaps to maintain some credibility.
Rasmussen is clearly doing the poll for political purposes instead of just to report current public opinion. Otherwise, why would you test a Senate matchup with a candidate who is not running (Tommy Thompson) but not test the frontrunner for the GOP nomination who is a declared candidate (Terrence Wall)? Michael Mathias asks.
The numbers they report don't make sense for anyone who has followed Wisconsin politics. They claim that 85% of statewide voters know Barrett well enough to rate him, 83% know Walker and 81% know Neumann. None of the three candidates are that well known according to other polls. In fact, about half of the respondents say they haven't haerd of Walker or don't know enough to rate him as favorable or unfavorable. (Ditto for Neumann and Barrett). It is very difficult to build statewide name recognition without paid television, which no one will be doing for awhile. Those numbers are just bogus.
Walker's purported popularity numbers (56% to 27% with 29% very favorable) would make him one of the most popular people statewide in WI. That is hard to believe for someone who has only run for election in Milwaukee County. There is no evidence from other polls that he is either that well known or that well-liked.
Unfortunately, many media outlets continue to report every poll as though they all are equally valid.
One last note: When the margin of error on a poll is plus or minue 4.5%, it means each number can be off by that much. So Rasmussen's amazing and unbelievable 48-38 lead for Walker could be 43.5 to 42.5 (close to what other polls say), and the 47-43 Tommy lead over Feingold for the Senate could be a 47.5-42.5 lead for Feingold. Probably not, but in one of 20 cases it's possible teh results could be even farther off.
Finally, did anybody say who's paying for all of these polls? Do we think they're a public service? Or do we think that some Republican money may be paying an admittedly GOP-leaning firm?