Nate Silver points out that Scott Rasmussen, the conservative activist and pollster, produced highly inaccurate polls that were strongly biased toward the GOP:
I did a quick check on the accuracy of polls from the firm Rasmussen Reports, which came under heavy criticism this year — including from FiveThirtyEight — because its polls showed a strong lean toward Republican candidates.
Indeed, Rasmussen polls quite consistently turned out to overstate the standing of Republicans tonight. Of the roughly 100 polls released by Rasmussen or its subsidiary Pulse Opinion Research in the final 21 days of the campaign, roughly 70 to 75 percent overestimated the performance of Republican candidates, and on average they were biased against Democrats by 3 to 4 points.
Rasmussen's reply is hilarious:
Asked about Silver's assessment, Scott Rasmussen told Salon today that "I don't respond to comments from bloggers or others." And he said he was on the whole proud of his midterm predictions.
"We showed the generic ballot numbers predicting a significant victory for Republicans all through 2010, and that worked out as we expected," Rasmussen said. Noting polls the firm did on Arlen Specter in Pennsylvania and Scott Brown in Massachusetts, he added: "We were the first to pick up the power of what later became the Tea Party movement."
He said he did not specifically remember the Hawaii poll showing Inouye ahead by just 13 points. "I'll have to go back and look at that one," Rasmussen said.
So first Rasmussen says he won't reply to "bloggers or others," following the convention of attempting to discredit your critics by noting that they publish their work online. Then he proceeds to reply, but in a way that consists of cherry-picking data points of success rather than addressing the overall inaccuracy of his work. First attack ad hominem, then sidestep the substance. Nicely done.