(Credit: Rasmussen Reports, screenshot)
Some Mitt Romney supporters have been arguing recently that most of the polls showing Obama surging in swing states, and even nationally, are wrong, because the polls have too many Democrats in the samples. These “skewed” polls, the critics feel, are giving a misleading picture of the electorate, and if those results were weighted to reflect the likely voter model used by the Rasmussen poll, Romney would be the clear leader.
You can find a couple of rebuttals – on Fox News by Doug Schoen, and at Gallup by Frank Newport – but the more interesting, and surprising, part of this story is the presumption that the Rasmussen Poll is especially accurate.
A year ago, I awarded Scott Rasmussen the “TRY HARDER NEXT TIME” Award, as part of the 2011 Top Ten “Dubious Polling” Awards, for his horrendous performance in the 2010 election cycle. Here’s part of the explanation for giving him the award:
As Nate Silver wrote after the 2010 mid-term election, “Every election cycle has its winners and losers: not just among the candidates, but also the pollsters. This year, Rasmussen has the honors. In the final three weeks of the campaign, Rasmussen released the results of 105 polls, with an average error of 5.8 percentage points and an average bias in the Republican direction of 3.9 percentage points – far worse than other polls.”
In its most outlandish poll, taken in mid-October, Rasmussen showed incumbent Democrat Daniel K. Inouye leading his Republican opponent Cam Cavasso by just 13 points, though Inouye won* by a 53-point margin, “the largest error ever recorded in a general election in FiveThirtyEight’s database, which includes all polls conducted since 1998” (writes Nate Silver).
*the U.S. Senate race in Hawaii
To be fair to Rasmussen, let me add what I also wrote:
…he hasn’t always been this bad. According to Silver, “Rasmussen’s polls — after a poor debut in 2000 in which they picked the wrong winner in 7 key states in that year’s Presidential race — nevertheless had performed quite strongly in 2004 and 2006. And they were about average in 2008. But their polls were poor this year.”
What’s the lesson? Relying on any single poll is not recommended. And claiming that there is a conspiracy among pollsters to boost Obama’s chances of victory is simply absurd.
But, seriously, we knew that already. Didn’t we?
