A piece in Politico claims "Democrats are turning their fire on Scott Rasmussen, the prolific independent pollster whose surveys on elections, President Obama's popularity and a host of other issues are surfacing in the media with increasing frequency." In fact the piece quotes only one Democratic pollster who's critical of Rasmussen. The complaints seem to come from lefty bloggers, and not particularly serious ones at that.
By contrast, one who is relatively serious, Nate Silver of fivethirtyeight.com, comments on his blog that "Rasmussen's election polling has tended to be quite accurate in the past." The fact is that in 2008, when Obama won by 7, Rasmussen's final poll had him ahead by 6. For the final five weeks or so of the campaign, Rasmussen showed Obama with a stable and solid lead, with between 50 and 52 percent of the vote -- which I know pretty much matched the McCain and Obama campaign's private polling. Other public polls fluctuated wildly. Pew matched Rasmussen's final result, but they had Obama +15 a week before. Gallup's final poll had Obama ahead by 11, and the Washington Post had Obama by 9 (and by 11 one night before the final poll). So Rasmussen came through in 2008.
And he was on the money -- as the Politico piece acknowledges -- in New Jersey in 2009. What about 2010? Silver explains that "Rasmussen has a different model of what the 2010 election is going to look like, one which will feature a more conservative electorate. But that model isn't necessarily wrong, nor does it necessarily reflect bias."
That different model of the electorate is produced by Rasmussen's screening for likely voters when most public polls, at this stage, survey all adults or registered voters. But Rasmussen's methodology enabled him to be early -- and accurate -- in catching pro-Democratic trends in 2006 and 2008. I remember saying on Fox News Sunday in, I think, January of 2006, that it was quite possible the GOP would lose control of the House that year. Juan Williams thought I was joking, and Republicans were upset. I remember talking with Scott, and he agreed that the data were moving in that direction. Indeed, that year he was, I believe, the first pollster to show Republican incumbents like Conrad Burns (MT) and George Allen (VA) in trouble -- and he said on TV in September that the Democrats would win control of the Senate.
Generally, because Rasmussen has a likely voter universe and polls so much, he seems to catch trends earlier -- and other polls eventually move toward him. If you assume likely voters pay more attention to politics and tend to move first, paying attention to them will allow you to see trends early. That's certainly been the case on Obama's job approval, where Rasmussen saw the downturn before everyone else. Rasmussen still has Obama's approval about 5 percentage points lower than other surveys, and that's due to his universe consisting only of likely voters. And while it's legitimate to say that it's as useful to know the approval rating of the president among all Americans as among likely voters, if you're interested in the 2010 results, history would suggest the likely voter numbers are more likely to be helpful.
And that's why serious people in Washington pay attention to Rasmussen's polls.
