[snarkytart]

Pew was actually tied with Rasmussen for most accurate poll in 2008. And tied with Battleground for most accurate in 2004.

http://www.fordham.edu/images/academics/graduate_schools/gsas/elections_and_campaign_/poll%20accuracy%20in%20the%202008%20presidential%20election.pdf

http://www.realclearpolitics.com/Commentary/blog_11_8_04_1018.html

And this report is saying Rasmussen badly missed in 2010 and was bias toward Republicans:

http://fivethirtyeight.blogs.nytimes.com/2010/11/04/rasmussen-polls-were-biased-and-inaccurate-quinnipiac-surveyusa-performed-strongly/

This is scaring me a little.

[Sooth2222]

"And this report is saying Rasmussen badly missed in 2010 and was bias toward Republicans"

2010 was an epic blowout for Republicans. As I read it -- in essence -- Nate Silver's criticism of Rasmussen in 2010 is that he predicted even more of a blowout for Republicans than actually happened. But that's not unusual. In 1996 Clinton/Dole, Gallup predicted 55% for Clinton but he actually won 49%. In 1964 Johnson/Goldwater, Gallup predicted 64% for Johnson and he actually won 61%. If the outcome is approaching certainty, some aren't going to bother to vote, and I suspect that's even more evident in off-year elections. Rasmussen predicted a Republican blowout in 2010 and there was a Republican blowout in 2010...

As I remember events in 2010, Nancy Pelosi was shocked when she learned that she was going to need to pack up her office and move. And she had access to the Democrat's internal polling. So if Rasmussen was inaccurate, so were a lot of other pollsters.

My own thought is that in 2012 many of the pollsters -- Gallup and Rasmussen are the main exceptions -- are "correcting" their results to match party ID in 2008, and they're doing it because "everybody else" is doing the same thing.