Most Recent Times Rasmussen has been wrong

13 thoughts on “Most Recent Times Rasmussen has been wrong”

1. Rasmussen called Ohio a tie in their last Pres-2008 poll, and Obama won by 4.  They had Pennsylvania closer than it really was, and they had Nevada a 4-point Obama advantage when he won by 12.

   But what all this really proves is not that Rasmussen is “bad,” but rather that it’s par for the course compared to other pollsters, correctly predicting at least what races are close and which ones are not, and pegging the wrong winner as often as other pollsters in the close ones.

   But those are just the close-to-election polls, when Rasmussen mysteriously conforms to everyone else.  That doesn’t explain Rasmussen’s extreme outliers on issue polls and on far-from-election-day horse race polls.

2. I have never understood the rationale for polls which are skewed in one direction or another.  Isn’t it always best for a candidates campaign to know where they truly stand in a contest?

   I understand the idea of wanting to have polls show that a candidate is closing on the leader but not this.

3. and its lack of neutrality…

   However, I don’t think a few cherry-picked polls is enough substantive evidence, especially when most of the polls you cite have each candidate within 3 or 4 points of the final result.

   Yes, while the toplines of the polls and the final results show different winners, (i.e. McCaskill performing 2 points worse than actual, and those 2 points being the difference between a win and a loss), statistically, that’s no different than say, a poll showing Obama at 59% in California instead of his actual 61%.

   Also by definition, on average, one out of every 20 polls will be completely off and further outside the MoE.

Comments are closed.