Jay Cost, Dick Morris, Unskewed Polls guy...

[_EAllusion]

Meanwhile, Sam Wang of PEC got the popular vote share exactly right. His electoral vote estimate was an artificial middle between 303 and 332 to reflect Florida being a near statistical tie, which also was dead on. So, on the other end of the spectrum, much props. Pollyvote, which I'm also a fan of, was off by .1%, which is catastrophic failure in the face of PEC's aggregation success.

If only we could've know beforehand who to trust: Scientists and paraprofessionals with a deep knowledge of statistics armed with the best poll reading aggregation and forecasting models political science can give us, or partisan hacks with an epic track-record of being wrong about things. If only there was some way to have known...

[_Kevin Graham]

I agree with everything you said EA but isn't this a very unusual election season for so many polls to be so accurate? It seems they are perfecting the science.

[_beastie]

It is truly satisfying to watch the toe-sucking fatso Dick Morris having to eat crow.

[_Kevin Graham]

I wish i could read ehat Scott Pierson is saying on Facebook. Last month he was telling us to mark his words Romney would win in a landslide.

But i bet in four years Dick Morris and Karl Rove will be at it again.

[_Res Ipsa]

I agree with everything you said EA but isn't this a very unusual election season for so many polls to be so accurate? It seems they are perfecting the science.

I don't know about the accuracy of individual polls, but the poll aggregators were very accurate.

[_Sethbag]

It is truly satisfying to watch the toe-sucking fatso Dick Morris having to eat crow.

I was somewhat impressed by how plainly Dick Morris came out and admitted that he was wrong. Eventually he spoiled it to a certain extent when he started with the excuses, but the first couple minutes of his "I was wrong" talk were pretty frank and honest.

What's funny is that I got from more than one side the whole idea that all the polls were wrong and that they were oversampling Democrats. I had a big argument with one Facebook friend over it on his wall, argued about it with my brother by email, and listened to it ad nauseum from some others I talk to or heard talking with each other at events I attended. My brother and this Facebook friend tried to convince me that Rassmussen was the only truly honest pollster out there, the most accurate, the least biased, etc. Turns out Rassmussen predicted a Romney win in the popular vote by 1%, and was off by 3% from what actually happened. Other than Gallup, I haven't yet seen anyone else do worse than that.

It will be interesting to me to see the detailed studies of the polls' performances this election cycle. I'd also be interested in knowing just what tripped up Rassmussen and the others who predicted Romney wins. They apparently over-sampled Republicans, but why? Did they just not believe the demographic changes that are working in the Democrats' favor? Are they seeing what they wish were the case, rather than what's really happening?

[_Res Ipsa]

I was somewhat impressed by how plainly Dick Morris came out and admitted that he was wrong. Eventually he spoiled it to a certain extent when he started with the excuses, but the first couple minutes of his "I was wrong" talk were pretty frank and honest.

What's funny is that I got from more than one side the whole idea that all the polls were wrong and that they were oversampling Democrats. I had a big argument with one Facebook friend over it on his wall, argued about it with my brother by email, and listened to it ad nauseum from some others I talk to or heard talking with each other at events I attended. My brother and this Facebook friend tried to convince me that Rassmussen was the only truly honest pollster out there, the most accurate, the least biased, etc. Turns out Rassmussen predicted a Romney win in the popular vote by 1%, and was off by 3% from what actually happened. Other than Gallup, I haven't yet seen anyone else do worse than that.

It will be interesting to me to see the detailed studies of the polls' performances this election cycle. I'd also be interested in knowing just what tripped up Rassmussen and the others who predicted Romney wins. They apparently over-sampled Republicans, but why? Did they just not believe the demographic changes that are working in the Democrats' favor? Are they seeing what they wish were the case, rather than what's really happening?

in my opinion, Rasmussen is one of the worst in terms of bias. He has a track record of reporting big R leads early in the polling, and then diving back toward reality at the very end. At this point, it's bad enough that it looks like an intentional effort to skew the poll aggregations. A good case can be made for eliminating his polls from the aggregations. (Silver handles it by weighting pollsters by their track records.)

I don't think the problem is oversampling anyone. I think it has to do with the likely voter screens that the pollsters are using. For example, I read about one company who asked "how likely are you to vote?" If you said "Certain," you were a likely voter. If you said "very likely," you were not. Gallup's Las Vegas screen is broken.

[_Brackite]

Here are a few Posts of mine about the Gallup Poll before the election:

The Gallup Poll among registered voters now has Romney at 48%, and it now has Obama at 48%.

http://www.gallup.com/poll/150743/Obama-Romney.aspx


The Gallup Poll among likely voters now has Romney at 51%, and it now has Obama at 46%.

http://www.gallup.com/poll/157817/elect ... omney.aspx


in my opinion, the Gallup Poll among registered voters is a little bit more accurate than the Gallup Poll among likely voters. I believe that Romney is up at about 2% over Obama Nationwide right now. However, Romney is still not winning in the Polls of the crucial swing States that he needs to win in order to win the Electoral college. It looks like now that it is probably likely that Romney will win the Popular vote while still losing the Electoral college.


http://www.mormondiscussions.com/phpBB3 ... 52#p650052

The Gallup Poll among registered voters now has Romney at 48%, and it now has Obama at 47%.

http://www.gallup.com/poll/150743/Obama-Romney.aspx


in my opinion, the Gallup Poll among registered voters is more accurate than the Gallup Poll among likely voters and the Rasmussen Poll. I think that Romney is up at about 2% over Obama Nationwide right now.


http://www.mormondiscussions.com/phpBB3 ... 18#p648918

Romney 49%, Obama 47% Among Likely Voters:
http://www.gallup.com/poll/157955/romne ... oters.aspx


http://www.mormondiscussions.com/phpBB3 ... 53#p644253

Gallup now has Obama at 48%, and it now has Romney at 46%.

http://www.gallup.com/poll/150743/Obama-Romney.aspx


Obama will get 47%, 48%, or 49% of the popular vote.
Romney will get 45%, 46%, or 47% of the popular vote.


http://www.mormondiscussions.com/phpBB3 ... 18#p638118

Gallup now has Obama at 48%, and it now has Romney at 45%.

http://www.gallup.com/poll/150743/Obama-Romney.aspx


Obama will get 47%, 48%, or 49% of the popular vote.
Romney will get 45%, 46%, or 47% of the popular vote.


http://www.mormondiscussions.com/phpBB3 ... 32#p635732

Gallup now has Obama at 49%, and it now has Romney at 44%.

http://www.gallup.com/poll/150743/Obama-Romney.aspx


Obama will get 47%, 48%, or 49% of the popular vote.
Romney will get 45%, 46%, or 47% of the popular vote.


http://www.mormondiscussions.com/phpBB3 ... 01#p635101

I Predicted here that Barack Obama would win the Popular vote before the first Presidential debate. Then I thought that Mitt Romney would probably win the Popular vote after the first Presidential debate. I thought also that the third Party Presidential candidates (especially Garry Johnson) would get more votes than they did.