For the second straight election, Pew Research's final poll was EXACTLY correct

[scheming daemons]

2004: 51-48 Bush

2008: 52-46 Obama

http://people-press.org/report/468/obama-leads-mccain-in-final-days


In 2000, they had it Bush-49, Gore-47.... and it ended up Gore-49, Bush-48.... Still pretty damn close.



Next time around.... maybe we should all pay closer attention to Pew.


Also.... Scott Rasmussen called it 52-46 this time also. Dead on. Kudos to him too.

[bullimiami]

1. you think they are taking into account the likely 6 million votes wiped out

by fraud, purges and outright thievery?

or did they just get lucky?

just like the weatherman and the economist. some will invariably hit it on the nose and they will be the new authority on how its done.

[scheming daemons]

2. Well... they've hit two in a row on the nose..... that's better than most weathermen

And I would say that.... being that the average of the final polls was 7.8 (according to RCP), and Obama won by a little more than 6......

That there wasn't "6 million votes wiped out".

The polls were off of Obama's win by about 1.5.... or less than 2 million votes.

[unpossibles]

3. I don't know...

I think when you look at the polls which included cell phone users, the margin went considerably up. And looking at the efforts which did get blocked as far as purging rolls, I bet they still purged a lot and stole more than a few, just not enough.

Either way, we won. We still need to address the problems in our election system.

[Awsi Dooger]

4. Their previous poll was weird, well into double digit margin

Then somehow they scraped together the correct final margin. I respect PEW tremendously but that one bizarre poll makes me skeptical of their effort this year.

This cycle wasn't particularly hard to estimate. My final Excel model late on election eve had it Obama 51.9, McCain 46.2, for a net of +5.7. I think it sits at +6.1 right now, which is exactly where Nate Silver had it. He obviously has a great system so I'm going to borrow (steal) some of his ideas. :)

The polling companies that need to be scoffed at the ones that wound up with a margin of 8+. There was never any threat of that, not that I ever saw in the logical models. I tried one night to get it there in my Excel model and I couldn't even approximate it, not without tipping white women toward Obama, which made no sense. I think he lost that block 53-46 according to the national exit poll.