What is more predictive of final vote in swing states, Early Voting or Polls?

[Erich Maria Remarque]

Absentee/early vote thread is very popular and many of you seems to pay a lot of attention to it. It has it's own heroes, most popular, probably, the guru of Nevada, Jon Ralston.

He bases his analyes on early voting patterns (in particular Clark?) and as I understand has some insider information from data polling firms [about demographics?].

But he stated himself a couple of times that his prediction holds if the following assumptions holds

a) the pattern on the election day ~about the same as 2012
b) there has not been a realignment of voters (crossovers)

I am intrested in the latter b) realignment. It is not necessary about Nevada, though NV is a good example.

I think it is a pretty strong assumption. I mean, we know that there was a realignment.

Two the most obvious crossovers:

Rep  white  educated women   --> vote D
Dem white non-educated men --> vote R

So leading in county X, would not explain a whole thing. One has to know education/gener etc of voters.

A good example about it is Upshot/Sienna's poll on Florida. http://www.nytimes.com/interactive/2016/10/30/upshot/florida-poll.html

You must be logged in to read this quote.

Clinton 55 vs 32 among white registered Democrats without a degree. Is it not a real realignment?

You may say: They have always voted R!
Common, do you really believe that 47% Romney would get the same numbers among white working class? Well, if you do, then this election is over. Congratulations!

You may then say: This poll is an outlier! Other polls would show a similar pattern, rrright?
Well, Upshot used voter file, and the 99% of polls don't. They just ask party identification and Trumpistas are plausible are more likely to say that they are Republicans OR Independents than Democrats. And yeah, 1% do ask. Among others "A+" Selzer (that by the way has consistently been Trump friendly, I wounder why) and TargetSmart (so those who likes TargetSmart more the Selzer: Congratulations! Hillary will win ). Probably, there are other polls that do it, but I am not aware of such.

And more about Selzer and Upshot:

You must be logged in to read this quote.

And of course BREXIT/Shy Trumpistas Meme!
https://morningconsult.com/wp-content/uploads/2016/11/MCModeStudyFinalToplines.pdf
https://morningconsult.com/2016/11/03/shy-trump-social-desirability-undercover-voter-study/

[eric82oslo]

Early voting is showing that women in particular and even more so, Hispanic voters have never ever been more highly energized in any single presidential since the very first one in 1789 when George Washington won with 100% of the vote and neither women nor Hispanics were allowed to vote. Though even if they had been allowed to vote, they wouldn't have been more energized.

[Erich Maria Remarque]

Nailed it

[SirMuxALot]

Nailed it

Yeah, I didn't answer, so that counts for "none of the above", right?