Remember folks, polls almost always tighten down the stretch.

[Bleacher Creature]

I don't blame anyone for getting nervous; the stakes are just too damn high. That said, people need to take a breath and remember that the polls almost always tighten the last couple of weeks in an election. It started before the Comey "bombshell" and will likely continue. It doesn't mean that the dynamics of the race have changed, and shouldn't dissuade anyone from continuing to fight!

For those who have forgotten, here's a gentle reminder that we've been here before (in 2008). I remember a similar freak out in 2012,and we all know how that turned out.

http://www.slate.com/articles/news_and_politics/explainer/2008/10/and_down_the_stretch_they_come_.html

[Farmgirl1961]

1. What I find interesting is that Daily Kos, Princeton Election Consortium and Huff Post

Show Hillary with 95% or great probability, but over at Nate 538, there's this downward drift. I know that folks have said to ignore 538 and that his polling has more "noise" and all of that, but it still makes little sense to me why there are such wide variances.

[Bleacher Creature]

2. Here's a good explanation from Sam Wang at PEC.

http://election.princeton.edu/2016/10/30/a-test-of-the-polarization-hypothesis/#comment-105703

[geek tragedy]

4. 538 takes the chance for systemic polling error more seriously than the others do nt

[Farmgirl1961]

3. Thank you

This helps reassure me, but still -- every time I go over to 538 it gives me the heeby-jibbies...I should just stop looking, but for some reason I click on the app from time to time and then start feeling a bit discouraged. Overall I'm quite encouraged, but feel like I'm on a ride with too many ups and downs.

[geek tragedy]

5. At least the stupid talk about landslide/blowout/400 EVs has gone away

People got overconfident in a big way.

This is an extremely close race--much closer than 2008 or 2012. More like 2004--a 2% race.

[Dawson Leery]

7. If it is 2 points, that speaks ill of this country.

[Va Lefty]

8. Very true

and I'm worried. Even if Clinton wins these assholes aren't going away.

[geek tragedy]

9. Indeed. For all the talk about how the GOP is about to blow itself up,

they will be in a pretty good chance to retake the White House in 2020.

The Democratic nominee getting ~290 EVs against a burning cross made of dog shit is really not a big show of strength on our part, either as a nation or as a party.

[Drunken Irishman]

6. 2012 actually broke late for Obama.

In the final week, he surged ahead in many national polls that either had the race tied or him down marginally.

Where it compares is the EC. Hillary, like Obama, had a decidedly strong lead there. Problem is, Hillary is probably not as much of a lock in two states Obama was not going to lose - Iowa & Ohio. Trump is probably going to win both. However, at this point four years ago, the media was writing Obama off in Florida and North Carolina, two states that are very favorable to Hillary.

[geek tragedy]

10. Florida isn't very favorable for Hillary.

Polls are not good, and the early voting numbers there are significantly worse than in 2012.

The big problem we're having is that African-American turnout is way down compared to 2012, not to mention 2008. Obama was a bigger inspiration to vote in a positive way than Trump is in a negative way.

NC is looking decent per the polls, but again early vote is not as good as it was in 2012.

[Drunken Irishman]

11. Her Florida polling numbers are better than Obama's.

I know what you're trying to do and it's extrapolate early voting. That's not very smart and I'm not going to take your view seriously on this if that's all you've got.

Fact is: In 2012, Obama trailed Romney by 1.5 points in the final average of polls. Currently, Hillary trails by just .5.

Nate Silver currently has Florida a very slight lean for Hillary.

The Princeton Election Consortium does as well.

[geek tragedy]

12. Early voting data is the best measure of actual turnout rather than

the hypothetical turnout in polls.

Florida is won and lost based on who turns out. Obama won Florida in 2012 because of his turnout.

Early vote in 2016 is more Republican, older, and whiter than it was at this point in 2012, despite the state's voter rolls becoming less white.

The danger for Clinton is that our base just doesn't turn out to vote for her. And that appears to be happening in Florida.

That's a red flag. It's a sign that the electorate really will look more like 2004 than 2008 or 2012.

[Drunken Irishman]

13. Not really.

No offense, and I respect your view, but I think Sam Wong and Nate Silver are probably looking at the same data as you are and they've concluded it's still a lean Hillary. I respect their opinion more, especially Wong's, than my own or some random person on the message board.

Is Florida a lock for Hillary? No. But it's still a very favorable state for her and she's out-polling Obama there.

[geek tragedy]

14. Neither Wang nor Silver's models are looking at early voting data. They look at polls

The only extent to which they look at early voting data is indirect--early voters will get picked up in most(!) likely voter screens for polls.

Recall Silver's big miss of the primary season--Michigan. Why was his model so wrong? Because all of the polling was really wrong.

Why was all the polling so bad? Because the pollsters all screwed up the turnout model.

[Drunken Irishman]

15. Primaries are not general elections. I'm sure you know this.

Primary polling has been unreliable for years and years. Case in point: New Hampshire in 2008. Every poll had Obama winning comfortably. Hillary won, tho. Didn't mean the general election polls were wildly off.

Florida will be close but Hillary should be considered the favorite there.

[geek tragedy]

16. I'm not saying Clinton won't win FL, just sayin' the data isn't necessarily pointing that direction

It's a state she SHOULD win (+6% more nonwhite than national electorate, trending nonwhite) but that somehow isn't translating into polls or early votes.

[Drunken Irishman]

17. You're using faulty logic to make a point that just isn't there.

There's a reason Wang and 538 aren't putting much emphasis on early vote totals.