Excellent, 2andy4 – thanks. I say this puts PA into the rock-solid part of Hillary’s firewall, along with Virginia, Michigan, and especially Wisconsin.
That just leaves NH, which we’ve been discussing…but also Colorado. Any recent updates on the trend in CO?
On the downside, ABC, which used to have Clinton up by double digits not that long ago, puts Trump ahead in a 4-way for the first time.
And I know 538 thinks that pollster is good, but they are showing McGinty+12, WAY outside the polling average. Is there any reason whatsoever to think that Pennsylvania has shifted that far blue, just now?
Clinton+2.5 is the number that holds the most significance today, IMO. Down from +2.9
538 polls only has Hillary up 4.7 right now (bigger than their estimate in 2012, I believe). They’ve got a better track record than RCP (if that’s where your number comes from).
A quite recent poll from a B-rated (by 538) pollster has Hillary up by 3 in CO. Her trend there has paralleled the full models – steadily downward in recent weeks, but hasn’t dipped below 70% (yet).
I’d keep an eye out for the next highly-rated CO poll. (And NH…and, as adaher mentioned, the next one from PA as well, in case that recent poll was an aberration of some sort).
I think +4.7 is quite reasonable, as an extrapolation, but there’s no actual polling average that shows that, and 538 admits uncertainty, which is why her chances are 75-80% despite that lead.
Right. If 3rd parties plus undecideds were at 2012 levels, she’d be at 90+% based on polling. I’m hopeful that the turnout/ground-game advantage makes up for that (and then some, hopefully).
As I predicted on the predictions thread, Trump would be essentially tied in the polls by midweek and some credible polls would show Trump with a lead. LA Times and Rasmussen are one thing but the ABC - Post poll seems to be in the relative middle ground in terms of polling IIRC.
I am now officially worried, although to be honest, I already knew that the best we could hope for after this scandal broke was a badly wounded Clinton as president-elect. Still, I’d be willing to take that over President Trump.
But as I predicted, this latest news is damaging not because it makes Trump stronger but it makes Hillary weaker. It’s a vote suppressor. And Hillary doesn’t have the kind of breathing room to have dampened enthusiasm.
“This scandal”? Which one would that be? All we have right now is that some e-mails might or might not exist, which might or might not have been sent to or from Clinton, and which might or might not have any problematic content.
Well that and a bunch of headlines/talking heads/people are saying/I’m hearing … about SCANDAL!
I agree with you that there is likely nothing there*, and certainly nothing solid at this point, but that’s between you and I in fact based reality. If they keep talking about it as a scandal, then a scandal it IS for the purposes of these weirdo “undecideds”.
Count me worried. But I have expected President Trump since way back in the primaries. In some abstract way, I accepted something like President Trump would happen ever since GWB was reelected.
*“nothing” outside the general scope of what they found with the rest of the emails they’ve been through.
NYTimes Upshot blog has a great experiment looking at the early vote numbers in NC – they’re combining early voting data with their polling from NC to see how it tracks. This includes poll results from people who have already voted (I presume they ask on the phone if they’ve voted already). So far, even with lower black turnout than in 2012, Hillary has a solid estimated lead of 5-6 points. This is the first time they’re doing this, I believe, but it’s a positive sign using actual voting data. If she wins NC, there’s pretty much no path for Trump. We might know by 7:30 on voting day, when NC polls close.
Other polls (even post-Comey announcement) have Clinton anywhere from -1 to +6 - and some indicate that she’s completely putting Pennsylvania away.
I’m not entirely sanguine and I’m not a stats guy, but what this suggests to me is that no one really knows how to rate likely voters, so take the aggregate, keep calm, and just stay drunk for the next week or so. If Sam Wang is to be believed, then no one is going to switch from Clinton to Trump or vice versa, it’s just a question of who gets their voters out - and she’s the one with the apparatus to do that.
North Carolina has a lot of white college educated voters, and losing those voters might be more damaging for Trump than losing some black voters is for Clinton. Especially since those black voters aren’t votes for Trump, they are votes for no one, whereas the white college educated voters are going for Clinton.
NC is also held back by a Republican government that overstepped. There’s been a lot of that going around the last few years, and wherever it happens there’s a price to be paid. NC Republicans are paying that price.
That might offer an explanation as to why Trump is campaigning in Michigan. The article I read about it suggested it’s an insane idea, but I wasn’t so sure.
Trump’s path to victory is obviously very difficult, but everyone assumes it amounts to “win every battleground state.” But Trump and his campaign may already recognize that it is unlikely he can do that, and conclude he actually has a better chance of cracking into the Democratic Firewall by courting angry white guys in states like Michigan. Rather than the traditional “just win all the battleground states” perhaps he’s throwing a Hail Mary at the “win the Rust Belt” target, to try to create a map like this:
If you switch PA for MI he has to pick up something else but you get the idea. This?
I’m just playing around, but the general idea is that if Trump recognizes he cannot win every battleground, he would be well advised to try to change where the battleground is.
Not that Trump has any idea what he’s doing, but it probably doesn’t matter. If Trump wins the popular vote by 2 points or more, he wins the electoral vote. The map only comes into play if the race is tied or within a point. I think Trump just wants to maximize turnout among his supporters, wherever they are, win the popular vote, and count on the electoral college taking care of itself.
Not to scare you more, but early voters aren’t well correlated with the final results. Fivethirtyeight found a very weak correlation, and if you remove the outlier (MD), there’s essentially no correlation at all. At least in the past, the polls have been much better at predicting the final result than early voting patterns.
Then it’s smart that 538 doesn’t adjust for early voting, but still, a lot of voting has taken place while Clinton led by 6.
That’s not what upshot is doing – they’re taking a very close and detailed look at the early vote numbers and breakdown by party registration, and comparing it to 2012, along with their polls. It’s not foolproof, but it’s real data, and it’s reasonable to try and extrapolate real voting data like this.
I don’t know how their model would incorporate it – but maybe future models will.
When it gets through to him that the e-mail scandal ain’t going to get it, Trump will open his pie-intake orifice. He’s got a whole week to fuck this up. In this respect, I totally trust him.
Don’t the polls actually incorporate it already by asking if people have already voted? If the polls do incorporate it then 538 doesn’t need to adjust anything.