The US Election Polling / Data Thread

[Senator Clay Davis]

New state polls from Quinnipiac , all four way races

PA: Hillary 44, Trump 39
OH: Trump 41, Hillary 37
FL: Hillary 43, Trump 43
NC: Hillary 42, Trump 38

Gary Johnson's disaster today is worth monitoring since when two ways become 4 ways, Hillary's numbers collapse more to him than to Trump. So if some of his soft supporters realize he's a foreign policy lightweight, they'll likely slide back into the Hillary camp.

[CorsiHockeyLeague]

The PA and NC numbers do not make sense with the OH number.

[Senator Clay Davis]

Well really they makes sense, simply because nothing in this election does.

[Iowa_Flames_Fan]

Originally Posted by CorsiHockeyLeague

The PA and NC numbers do not make sense with the OH number.

I believe that Ohio may have proportionately more white male voters without a college degree than PA, so that may account for it partly. Also, I think that Ohio leans somewhat more republican than PA does in general. Republicans can actually win Ohio, but PA has been something of a white whale for them in the past few decades.

[CorsiHockeyLeague]

It's probably just error bars. Just knowing what I know about the way those states run, a +4 in NC and +5 in PA do not line up with a -4 in OH. At least one of those is off. If Clinton wins NC by even a slim margin, the worst case scenario should be an even slimmer loss in OH. This is a weird election, sure, but there is no circumstance where there's an 8 point swing between those two outcomes.

[Iowa_Flames_Fan]

Originally Posted by CorsiHockeyLeague

It's probably just error bars. Just knowing what I know about the way those states run, a +4 in NC and +5 in PA do not line up with a -4 in OH. At least one of those is off. If Clinton wins NC by even a slim margin, the worst case scenario should be an even slimmer loss in OH. This is a weird election, sure, but there is no circumstance where there's an 8 point swing between those two outcomes.

Normally I'd agree, but the demographics of Trump supporters do throw that off a bit. That is why Trump is running even or slightly ahead in Maine's third district, which Romney lost by 12 points.

EDIT: I should clarify that I'm not suggesting your explanation that we are seeing sampling error is wrong. That's absolutely possible too.

[wittynickname]

Originally Posted by CorsiHockeyLeague

It's probably just error bars. Just knowing what I know about the way those states run, a +4 in NC and +5 in PA do not line up with a -4 in OH. At least one of those is off. If Clinton wins NC by even a slim margin, the worst case scenario should be an even slimmer loss in OH. This is a weird election, sure, but there is no circumstance where there's an 8 point swing between those two outcomes.

NC is an interesting beast, because you have pockets of very well educated, very liberal demographics, especially in the Triangle region (Raleigh/Durham/Chapel Hill). Orange County in NC is pretty heavily populated and is one of the top 10 most educated counties in the entire US. There are plenty of transplants from more liberal Northeastern States, as well as a not-insignificant Latino population.

If the cities in NC get out and vote in large numbers, they could possibly push the state blue despite the red rural areas. Also I feel like the GOP has made a lot of enemies, with the disastrous voter suppression tactics and the transphobic legislation that has been passed there, it could easily motivate more progressive voters to get out there and try to get rid of McCrory before he brings any further embarrassment to the state.

Also if my census checking is correct, Ohio is around 80% white, North Carolina is only 62% white. Ohio has 12 and 3% Black and Latino respectively, whereas North Carolina has 21 and 10%. Demographics are the hill that Trump is going to die on. He's not pulling minority voters. So he might snag Ohio, but he'll have a hard time in more diverse states.

[nfotiu]

Originally Posted by wittynickname

NC is an interesting beast, because you have pockets of very well educated, very liberal demographics, especially in the Triangle region (Raleigh/Durham/Chapel Hill). Orange County in NC is pretty heavily populated and is one of the top 10 most educated counties in the entire US. There are plenty of transplants from more liberal Northeastern States, as well as a not-insignificant Latino population.

If the cities in NC get out and vote in large numbers, they could possibly push the state blue despite the red rural areas. Also I feel like the GOP has made a lot of enemies, with the disastrous voter suppression tactics and the transphobic legislation that has been passed there, it could easily motivate more progressive voters to get out there and try to get rid of McCrory before he brings any further embarrassment to the state.

Also if my census checking is correct, Ohio is around 80% white, North Carolina is only 62% white. Ohio has 12 and 3% Black and Latino respectively, whereas North Carolina has 21 and 10%. Demographics are the hill that Trump is going to die on. He's not pulling minority voters. So he might snag Ohio, but he'll have a hard time in more diverse states.

NC and Ohio numbers make sense. Trump has lost a lot of the southern, white, old school, wealthy republicans in NC. But he's probably stole some support from the white, blue collar democrats in Ohio. It's always been my theory that he's going to move the needle in both directions in a lot of states. In some states the gains and losses probably cancel each other out.

Forecasts that rely too much on past voting patterns probably aren't going to be very reliable this time around.

[Parallex]

Originally Posted by Senator Clay Davis

Hillary's numbers collapse more to him than to Trump. So if some of his soft supporters realize he's a foreign policy lightweight, they'll likely slide back into the Hillary camp.

Yeah that's one of two advantages that Clinton is going to have on Election Day. (the first and more important is her extensive ground game), the Third Party vote is never as high as Third Party polling so that vote will likely leech back to the GOP/DNC candidates and since Clinton Polls better in two choice polls it's more likely that it'll go to her.

It's not a huge advantage probably not more then a point where it counts but it's something.

[Senator Clay Davis]

A plethora of state polls today, everything pretty tight. NBC/WSJ/Marist for the first four, CBS/YouGov for the last two. Overall I'd say Hillary should like these numbers slightly more than Trump.

Nevada is Trump +1 in a four way, Hillary +1 in a two way.
Arizona is Trump +1 in a two way, +2 in a four way.
New Hampshire is Hillary +1 in a two way, +2 in a four way.
Georgia is Trump +3 in a two way, +2 in a four way.
Florida is Hillary +2 in a four way
Ohio is Hillary +7 in a four way.

[afc wimbledon]

I'm deeply disturbed by the thought of Hilary and trump in a four way.

[peter12]

Originally Posted by afc wimbledon

I'm deeply disturbed by the thought of Hilary and trump in a four way.

With you there, buddy. Watching with eyes partially covered, I mean.

[Drak]

Not sure if this was posted. It is from a few days ago.



http://abcnews.go.com/Politics/quali...ry?id=41984690

The latest WaPo/ABC News poll shows Hillary Clinton leading Donald Trump 51-43 percent in a two-way race. That's good news for Clinton, but the rest of the poll is even better news. Asked who they believe will win, Clinton leads by a whopping 58-29 percent. Historically, this is a pretty predictive indicator. President Obama's approval rating is up to 58 percent, which is good news for the candidate of the same party. Clinton also leads on all four questions about character and all five questions about issues.

[troutman]

Betting Odds are usually a strong predictor too, and last I checked Clinton was at about 70%.

https://electionbettingodds.com/

Down 8% to 61% in the last week.

[Looch City]

Originally Posted by Drak

Not sure if this was posted. It is from a few days ago.
http://abcnews.go.com/Politics/quali...ry?id=41984690

Haha I can only imagine what Trump's answers to those "five issues" were.

"I will make it better."

Yes, but how?

"I will gather the smartest people, best people I know. General, doctors, colonels. We're going to make America great again."

Repeat x5.

[Izzle]

Originally Posted by Iggy City

Haha I can only imagine what Trump's answers to those "five issues" were.

"I will make it better."

Yes, but how?

"I will gather the smartest people, best people I know. General, doctors, colonels. We're going to make America great again."

Repeat x5.

Ok, I read the bolded in Trump's voice before I even realized I did it.

Your Trump imitation game is on point.

[octothorp]

Originally Posted by Drak

Not sure if this was posted. It is from a few days ago.

http://abcnews.go.com/Politics/quali...ry?id=41984690

I'm curious to know why 'who you think will win' would be an effective indicator. All it's doing, IMO, is taking all the sampling and inaccuracy issues that even good polls have, and adding a level of individual interpretation on top of it... is it because people are in theory answering about who their local/regional community seems likely to vote for, so it's kinda sampling a larger audience than just one person? Is it because the difference between someone's win-expectation and their support percentage forcasts a lack of voter enthusiasm? (ie. a lot of Trump supporters don't believe he'll win and hence won't actually vote?)

I'm not disputing that this may be a predictive indicator, I'm just confused as to why.

[Iowa_Flames_Fan]

Originally Posted by octothorp

I'm curious to know why 'who you think will win' would be an effective indicator. All it's doing, IMO, is taking all the sampling and inaccuracy issues that even good polls have, and adding a level of individual interpretation on top of it... is it because people are in theory answering about who their local/regional community seems likely to vote for, so it's kinda sampling a larger audience than just one person? Is it because the difference between someone's win-expectation and their support percentage forcasts a lack of voter enthusiasm? (ie. a lot of Trump supporters don't believe he'll win and hence won't actually vote?)

I'm not disputing that this may be a predictive indicator, I'm just confused as to why.

It's an interesting question. I like your hypothesis that it indirectly samples a larger group. It also suggests that voters are pretty good at taking the temperature of a race, which is in itself surprising.

[octothorp]

Slate and a new company called Votecastr are going to do live election-day real-time projections, based on voter turnout in swing states. That's going to make for a fun day of number-watching.

[Buster]

Is anyone doing any polling on whether or not a Biden replacement would do better?