[RD] US 2016 election: Poll watching thread

Early voting is baked into the poll numbers, in the sense that pollsters should be asking respondents if they voted, and if so, for whom. And then it simply goes into the poll number as part of the sample response. So the polling numbers and, by extension, the forecast ought to include the early vote.

But I think that exposes a weakness of an exclusively poll-based statistical model. They currently have Nevada - whose EVs are of vital importance to Trump given how few opportunities he has - as little better than a toss-up. But Jon Ralston, who knows Nevada voting perhaps better than anyone, says that early voting patterns are a close match for 2012, where Obama won by nearly 7 points. He says barring some huge shakeup in the voting patterns of party registrants, i.e. either lots of registered Ds or Is voting Trump - which polling says is not going to happen - Clinton is going to win the state.

Early voting in most other places seems to be showing far less clear results - except for Texas, where most commenters still seem to give a wink and a nudge but I think Clinton has a real chance at winning.
Nevada is not of critical importance. Rather, it is a lesser included state. He has to get at least one state that is much more difficult. If he gets that state, say Michigan, Nevada will win comfortably.

You are correct that poll base prediction models have flaws. One is that they are always yesterday's news--if you are very fortunate. More often, they are week old news, which can be like week old fish. OTOH using a previous election for comparison is a different minefield. Using any previous election as a standard for this one is unusually problematic.

I am not sure why you think Clinton has a chance in Texas. Trump has as much chance to win New Jersey. In either case, there will 35-40 states in front of them.

J
 
Most Texas polls are between +3 and +7 for Trump. Indiana and Missouri are leaning more Trump than that
 
Nevada is not of critical importance. Rather, it is a lesser included state. He has to get at least one state that is much more difficult. If he gets that state, say Michigan, Nevada will win comfortably.

Nope. Simply not true. There is little correlation between the results in Michigan and the results in Nevada. Nevada is an outlier in many ways; it is quite conceivable that Trump picks off a Michigan or Wisconsin but loses Nevada. They are totally different states, so it is incorrect to say that they rise and fall together.

You are correct that poll base prediction models have flaws. One is that they are always yesterday's news--if you are very fortunate. More often, they are week old news, which can be like week old fish. OTOH using a previous election for comparison is a different minefield. Using any previous election as a standard for this one is unusually problematic.

I am not sure why you think Clinton has a chance in Texas. Trump has as much chance to win New Jersey. In either case, there will 35-40 states in front of them.

J

Demographics, for one. Texas is a well-educated state overall. It also has a high share of people of color, and Latinos in particular. In a year where education is an unusually large dividing line between voters for each candidate, a well-educated red state is naturally going to be more competitive.

Probably the biggest reason, though, is the lack of any other races of note. The Congressional districts are almost all gerrymandered beyond any reasonable competitiveness. There is no Senate race. I'm guessing the large majority of local races are similarly noncompetitive, not that local races drive much enthusiasm anyways. The presidential race is the whole ball of wax, and unlike many of its neighbors Texas lines up well demographically for Clinton to win. It's still a very hard lift because much of the state is culturally conservative, but there is a perfect enough storm that the state just looks like one that could go blue. I wouldn't put money on it at less than 4-1 odds, but I'd say it's more likely than most people think.

New Jersey is a diverse, wealthy, well-educated blue state. There just isn't any rational basis for thinking it will be close, as the particulars of the state play very poorly for Trump. There is a much better argument for Texas being close, and the polling indicates as much. The majority of news organizations have taken it out of "Safe R" and moved it to likely or even leans R. So I'm not the only one looking at it as potentially competitive.
 
Last edited:
It's also worth nothing that it's currently much more likely that Trump will win the electoral college while losing the popular vote than the other way around. Clinton has lead by about 2 percentage points to win the electoral college reliably. She still is doing this by most polling averages and models, but this election is a genuine nailbiter with a slight-to-moderate favorite at this point, rather than any kind of foregone conclusion.
 
Nope. Simply not true. There is little correlation between the results in Michigan and the results in Nevada. Nevada is an outlier in many ways; it is quite conceivable that Trump picks off a Michigan or Wisconsin but loses Nevada. They are totally different states, so it is incorrect to say that they rise and fall together.

Demographics, for one. Texas is a well-educated state overall. It also has a high share of people of color, and Latinos in particular. In a year where education is an unusually large dividing line between voters for each candidate, a well-educated red state is naturally going to be more competitive.

Probably the biggest reason, though, is the lack of any other races of note. The Congressional districts are almost all gerrymandered beyond any reasonable competitiveness. There is no Senate race. I'm guessing the large majority of local races are similarly noncompetitive, not that local races drive much enthusiasm anyways. The presidential race is the whole ball of wax, and unlike many of its neighbors Texas lines up well demographically for Clinton to win. It's still a very hard lift because much of the state is culturally conservative, but there is a perfect enough storm that the state just looks like one that could go blue. I wouldn't put money on it at less than 4-1 odds, but I'd say it's more likely than most people think.

New Jersey is a diverse, wealthy, well-educated blue state. There just isn't any rational basis for thinking it will be close, as the particulars of the state play very poorly for Trump. There is a much better argument for Texas being close, and the polling indicates as much. The majority of news organizations have taken it out of "Safe R" and moved it to likely or even leans R. So I'm not the only one looking at it as potentially competitive.
Of course, there is a significant correlation between widely separated states. It's the nationwide sentiment. States will tend to maintain the same relationship to the national mean. Trump cannot win Michigan without moving the nationwide numbers. Nevada will reflect the same shift. Perhaps this should be in the past tense. Trump has managed to move the underlying Statistic. Individual states are slower to report but signs of the expected movement are widespread.

Texas and New Jersey occupy similar positions at the opposite extremes of the spectrum, with NJ being a bit further left. All of your demographic arguments have weight. So, this year Texas has about fifteen states to the right and NJ has only ten to its left. In 2012 it was more even, which means you are correct. The Hispanic vote and the education factor are working against Trump, just not enough to come close to swinging the state. The lack of other races is also true, but I think that helps Republicans. Trump is a Republican, in name at least.

J
 
Several polls have Texas at Trump +3. That is most definitely close to swinging the state. I'll concede that Clinton getting those last 3 is not likely to happen, but the result looks to be far closer than what we've seen in a long time.

You can't on the one hand say it's dangerous to model things off of prior years, and then proceed to argue Michigan v. Nevada based on what has happened in prior years. I agree that in a year similar to 2012, Nevada would tip well before Michigan. However the demographics are working differently this year, and Nevada is rather idiosyncratic to begin with. Plus, you have to factor in early voting - the noise and movement that has happened over the last few days matters less in Nevada because people have been voting now for a couple of weeks, and a lot of votes are already in. Michigan doesn't have early voting, so stuff like Comey has more of an impact there, if it has any at all.

Granted, the most likely scenario is that Clinton wins Michigan and Nevada, and Michigan by a wider margin, but things are rather uncertain this year. Jon Ralston thinks Nevada is going to end up around where it did in 2012, while both campaigns are making final week stops in Michigan. So it's not outlandish to think they could occupy different slots than normal.
 
The #1 reason Clinton probably won't take Texas is the campaign's lack of an attempt at it. It's expensive to try and take it. Trump is campaigning there; Clinton is not. Grass roots organizations are really stepping up, particularly in the south and in urban centers, but it's hard to overcome the Republican governor's Trump endorsement and Texas' dependence on petroleum. States like that just love screw-global-warming candidates. Look at West Virginia and coal.
 
The hope is that Clinton makes it close, and that forces the GOP to clean house. I mean, once we get to a point where Texas isn't reliably red, they have no choice if they don't want to concede the presidency forever.
 
Early voting Exit polls in Florida have Clinton up Big by the responses still though. She should focus there. If Clinton wins Florida, game is over.

My understanding that Dem. Hispanics are showing a great turn out while black turnout is disappointing. This is worrisome for North Carolina, Florida and Virginia.
 
The hope is that Clinton makes it close, and that forces the GOP to clean house.
She might. At this point, she can only win a close election. Trump could also win a close election but I would not bet against a Trump landslide.

Why Clinton losing a close election for President will force a house cleaning is unclear.

Early voting Exit polls in Florida have Clinton up Big by the responses still though. She should focus there. If Clinton wins Florida, game is over.
This is true and you could add North Carolina. Either would be sufficient. Right now, 538 has both as dead heats. The key is that even with both FL and NC, Trump needs another big state or combination of smaller ones.

Right now the RCP no tossup map has Clinton 273-265. In the map, Trump owns FL, NC, and NV. His best chance for more votes is NH which would be a tie. Most likely, he would also get one EV from Maine since it is polling closer than NH. As Nate Silver says, Trump has a path. It is not an easy one.

J
 
Last edited:
Last edited:
So despite Hilary leading in the polls, it is possible for Trump to win by a landslide and not her?
Absolutely. Haven't you ever seen a ballgame where a team gave multiple scores late in what had been a close game. Like that. Clinton's supporters are being stressed for the first time. Anyone at all scared of looking foolish has already abandoned Trump. His numbers are fully vetted. All the faint of heart are in the Clinton camp. Even if they just stay home or vote Stein, Trump could win big. Watch the heavy blue states for Stein's numbers. If she is over 5%, the panic is on.

Early Exit polls show at 16 point advantage for Clinton. A lot of this is stemming from like more than 25% of Republicans in FL have voted for Clinton, which is really surprising. http://www.msnbc.com/the-last-word/watch/exclusive-stunning-clinton-number-in-fl-798824003656 If this is true then in the bout of presidential candidate, Trump has just been hit with a devastating roundhouse kick to the face.
Meh. A lot of Trump's support is former Democrats and early voting is not a great predictor. That said, it could bail Clinton out in some states.
http://www.realclearpolitics.com/ar...voting_a_poor_predictor_of_final_results.html

J
 
Absolutely. Haven't you ever seen a ballgame where a team gave multiple scores late in what had been a close game. Like that. Clinton's supporters are being stressed for the first time. Anyone at all scared of looking foolish has already abandoned Trump. His numbers are fully vetted. All the faint of heart are in the Clinton camp. Even if they just stay home or vote Stein, Trump could win big. Watch the heavy blue states for Stein's numbers. If she is over 5%, the panic is on.
There are plenty of faint of heart in Trump's camp, which is why he crashes in the polls whenever something damaging comes out. There are lots of people, including most of the Rep establishment, who dislike him but favor him mostly for the tax cuts, SC nominees, and other perks of having a Republican president. And no, as a Stein voter myself, I have no illusions that she'll do better than the low 2% range that she is currently polling at. Given that third party candidates usually underperform their polls, I'd expect a result in the 0.9-1.8% range.
 
That would defeat the purpose.

Two weeks out and the gap has stopped widening. It could be the debate. It could be the wiki-leaks. It could be the live mike story running its course. It could all of above of something else. Regardless, there is a slight twitch toward Trump in the data. This is somewhat akin to having a summer day 39° of instead 40° (100°F instead of 102°F)

The final debate will drop its deposit shortly.

J
16 Oct

You'd think J would know better not to judge a single outlier poll, which Clinton +4 is, but apparently not.

Holding steady at +11 in NBC News.
16 Oct

Would probably only have been interesting if it were a tie-breaker.

In answer to the original question, no. Trump's further coming slide in the polls will be utterly unrelated to his miserable performance in tonight's debate.
19 Oct.

Depending on how you score it, the Clinton lead peaked on 15 October or 18 October. In between was a dip, then recovery, followed by the long decline. From 15 Oct - 19 Oct things were very flat, then the Trump surge. I'll count this one as a win.

J
 
Hm, is the Scylla/Charybdis going to be a very close race or an upset?

Got to say, things look yugely troubling in both cases (win for Scylla or Charybdis -- Scylla can be Trump if too SJW-triggered, yet recall that Charybdis is the even worse one so i wouldn't set anything there name-wise...).

Other than the incredibly negative campaigns, i think it is bigly alarming that regardless of which awful candidate wins the other side looks determined to not concede.

Do you feel the Bern by now? :/
 
Top Bottom