US Election

Discussions about serious topics, for serious people
Locked
User avatar
dyqik
Princess POW
Posts: 7526
Joined: Wed Sep 25, 2019 4:19 pm
Location: Masshole
Contact:

Re: US Election

Post by dyqik » Tue Sep 29, 2020 6:38 pm

El Pollo Diablo wrote:
Tue Sep 29, 2020 6:33 pm
Has something happened in Georgia? Because all of a sudden Raphael Warnock is polling ahead of the two republican candidates in the special senate election.
Quinnipiac has a bit of a pro-Dem house effect in the south. The other thing is that Lieberman, the nominally D spoiler, is falling back.

But you could also be seeing fall-out from the Senate prioritizing a SCOTUS judge over bailouts, particularly for airlines (Atlanta is headquarters of Delta, and support for airline workers expires tomorrow, with the Senate not wanting to pass anything to renew it).

User avatar
Grumble
Light of Blast
Posts: 4746
Joined: Mon Nov 11, 2019 1:03 pm

Re: US Election

Post by Grumble » Tue Sep 29, 2020 6:51 pm

lpm wrote:
Tue Sep 29, 2020 4:43 pm
Of course it isn't. Polling resources are limited, so deploy resources on a state that also reveals other states.

If you were only allowed one poll, you'd poll Pennsylvania. If allowed a second, then Florida. From those two you can pretty much get all the marginals.
That's how correlation works. If you poll Georgia as well then that only gives extra indications for Louisiana or Alabama that you don't care about. North Carolina gives indications for Tennessee and Kentucky.

The only benefit to polling Alaska or Iowa is for the Senate races.
But you only know it’s correlated because of polls in those other states...
where once I used to scintillate
now I sin till ten past three

Squeak
Catbabel
Posts: 636
Joined: Tue Dec 03, 2019 6:27 am

Re: US Election

Post by Squeak » Tue Sep 29, 2020 6:56 pm

And states don't move in lockstep from one election to the next. It's through polling that you can see how well the last election.'s correlations are holding up.

monkey
After Pie
Posts: 1906
Joined: Wed Nov 13, 2019 5:10 pm

Re: US Election

Post by monkey » Tue Sep 29, 2020 7:24 pm

dyqik wrote:
Tue Sep 29, 2020 6:38 pm
El Pollo Diablo wrote:
Tue Sep 29, 2020 6:33 pm
Has something happened in Georgia? Because all of a sudden Raphael Warnock is polling ahead of the two republican candidates in the special senate election.
Quinnipiac has a bit of a pro-Dem house effect in the south. The other thing is that Lieberman, the nominally D spoiler, is falling back.

But you could also be seeing fall-out from the Senate prioritizing a SCOTUS judge over bailouts, particularly for airlines (Atlanta is headquarters of Delta, and support for airline workers expires tomorrow, with the Senate not wanting to pass anything to renew it).
There were calls for Liebermann to quit the other day, and Warnock has picked up a couple of ex-presidential backings (Obama and Carter, Carter after the poll). Seems like there's a bit of a campaign to get the Ds to coalesce behind a single candidate, which seems to explain that. I suspect it may have been spurred on by having to pick a new judge. EDIT: 'cos if they don't get over 50%, they won't be a senator till December (if at all), which might be too late.

monkey
After Pie
Posts: 1906
Joined: Wed Nov 13, 2019 5:10 pm

Re: US Election

Post by monkey » Tue Sep 29, 2020 7:45 pm

monkey wrote:
Tue Sep 29, 2020 7:24 pm
dyqik wrote:
Tue Sep 29, 2020 6:38 pm
El Pollo Diablo wrote:
Tue Sep 29, 2020 6:33 pm
Has something happened in Georgia? Because all of a sudden Raphael Warnock is polling ahead of the two republican candidates in the special senate election.
Quinnipiac has a bit of a pro-Dem house effect in the south. The other thing is that Lieberman, the nominally D spoiler, is falling back.

But you could also be seeing fall-out from the Senate prioritizing a SCOTUS judge over bailouts, particularly for airlines (Atlanta is headquarters of Delta, and support for airline workers expires tomorrow, with the Senate not wanting to pass anything to renew it).
There were calls for Liebermann to quit the other day, and Warnock has picked up a couple of ex-presidential backings (Obama and Carter, Carter after the poll). Seems like there's a bit of a campaign to get the Ds to coalesce behind a single candidate, which seems to explain that. I suspect it may have been spurred on by having to pick a new judge. EDIT: 'cos if they don't get over 50%, they won't be a senator till December (if at all), which might be too late.
Sorry, ignore the edit, getting confused between what happens with the Senate seat and John Lewis's old congressional seat.

User avatar
lpm
Junior Mod
Posts: 5944
Joined: Mon Nov 11, 2019 1:05 pm

Re: US Election

Post by lpm » Tue Sep 29, 2020 7:51 pm

Squeak wrote:
Tue Sep 29, 2020 6:56 pm
And states don't move in lockstep from one election to the next. It's through polling that you can see how well the last election.'s correlations are holding up.
The correlations aren't based on polling.

They are calculated from % Hispanic, rural white without college ed, proportion >65 years, that sort of thing.

Pennsylvania has a bit of everything. Poll in detail, then do stuff like "Clinton won suburban women with college ed by +12, Biden is now polling +18", then apply that segment shift to other states.

The states don't move in lockstep because they have different proportions of each segment. But the segments generally move together.
⭐ Awarded gold star 4 November 2021

User avatar
El Pollo Diablo
Stummy Beige
Posts: 3323
Joined: Wed Sep 25, 2019 4:41 pm
Location: FBPE

Re: US Election

Post by El Pollo Diablo » Tue Sep 29, 2020 8:07 pm

dyqik wrote:
Tue Sep 29, 2020 6:38 pm
El Pollo Diablo wrote:
Tue Sep 29, 2020 6:33 pm
Has something happened in Georgia? Because all of a sudden Raphael Warnock is polling ahead of the two republican candidates in the special senate election.
Quinnipiac has a bit of a pro-Dem house effect in the south. The other thing is that Lieberman, the nominally D spoiler, is falling back.

But you could also be seeing fall-out from the Senate prioritizing a SCOTUS judge over bailouts, particularly for airlines (Atlanta is headquarters of Delta, and support for airline workers expires tomorrow, with the Senate not wanting to pass anything to renew it).
It's not just quinnipiac, it's civiqs and Monmouth University as well
If truth is many-sided, mendacity is many-tongued

User avatar
dyqik
Princess POW
Posts: 7526
Joined: Wed Sep 25, 2019 4:19 pm
Location: Masshole
Contact:

Re: US Election

Post by dyqik » Tue Sep 29, 2020 9:22 pm

El Pollo Diablo wrote:
Tue Sep 29, 2020 8:07 pm
dyqik wrote:
Tue Sep 29, 2020 6:38 pm
El Pollo Diablo wrote:
Tue Sep 29, 2020 6:33 pm
Has something happened in Georgia? Because all of a sudden Raphael Warnock is polling ahead of the two republican candidates in the special senate election.
Quinnipiac has a bit of a pro-Dem house effect in the south. The other thing is that Lieberman, the nominally D spoiler, is falling back.

But you could also be seeing fall-out from the Senate prioritizing a SCOTUS judge over bailouts, particularly for airlines (Atlanta is headquarters of Delta, and support for airline workers expires tomorrow, with the Senate not wanting to pass anything to renew it).
It's not just quinnipiac, it's civiqs and Monmouth University as well
I wrote that before I'd seen those.

User avatar
Little waster
After Pie
Posts: 2385
Joined: Tue Nov 12, 2019 12:35 am
Location: About 1 inch behind my eyes

Re: US Election

Post by Little waster » Wed Sep 30, 2020 8:57 am

Here's a fun fact:-
29. Who do you trust more on information about the coronavirus: President Trump or scientists at the CDC?
LIKELY VOTERS..............................................

Tot Rep Dem Ind

Trump 21% 51% 1% 17%
CDC scientists 70 36 97 73
DK/NA 9 13 2 10
Or to spell it out as the formatting is a bit borked, 51% of likely voters who are Republican trust Trump more than the CDC over COVID information. :shock:

I'm not sure how you would begin to crack a layer of unreason that thick. :?
This place is not a place of honor, no highly esteemed deed is commemorated here, nothing valued is here.
What is here was dangerous and repulsive to us.
This place is best shunned and left uninhabited.

bmforre
Snowbonk
Posts: 504
Joined: Mon Nov 11, 2019 1:15 pm
Location: Trondheim

Re: US Election

Post by bmforre » Wed Sep 30, 2020 10:56 am

Little waster wrote:
Wed Sep 30, 2020 8:57 am
... 51% of likely voters who are Republican trust Trump more than the CDC over COVID information. :shock:

I'm not sure how you would begin to crack a layer of unreason that thick. :?
Have you considered https://en.wikipedia.org/wiki/Tunnel_boring_machine granite rated tunnel boring machines?

User avatar
dyqik
Princess POW
Posts: 7526
Joined: Wed Sep 25, 2019 4:19 pm
Location: Masshole
Contact:

Re: US Election

Post by dyqik » Wed Sep 30, 2020 2:50 pm

Adding to Senate fun, a new poll of Kansas is showing the Democrat ahead there (well within MoE though). Democrats already won the governorship in 2018.

User avatar
Vertigowooyay
Snowbonk
Posts: 423
Joined: Mon Nov 11, 2019 1:25 pm

Re: US Election

Post by Vertigowooyay » Thu Oct 01, 2020 8:43 am

Calm yourself Doctor NotTheNineO’ClockNews. We’re men of science. We fear no worldly terrors.

User avatar
lpm
Junior Mod
Posts: 5944
Joined: Mon Nov 11, 2019 1:05 pm

Re: US Election

Post by lpm » Thu Oct 01, 2020 11:58 am

I hadn't realised just how strongly the 538 model adjusts for time left. I'd thought with only 5 weeks to go the time effect was already declining towards zero, and the Biden % chance would only drift up a bit from the current level.

But in fact 538's underlying Trump chance on current polling is only 9%, with the time left factor bumping that up to 21%.

In other words, if the election was today, the model would give Trump only 9% chance.

The Economist model gives Trump 12%, because they have a far lower time left factor.
⭐ Awarded gold star 4 November 2021

User avatar
El Pollo Diablo
Stummy Beige
Posts: 3323
Joined: Wed Sep 25, 2019 4:41 pm
Location: FBPE

Re: US Election

Post by El Pollo Diablo » Thu Oct 01, 2020 12:20 pm

Yeah, obviously the more time there is, the more chance there is for malicious actors like the head of the FBI to f.ck everything up.
If truth is many-sided, mendacity is many-tongued

User avatar
dyqik
Princess POW
Posts: 7526
Joined: Wed Sep 25, 2019 4:19 pm
Location: Masshole
Contact:

Re: US Election

Post by dyqik » Thu Oct 01, 2020 12:32 pm

Something like 10% of the number of votes cast in 2016 have already been cast in some swing states. I don't think the 538 model was adjusted for the massive use of early absentee voting, and a possible big uptick in turnout.

User avatar
lpm
Junior Mod
Posts: 5944
Joined: Mon Nov 11, 2019 1:05 pm

Re: US Election

Post by lpm » Thu Oct 01, 2020 1:09 pm

El Pollo Diablo wrote:
Thu Oct 01, 2020 12:20 pm
Yeah, obviously the more time there is, the more chance there is for malicious actors like the head of the FBI to f.ck everything up.
21% to 9%, 12 percentage points, seems high?

That's just Trump side - 12% scenarios that help Trump - with presumably an effect equally the other way, e.g. events where the FBI arrest Don Jr and Eric.

The coming Russian hacked docs will hit soon. But on the other hand, the Biden lead has been frozen for months and early voting locks it in.

538 must be basing the time effect on previous elections. Some average movement calc. But 2020 is a massive outlier in terms of stability of polling.
⭐ Awarded gold star 4 November 2021

Squeak
Catbabel
Posts: 636
Joined: Tue Dec 03, 2019 6:27 am

Re: US Election

Post by Squeak » Thu Oct 01, 2020 1:38 pm

I think a fairly hefty chunk of the uncertainty in their model will resolve when the October jobs report comes out.

They put in a big fudge factor for the pandemic and economy jumping in either direction and the October jobs report is the last big piece of economic data to come in before the election.

They also put in a fudge factor of "how many full width headlines in the NY Times since January, as a proxy for a turbulent news year, and this year blew out all previous years (on top of an ongoing trend towards big headlines).

User avatar
lpm
Junior Mod
Posts: 5944
Joined: Mon Nov 11, 2019 1:05 pm

Re: US Election

Post by lpm » Fri Oct 02, 2020 5:10 am

Ha ha ha ha please please please please ha ha ha ha ha ha please please please please please please oh my god it would be so amazing
⭐ Awarded gold star 4 November 2021

User avatar
Vertigowooyay
Snowbonk
Posts: 423
Joined: Mon Nov 11, 2019 1:25 pm

Re: US Election

Post by Vertigowooyay » Fri Oct 02, 2020 5:13 am

lpm wrote:
Fri Oct 02, 2020 5:10 am
Ha ha ha ha please please please please ha ha ha ha ha ha please please please please please please oh my god it would be so amazing
Oh gosh. Oh my. Who could possibly have foreseen this?
Calm yourself Doctor NotTheNineO’ClockNews. We’re men of science. We fear no worldly terrors.

User avatar
lpm
Junior Mod
Posts: 5944
Joined: Mon Nov 11, 2019 1:05 pm

Re: US Election

Post by lpm » Fri Oct 02, 2020 5:13 am

Twitter:

Donald Trump tests positive for karma.
⭐ Awarded gold star 4 November 2021

User avatar
lpm
Junior Mod
Posts: 5944
Joined: Mon Nov 11, 2019 1:05 pm

Re: US Election

Post by lpm » Fri Oct 02, 2020 5:19 am

Some people want to go high. "Nobody deserves this terrible disease." I don't. He totally f.cking deserves it.
⭐ Awarded gold star 4 November 2021

User avatar
lpm
Junior Mod
Posts: 5944
Joined: Mon Nov 11, 2019 1:05 pm

Re: US Election

Post by lpm » Fri Oct 02, 2020 5:23 am

They were quite far apart at the debate. But Trump was pretty shouty. Two nights ago.

Is there a chance he infected Biden?
⭐ Awarded gold star 4 November 2021

secret squirrel
Snowbonk
Posts: 551
Joined: Wed Nov 13, 2019 12:42 pm

Re: US Election

Post by secret squirrel » Fri Oct 02, 2020 5:29 am

lol

User avatar
Martin_B
After Pie
Posts: 1614
Joined: Mon Nov 11, 2019 1:20 pm
Location: Perth, WA

Re: US Election

Post by Martin_B » Fri Oct 02, 2020 6:33 am

What do the polls say if it's a choice between Biden and Pence?
"My interest is in the future, because I'm going to spend the rest of my life there"

User avatar
Vertigowooyay
Snowbonk
Posts: 423
Joined: Mon Nov 11, 2019 1:25 pm

Re: US Election

Post by Vertigowooyay » Fri Oct 02, 2020 7:05 am

The only 'good' outcome from this is long covid. A weakened, debilitated Trump can't lie about the severity of the virus. Plus, frailty doesn't play well with his base.

Mild case - means he can say how trivial Covid is, look at me I beat it easy.

Death - Pence wins Nov 3rd.
Calm yourself Doctor NotTheNineO’ClockNews. We’re men of science. We fear no worldly terrors.

Locked