|
Actually that +14/18 PRRI poll was done a week ago. I think it's extremely unlikely Biden will actually win 14-18 but it's certainly not completely absurd given other polling, and 538 seems to rate the pollster fairly highly.
|
#
?
Oct 19, 2020 13:17
|
|
|
|
| # ? Jun 9, 2024 03:12 |
|
|
Chinese Gordon posted:It's a display error on 538: 70% turnout at 56-38% would shatter so many gerrymanders. How many senate seats would we win? 11? 12?
|
|
#
?
Oct 19, 2020 13:18
|
|
|
You'd expect Biden's national numbers to be falling back in line with the state averages by now, but that hasn't really happened. So something still doesn't add up. Either the national margins are overestimating Biden's support by 2-3 points, or the state-level polling is missing some significant shift. Or Biden's just running up the score in high population states that are voting Democratic anyway.
|
|
#
?
Oct 19, 2020 13:23
|
|
|
exquisite tea posted:Apparently that polling period was from over a month ago, in September. Just appears to be uncommonly favorable towards Biden. Well given we have seen anything up to Biden +21 (LOLNO) I'm not so sure +15 is out of the question anymore. quote:You'd expect Biden's national numbers to be falling back in line with the state averages by now, but that hasn't really happened. So something still doesn't add up. Either the national margins are overestimating Biden's support by 2-3 points, or the state-level polling is missing some significant shift. Or Biden's just running up the score in high population states that are voting Democratic anyway. Funny, was just pondering that. And I think the likely answer is becoming 2) State level is missing something in Biden's favour. CAT INTERCEPTOR fucked around with this message at 13:26 on Oct 19, 2020 |
|
#
?
Oct 19, 2020 13:24
|
|
|
exquisite tea posted:You'd expect Biden's national numbers to be falling back in line with the state averages by now, but that hasn't really happened. So something still doesn't add up. Either the national margins are overestimating Biden's support by 2-3 points, or the state-level polling is missing some significant shift. Or Biden's just running up the score in high population states that are voting Democratic anyway. You'd have to think that option 3 is a natural consequence of hyperpolarization in general , right?
|
|
#
?
Oct 19, 2020 13:25
|
|
|
exquisite tea posted:You'd expect Biden's national numbers to be falling back in line with the state averages by now, but that hasn't really happened. So something still doesn't add up. Either the national margins are overestimating Biden's support by 2-3 points, or the state-level polling is missing some significant shift. Or Biden's just running up the score in high population states that are voting Democratic anyway. Wasserman says the GOP congressional internals back up the national numbers
|
|
#
?
Oct 19, 2020 13:26
|
|
|
Given the hyperpolarization in our politics right now I don't think it's at all out of line for Biden to get +10 or so nationally but to only get mid-to-high single digit victories in the competitive states that he's leading in. In fact, that's why Trump claiming he's gonna totes win New York this time guys (swearsie realsies!) is even funnier than it was in 2016.Charlz Guybon posted:Wasserman says the GOP congressional internals back up the national numbers Well, that's the thing. The Senate and Congressional polling is as equally awful for the GOP as the Presidential race is for Trump. We didn't see this phenomenon in 2016; if anything, we should have smelled a rat when in the closing weeks of the race, the Senate and House polling all shifted hard Republican even as the Presidential race didn't appear to move at all. Fritz Coldcockin fucked around with this message at 13:32 on Oct 19, 2020 |
|
#
?
Oct 19, 2020 13:28
|
|
|
Yeah I definitely remember doing some mental bargaining with the fact that polling seemed to indicate we were gonna lose the election in the house and the senate even if Clinton won. I never really thought about how that should have meant Clinton was going to lose also.
|
|
#
?
Oct 19, 2020 13:37
|
|
|
Zwabu posted:As I understand it, 538 actually runs a bunch of simulated elections on their computers to derive the numbers. So called Monte Carlo style. So the 18 percent means that in a thousand simulations, Trump was winner in 180 and Biden in 920, etc. Great math here.
|
|
#
?
Oct 19, 2020 13:42
|
|
|
Charlz Guybon posted:Wasserman says the GOP congressional internals back up the national numbers The thing that makes me hesitate to declare an overwhelming Democratic victory is that the generic congressional ballot, for as long as it has been polled this cycle, is still D+7 or D+8. If Biden were really up on Trump by double digits, you'd expect that number to shift as well. But it's remained exactly the same for months.
|
|
#
?
Oct 19, 2020 13:44
|
|
|
cant cook creole bream posted:The probability value isn't just something arbitrary. It seems to me, not being an expert at statistical modeling of the behavior of populations, that fundamentally the tools being used here are based on approximating random chance. That is, the basic assumption is that the error in the polls is random and can be modeled using the tools created for modeling Bernoulli trials over random populations. The basic idea is that sometimes people say they want Biden, sometimes they say they want Trump based on recent input stimulus and with time and hysteresis and they change their minds but over multiple polls the random noise cancels out and you are left with a strong signal of intent and a dominant trend. Except fundamentally, none of this is actually random. It's actually more like where somethings are relative to others and fluctuate with changes to inputs and each atomic element has it's own unique inputs, that are often imperfect, and an internal transfer function that is 'fuzzy' and can provide widely variable outputs.
|
|
#
?
Oct 19, 2020 13:48
|
|
|
538 just updated with some more recent Public Religion Research Institute polls (Oct 9-12) which have Biden at +14 and +18. I have a hard time believing these double digit national numbers when state margins are much closer and as Nate pointed out there has been a lack of high quality state polling recently.
|
|
#
?
Oct 19, 2020 13:58
|
|
|
exquisite tea posted:The thing that makes me hesitate to declare an overwhelming Democratic victory is that the generic congressional ballot, for as long as it has been polled this cycle, is still D+7 or D+8. If Biden were really up on Trump by double digits, you'd expect that number to shift as well. But it's remained exactly the same for months. yea I super don't trust the whole 'there's gonna be a huge wave' thing just because no generic congressional polls have actually said that. Like, yea I'm pretty sure Joe's gonna win and there will be some flips but I really don't think there's gonna be some massive blowout even with how bad trump's flailing.
|
|
#
?
Oct 19, 2020 13:59
|
|
|
If the state polls are that behind though, is it because Biden is running up the totals in already blue states, and the battleground states are polarized to the point of barely any shift at all? Like what otherwise explains the phenomena? Is it just there aren't any real undecideds in battleground states?
|
|
#
?
Oct 19, 2020 14:03
|
|
|
TulliusCicero posted:If the state polls are that behind though, is it because Biden is running up the totals in already blue states, and the battleground states are polarized to the point of barely any shift at all? There have also been a few polls of places like west Virginia and the like where biden is *only* behind by 10-15 points and the like. If those shifts are happening it could inflate the popular vote numbers while having no electoral college impact.
|
|
#
?
Oct 19, 2020 14:05
|
|
|
Biden isn’t up 18. But he’s up double digits and a poor pollster could then get a result like +18.
|
|
#
?
Oct 19, 2020 14:08
|
|
|
https://twitter.com/wsteaks/status/1318157804939640839?s=20
|
|
#
?
Oct 19, 2020 14:09
|
|
|
Cicero posted:Though the fact that it still shows polling firms having multiple polls covering the exact same dates seems odd. I think they do this when the poll reports multiple turnout models. A lot of polls have been reporting "high turnout" "low turnout" etc. models, plus a registered voter result; I'm pretty sure 538 just adds them all to the average (I'm not sure if they're weighted differently or not).
|
|
#
?
Oct 19, 2020 14:09
|
|
|
exquisite tea posted:You'd expect Biden's national numbers to be falling back in line with the state averages by now, but that hasn't really happened. So something still doesn't add up. Either the national margins are overestimating Biden's support by 2-3 points, or the state-level polling is missing some significant shift. Or Biden's just running up the score in high population states that are voting Democratic anyway. Biden is also doing much better in blood red States that he will never carry, like West Virginia.
|
|
#
?
Oct 19, 2020 14:10
|
|
|
We're also in a bit of a polling lull right now, I wouldn't say there hasn't been any state movement until we get some more results. Hopefully more come out this week, otherwise we'll probably have to wait a few days after the debate.
|
|
#
?
Oct 19, 2020 14:14
|
|
|
TwoQuestions posted:Biden is also doing much better in blood red States that he will never carry, like West Virginia. Which bodes well for states that have West Virginia-like areas in them, like PA, MI, WI and OH. It shows Biden can cut into Trump's margins even in rural areas.
|
|
#
?
Oct 19, 2020 14:17
|
|
|
Mainwaring posted:There have also been a few polls of places like west Virginia and the like where biden is *only* behind by 10-15 points and the like. If those shifts are happening it could inflate the popular vote numbers while having no electoral college impact. This is why the argument for the continuing existence for the EC is such bullshit: those shifts in voters in West Virginia should loving matter, those people should be heard, and in a pure popular vote they would be Instead a bunch of indecisive fucks in the Midwest and Florida decide every election depending on who get called to either side like a puppy choosing it's owner based on how many treats it gets 1/4th of a country's populace should not get to decide who becomes President
|
|
#
?
Oct 19, 2020 14:23
|
|
|
Second level analysis would be monitoring the linked in status of people working for the campaigns. If they are actively looking for new work in November (i.e. increased activity, updating their pages, creating new contacts and etc...) you would get a good indicator of confidence.
|
|
#
?
Oct 19, 2020 14:25
|
|
|
Fritz Coldcockin posted:Which bodes well for states that have West Virginia-like areas in them, like PA, MI, WI and OH. It shows Biden can cut into Trump's margins even in rural areas. Thats the reason why Texas is so hard, they do have some very large blue cities, but also absurdly enormous rural areas where the Democrat gets just annihilated 75-20 or even 80-15. Getting up to just losing by 25% in chud-land would flip the state.
|
|
#
?
Oct 19, 2020 14:29
|
|
|
exquisite tea posted:The thing that makes me hesitate to declare an overwhelming Democratic victory is that the generic congressional ballot, for as long as it has been polled this cycle, is still D+7 or D+8. If Biden were really up on Trump by double digits, you'd expect that number to shift as well. But it's remained exactly the same for months. If you've got a statistically significant number of Republican voters who are happy with GOP/Trump policies but can't stand the man himself, you'd expect downballot approvals to stay steady while Ttump kept falling further and further behind.
|
|
#
?
Oct 19, 2020 14:40
|
|
|
Yard sign report: here in NE FL, almost all of the Trump signs are..."evolved" I guess. They're like 4' x 8' and framed in wood. I guess so you can't steal them (?). I dunno but they're all over the place here. About to go vote and did my sample ballot but I wish there were a better way to research judges and poo poo like "Mosquito Control Board".
|
|
#
?
Oct 19, 2020 14:52
|
|
|
Shimrra Jamaane posted:Biden isn’t up 18. But he’s up double digits and a poor pollster could then get a result like +18. According to the excerpt from the poll's methodology upthread the +18 result was under an assumption of 70% turnout, which just flat ain't happening. The same poll gave +14 under a more conservative 55% turnout estimate.
|
|
#
?
Oct 19, 2020 14:52
|
|
|
Seems like a good sign https://mobile.twitter.com/tbonier/status/1318180016786726914
|
|
#
?
Oct 19, 2020 14:55
|
|
|
Rigel posted:Thats the reason why Texas is so hard, they do have some very large blue cities, but also absurdly enormous rural areas where the Democrat gets just annihilated 75-20 or even 80-15. Getting up to just losing by 25% in chud-land would flip the state. Well, rural areas in the Sun Belt and rural areas in the Rust Belt are differently predisposed towards Biden. While it's apparent that he appeals more to the OH/PA/WV/MI/WI rural voter, it remains to be seen how he does in the AZ/TX/FL/NC rural areas. If he can cut into them BOTH by 5-10%, then he will sweep them all.
|
|
#
?
Oct 19, 2020 15:00
|
|
|
Charlz Guybon posted:Seems like a good sign
|
|
#
?
Oct 19, 2020 15:04
|
|
|
The guy who runs the early vote analysis Twitter was taking shots at the poll pundits over the claim that you can’t interpret anything from the early days.
|
|
#
?
Oct 19, 2020 15:06
|
|
|
TwoQuestions posted:Does anyone know who the Public Religion Research Institute is? Clicking on the poll gave a 404 for me. In case you're curious, it's a research institute founded by a Baptist minister and former professor of Religious Studies named Dr. Ron Jones. It's most famous for the American Values Atlas, which looks at religious affiliation and beliefs and how it affects public policy. Jones, who's an Evangelical himself, is critical of a lot of the modern white Evangelical movement, arguing that it's tied up with white supremacy and a blind embrace of capitalism, and more generally, that white Christianity in America has never come to terms with a legacy of white supremacy, and until it does that, the US can't really get rid of racism. He's also, as you can imagine, not a big fan of Trump. Regardless of the PRRI founder's personal views, its pretty well respected.
|
|
#
?
Oct 19, 2020 15:09
|
|
|
Shimrra Jamaane posted:The guy who runs the early vote analysis Twitter was taking shots at the poll pundits over the claim that you can’t interpret anything from the early days. which early vote analyst?
|
|
#
?
Oct 19, 2020 15:11
|
|
|
Epicurius posted:In case you're curious, it's a research institute founded by a Baptist minister and former professor of Religious Studies named Dr. Ron Jones. It's most famous for the American Values Atlas, which looks at religious affiliation and beliefs and how it affects public policy. Jones, who's an Evangelical himself, is critical of a lot of the modern white Evangelical movement, arguing that it's tied up with white supremacy and a blind embrace of capitalism, and more generally, that white Christianity in America has never come to terms with a legacy of white supremacy, and until it does that, the US can't really get rid of racism. He's also, as you can imagine, not a big fan of Trump. That's very interesting. Thank you.
|
|
#
?
Oct 19, 2020 15:14
|
|
|
abelwingnut posted:which early vote analyst? Michael McDonald.
|
|
#
?
Oct 19, 2020 15:16
|
|
|
Epicurius posted:In case you're curious, it's a research institute founded by a Baptist minister and former professor of Religious Studies named Dr. Ron Jones. It's most famous for the American Values Atlas, which looks at religious affiliation and beliefs and how it affects public policy. Jones, who's an Evangelical himself, is critical of a lot of the modern white Evangelical movement, arguing that it's tied up with white supremacy and a blind embrace of capitalism, and more generally, that white Christianity in America has never come to terms with a legacy of white supremacy, and until it does that, the US can't really get rid of racism. He's also, as you can imagine, not a big fan of Trump. Thanks! I tried looking at their site when the (incorrect) polls were posted, but I think a couple million others were too.
|
|
#
?
Oct 19, 2020 15:17
|
|
|
These ghouls will have no problem getting employed and getting invited on to all sorts of media over and over again, same as every ex-GWB staffer and, hell, every ex-Trump staffer. Anyway, wrap it up, Bidenaliures. https://twitter.com/PpollingNumbers/status/1318193893448519680
|
|
#
?
Oct 19, 2020 15:19
|
|
|
Wanna talk about some good international polling news: MAS landslide victory in Bolivia tossing out the fascist coup 
|
|
#
?
Oct 19, 2020 15:20
|
|
|
You had me excited for this guy: https://www.youtube.com/watch?v=MJTiDMfDeyE 
|
|
#
?
Oct 19, 2020 15:25
|
|
|
|
| # ? Jun 9, 2024 03:12 |
|
|
brugroffil posted:These ghouls will have no problem getting employed and getting invited on to all sorts of media over and over again, same as every ex-GWB staffer and, hell, every ex-Trump staffer. I was just reading this: https://twitter.com/NateSilver538/status/1318191837262319620
|
|
#
?
Oct 19, 2020 15:26
|
|