2024 US Presidential Elections Megathread

[DGuller]

Nate Silver's latest free newsletter claims, quite plausibly to me, that pollsters are herding big time in battleground states.  He suspects that pollsters are afraid to buck the polling average, and cherrypick their polls or methodologies until their results match the average.  Given the margin of error, the polls can't be as close to each other as they are by chance.  The results is that the polling average itself is unmoored from reality, and no longer has the law of large numbers going for it.  In a way, the polling average is what the pollsters think it should be, not the best estimate that is a result of polling errors cancelling each other out.

[Admiral Yi]

Very interesting.

[jimmy olsen]

85-8. There are a lot of Puerto Rican voters out there. This could definitely help in some states

[FunkMonk]

Nate Silver's latest free newsletter claims, quite plausibly to me, that pollsters are herding big time in battleground states.  He suspects that pollsters are afraid to buck the polling average, and cherrypick their polls or methodologies until their results match the average.  Given the margin of error, the polls can't be as close to each other as they are by chance.  The results is that the polling average itself is unmoored from reality, and no longer has the law of large numbers going for it.  In a way, the polling average is what the pollsters think it should be, not the best estimate that is a result of polling errors cancelling each other out.

No one wants a poor rating from the poll aggregators so they all come up with roughly the same numbers. And then those numbers go into the aggregators lol

It's been funny to see the polling industry eat itself this cycle. Haven't trusted the polls at all for good reason. They're shit.

[Josquius]

As said I've heard the best thing to do with polls is to just look a handful which do have a good record of calling things right behind them.
NYT Siena is said to be one of the better ones here.

I did read something curious this morning. That trumps campaign is really heavily online targeted whilst Harris is very traditional. This could well explain why Trump seems all over whilst Harris is quite invisible - the Internet is the main place I see what's going on.

[jimmy olsen]

Seltzer, the Iowa gold standard has Harris up 3 in that state!

https://www.desmoinesregister.com/story/news/politics/iowa-poll/2024/11/02/iowa-poll-kamala-harris-leads-donald-trump-2024-presidential-race/75354033007/

[Crazy_Ivan80]

There's  that old professor who's been predicting outcomes since Reagan and he's going for Harris. Given that he hasn't really been wrong yet...
Not sure what his name is, which is annoying me

[Admiral Yi]

Allan Lichtman

I like that too.

[Crazy_Ivan80]

Allan Lichtman

I like that too.

That's the one. Thanks

[PJL]

I think some pollsters are now starting to see a last minute surge for Harris which the ears on the ground have been saying for a few days now. Which does indicate independants and sane Republicans are breaking for Harris now.

[OttoVonBismarck]

I've avoided excess prognostication over this year's election, as many of you know I have no moral choice other than to vote for Trump as a defense of the Hamas elements that will almost certainly attempt to weaponize a Harris administration against Israel, but as someone who most election cycles is waist deep in degenerate poll analysis and election gambling I have become very pessimistic about Trump's chances.

I think the state by state polls ran by national polling firms have a systemic defect in how they are being constructed, and possibly in a way that has overstated Trump's chances.

I've been growing apprehensive the last 3-4 days based on a number of data points coming in that look very bad for Trump but don't jive with the state by state polling numbers (which are still close, but with a slight Trump lean--as they have had for 3 weeks.)

The "shoe drop" moment for me is Ann Seltzer (a highly reputable pollster who only polls Iowa, and almost always more accurately than any other pollster), released  her final pre-election poll, which shows Harris up +3 in Iowa over Trump. (Note that the last 3 Presidential elections, her poll was within MOE, and in fact the only recent Iowa poll of hers to end up outside MOE was a governor's race in 2018.)

Even more concerning is the deep dive analysis Seltzer's poll has found--which is an unprecedented swing in the female vote to Harris.

Seltzer's poll has found that Seniors (65+) favor Harris, senior women by an unbelievable 63% to 28% margin, Harris was also shown to be winning amount Senior males in Iowa, 47-45.

This senior vote is particularly concerning for Trump, because there have been confounding data points out of states like PA that also show Trump losing the overall senior vote--a vote he won with 53% share in Pennsylvania in 2020.

Now yes, Iowa is one state--but to me it's the total picture that makes me worried that not only could Trump lose, I think it is possible we are looking at a perhaps unprecedented systemic polling "malfeasance" of some sort, and Trump could lose dramatically across the country.

The other issue at its core--win or lose, if Harris is close in Iowa, it is highly unlikely Trump would lose or barely win Iowa, but win WI / MI / PA, which have been consistently worse polling for him than Iowa.

[jimmy olsen]

Good Lord!
Trump only leading by 5 in Kansas
https://x.com/PollTracker2024/status/1850973817213984974?

[PJL]

If those two state polls are anything near accurate it means that it's more like Harris is about 56% nationally and Trump about 44%, a swing of about 5% in favour for Harris. Or to put it another way you're in 1992 UK election territory for poll credibility - i.e. shattered.

[Admiral Yi]

If those two state polls are anything near accurate it means that it's more like Harris is about 56% nationally and Trump about 44%, a swing of about 5% in favour for Harris. Or to put it another way you're in 1992 UK election territory for poll credibility - i.e. shattered.

Alternatively a lot of people recently made up or changed their mind.

[PJL]

If those two state polls are anything near accurate it means that it's more like Harris is about 56% nationally and Trump about 44%, a swing of about 5% in favour for Harris. Or to put it another way you're in 1992 UK election territory for poll credibility - i.e. shattered.

Alternatively a lot of people recently made up or changed their mind.

Well we should know for sure in the next couple of day when news polls come out. If they're still in dead heat territory than something is off. Even Nate has commented how the polls have been herding recently.