I don’t recall seeing a definitive comment, but the folks at 538 do make reference to “correlated state error” in the published summary of their model. I assume that means state level correlation of |(actual results) - (polling indication)|, perhaps with some weighting/judgment of the polls and elections used in developing the matrix.
I think the analysis is that he’s favoring Trump for AZ, GA, and NC, and favoring Harris for MI, WI. Then PA and NV are toss ups. But if there are any surprises then it mostly breaks towards one of the candidates. If there aren’t any surprises then the state level outcomes are close to independent (notice the CDF is linear in the middle)?
I should add that I’m sticking with the guess I made up-thread: Harris gets NV and MI, and Trump gets the other swing states.
I’m assuming that Trump’s out-performance versus the polls still holds true this year (admittedly a risky assumption, given that several talking heads seem to think that might not actually be the case this year), but I deviate from the implications of that in Nevada, because there is an abortion measure on the ballot there and there’s history of blue folks over-delivering when such referenda are held.
I’m still hoping I’m wrong (and I almost certainly am in my added tweaks for faithless electors), but I’m sticking with that guess.
I’ll also add a prediction that Trump declares victory in Pennsylvania Tuesday night…but that the media doesn’t call PA before the weekend.
Here’s a cartogram on when AP called the state-level results in 2020:
Source: https://www.axios.com/2024/10/25/how-long-states-called-election
Maybe inside a state-space model?
I’m clearly not a Trump fan, but he doesn’t seem like the same candidate as in prior years. I’m wondering if Trump underperforms this year. I keep seeing stories about lots of empty seats at rallies this time around and people walking out. Listening to video clips of his speeches from the last few days he just seems half asleep.
I don’t remember that in prior elections. When you get down to it, Trump is in his third presidential election there people have been there, done that, just not as enthusiastic as the first one. People voting for the first time this year were six when he first ran for President. He’s been a politician running for president most of their lives. Similar for people in their early 30s.
ETA: whoops, messed up that math. Feels like forever, but only actually 9-10 years. I was adding on thel upcoming term
1 Like
He’s taken hand waving, sorry weaving, to a whole new level.
I don’t think that rally size necessarily correlates with the proportion of people who prefer a candidate, but do find it plausible that it correlates with enthusiasm and turnout. And I do think that having a successful rally that excites the attendees will make them more likely to vote and to encourage their friends to vote, while a weird, low energy rally will do the opposite. If this cycle’s rallies, combined with the Democrats having a good ground game and Trump having virtually none, gives Harris a 1-2% bump in vote count, that is huge.
Trump himself is weaker, but the Trump machine is much stronger (worse)
I’m not sure about the machine. I’ve seen repeated stories that their ground teams are a mess.
It also makes lying about it more critical.
2016 he had a ton of momentum at his rallies which were for the most part packed. And possibly the most hated women to conservative talk radio that the Democrats could possibly have run up against him.
2020 - Pretty sure we had lock downs and I don’t remember how many rallies he could actually do let alone whether they were “packed” of not.
2024 - His rallies seem less frequent than 2016, more meandering, less well attended, and some folks don’t seem to be staying all the way through this time.
What that means for the turnout and vote tomorrow, I guess we’ll know “soon”.
I’m hoping, but not expecting, a Harris blowout of such proportion that they actually call it for Harris tomorrow night which might go a long way to forestalling Trump’s inevitable challenges should he lose.
But what we’ll probably get is a close race that can’t be called for days or a week like last time.
1 Like
I think that’s pretty much in line with my expectations.
I’ve got one friend in the U.S. who is contemplating moving up here.
I believe that some of the midwest states won’t count early voting until after all the same-day votes are counted. It is also expected that those votes will skew more towards Harris than for same-day votes. Given that nearly half of expected votes are early (78 million so far, compared to 160 million for 2020), it might seem that Trump is doing much better in those states than ultimately for those states.
You new to this country??
That shit is on the low burner already.
1 Like
I can’t decide which is better/worse, a state releasing partial results or waiting to release until 99.x% is tallied. Both would have the whack-jobs saying “proof of manipulation” to anything not going their way.
I am of the mind that the outcome of the election versus polling results (assuming the polls are as accurate as the past two elections / assuming there’s no new bias due to polling firms being biased towards a toss-up, and ignoring potential court intervention) will come to:
- The propensity for Trump and Harris voters to accurately disclose their intended candidate; and
- The ability for the two campaigns to unreliable/unlikely voters to actually go out and vote
I’ll admit that my prediction/guess on the outcome assumes that the first point is unchanged from 2020, but…
- I can imagine Trump supporters may have become more likely on average to be willing to profess their allegiance, with the possible exception of folks in the next bullet.
- I can believe that some people who might otherwise be expected to support Trump, may actually vote for Harris but be unwilling to admit it (c.f. women being reminded they can lie to their husbands about who they voted for).
- I have no clue about whether pro-Trump folks’ disregard for mainstream media impacts their willingness to respond honestly to polling firms…or whether extreme election fatigue may be distorting folks’ willingness to respond to polling firms as compared to 2016 and 2020.
And of course, the GOTV consideration is a factor in any close election. Both sides seem very aware of the importance. I even heard some commentators observe that part of the reason that Trump has been doubling down on extreme comments may be that he is trying to excite and motivate folks in his core demographic that have never before voted to come out and do so (in contrast to Harris intentionally presenting a more moderate image, in an attempt to attract folks who would normally vote Republican, but are dissatisfied with Trump).
I’m not sure what the motivation for releasing ANY results before Wednesday, to be honest.
Scientific American showed me a graph of how votes have been counted over the past 24 years or so. let me see if I can paste it…
link to article:
I’m aware the shit show challenges real and imagined will fly. But if Harris win like 286 and it takes a week to decide his followers are going to be full frenzie mode by the time she’s declared winner. Where as if she’s in the 319+ area and a surprise state like say Florida gets called for her on election night, a good chunk of the country will say “oh there goes bat shit crazy Don-old airing his imaginary grievances again”
1 Like