Suppose that, due to global warming or any other factor, the predicted snow accumulation becomes, on average, an inch too high. An expert can correct me if wrong, but I think the weather forecasts will be changed to predict an inch less.
Polling is a little different.
The predicted Trump November election result has twice beat final polls. Ethical pollsters cannot just add in three points to Trump’s numbers. They can try adding in another control, such as normalizing on education level. But they did most of that after 2016, and Trump still beat his polling average in actual November 2020 voting.
From what I have read, the reasons why Trump beat his polling averages, in November of 2016 and 2020, remain uncertain. A sample size of two elections is so small that one cannot be sure it will happen again. But I have not read anything indicating pollsters learned some big lesson and changed their methods accordingly.
P.S. I will give you this much, in terms of similarity between polling and weather: Two elections, or two snowstorms, are not enough to justify changing the forecast model. Therefore, although I consider Biden an underdog, I wouldn’t think Trump has more than a slight probability, above 50 percent, of winning.
Then you’re idea of an “underdog” is different than mine. Not a problem – just different! (For me, the less-likely-to-win person or team starts being an “underdog” when they have less than about a 35% chance of winning).
Answering the last post: Infinite. Reputable pollsters cannot ever put in a fudge factor where they add points to a candidate. If they did, polling would be biased.
There is no party preference skew problem. Per 538:
So the problem, if not just a 2016/2020 coincidence, is Trump-specific. If he loses this November, maybe he’ll be nominated again in 2028, but, IMHO, that would be the end. So, at most, pollsters could underestimate his support a total of four times. We’d just have to accept that they were imperfect four times, if it worked out that way.
There is not really data on any populist candidate of the Trump type. All the ones before were third party populists. So I think the base is so faithful that there is a large preference for them to answer polls. Vs Biden voters. “Are you excited about president Biden?” No. “Will you vote for him?” Yes.
Maybe? I’d even say Probably. But polls have been missing a lot on special elections since Roe was overturned. And the underlying problems re: cell phones and selection bias haven’t gotten any easier to solve. Underlying demographic trend may have gotten stronger or weaker, it’s hard to say.
But after this last batch of NYT/Sienna polls (almost entirely bad news for Biden) I think I’ve seen enough to feel that even if polls happen to be wrong (and wrong in the opposite direction they’ve been wrong the last two times Trump ran) Biden is still almost certainly behind at this point. He really does seem to be hemorrhaging support amongst Hispanic (oof those Nevada and Arizona numbers) and young Black voters (bye-bye Georgia) as well as younger voters in general.
If the election were today and these were final polls I’d think Trump would be >75% to win, with Biden needing a polling error similar to what happened in Trump’s direction in 2016.
Maybe you meant WI and MI (along with PA)? That is likely the easiest path (hold the rust belt while losing the Southwest). AZ and GA look very tough, but aren’t strictly necessary. Biden can lose NV, AZ and GA from 2020 and still squeak it out.
I think the number of Trumpsters answering polls seems to fluctuate, whereas the democrats they manage to reach stays the same. The graph at the bottom of this had the two candidates tied a few weeks ago.
Now Trump has a slight increase. Probably Trumpsters eager to say Trump has been wronged in court.
After the State of the Union address in early March, Biden’s poll numbers were gradually drawing even with Trump’s. By mid-April, the two were at a statistical dead heat – 0.2% (yes, zero-point-two) percentage points apart in the RealClear Polling average.
Then, after several months of keeping hard-right Rasmussen out of their polling average, RealClear decided to bring in a Rasmussen Trump +8 poll. After that, the “hush money” trial started, and it looks like that woke up some right-leaning poll respondents. Meanwhile “liberal” outlets like CNN and the New York Times rush breathlessly to report about how much trouble Biden is in right now.
On The 2024 Election - I’ve been diving into polls and the election quite a bit in the last few days (here, here, here) so just two points today:
in the last few weeks we’ve seen Biden’s best national polls of the year, and as I write in my posts, Biden’s emerging strength with LIKELY VOTERS is a very important new development in our understanding of the 2024 election (my caps - b).
For Trump to be winning the 2024 election he needs to be clearly ahead, outside the margin of error, in MI, PA or WI. If President Biden wins these three states, assuming we also gain the single Nebraska electoral college vote (likely), we hit 270. In the new NYT poll Trump isn’t ahead outside of the margin of error in any of these states among LIKELY VOTERS(my caps - b). Thus, in this poll, despite the headline and framing, Trump is not ahead or winning the election.
Until I am shown otherwise I assume Rosenberg knows what he’s talking about. Pity it doesn’t fit the “Biden is in huge trouble!” narrative the NYT is pushing.
It’s more than that actually. I shared this in the President Biden and the Runup to the 2024 Election thread so I won’t go into it here other than to say Sulzberger and the Times have a bug up their ass about Biden and his team. It looks like the bad feelings are mutual.
Let me say this plainly: The Times poll is intentionally Trump- and conservative-leaning, both in how it was structured and how it was promoted. There may be fine reasons for this, but they are not discussed anywhere. And its data contains some eyebrow raising anomalies that it either glosses over or ignores completely.
Like Rosenberg above, Kuo rails on the registered voters vs. likely voters thing:
What this headline didn’t mention is that the same poll looked at both registered voters and likely voters, but the headline ran with the numbers out of its registered voters sample.
Reading further into Kuo’s 5/14/2024 article, regarding the recent NYT/Siena poll data set:
There’s another problem with the sample set: It’s heavily biased toward conservatives. In fact, a full 80 percent of its respondents described themselves as something other than liberal, with nearly twice as many conservatives answering:
ANSWERS [ALL STATES]:
Very liberal: 10%
Somewhat liberal: 10%
Moderate: 39%
Somewhat conservative: 19%
Very conservative: 17%
Don’t know/Refused: 5%
Was this oversample of conservatives on purpose? Any mention of why there’s a conservative lean by the Times? Nope.
EDIT: Still more.
Election observers poring over the details of the polls were quick to point out some rather ridiculous conclusions. I should first note that every poll has some amount of bad data in it, and we shouldn’t ever draw broad conclusions from small sample sets. But when a poll is riddled with data that defies reality and all probability, we need to at least take note.
Take Pennsylvania. The NYT poll has Biden winning Philadelphia by 54-30, which is a big drop from how he did in 2020, when he won 81.4 percent of the vote there. That means either that there is a massive collapse in Black voter support for Biden in just four years, or that the poll is not picking up the right data. And in the suburbs, the NYT poll shows a suburban shift to the right since 2020, despite trending to the left in every election since then, including just a few months ago.
We can also look at Nevada. The NYT poll has Biden actually losing Clark County, a Democratic stronghold, by eight points. As Democratic consultant Tim Hogan noted, “That is not going to happen. Insane that outlier numbers are driving so much coverage.” Reporter Jon Ralston, who crunches Nevada numbers for a living and is my go-to guy on this, remarked that something is amiss in these Nevada numbers. He tweeted,
I hesitate to post more on NYT/Siena poll, but regionals are bizarre:
Trump up 8 in Dem bastion of Clark where Biden beat him by 9 in 2020. 17-pt swing!
In Washoe, where Biden won by 4, Trump is up 14! 18 pts!
No problem looking at it, but using that as the headline number, vice likely voters, is a choice they’re making, and they’ve consistently made such choices to portray the race in a particular way.
I wouldn’t get too hung up on those cross-tabs though. Those are likely very small samples, so you can easily get big swings like that and the error bars are very large.
The best way to deal with these things, generally, is to average them all out and then slap some error bars on it. Understanding that polling errors happen, and perhaps taking some comfort in the fact that since Dobbs most of the errors in special election polling have shown them underestimating Democratic strength.
I think, historically, most pollsters don’t switch to likely voters until later in the cycle. Having a pollster do both and then emphasize the RV number is a bit odd, but I can see why they would do it to drive clicks.
It’s really an odd election, since both candidates are basically incumbents and both are disliked at historically-high rates.
I think the one thing that Biden has going for him is Trump’s mierdas touch when it comes campaign spending and the RNC.
Much of the money collected by Trump and the RNC for the campaign is going to pay for Trumps legal bills and into Trump’s pocket. This will of course reduce the amount that they spend on the campaign, but furthermore will reduce the willingness for donors to give money.
Also with Trumps take over of the RNC they fired a whole bunch of key staff who had decades of experience running campaigns, analylzing data, planning strategy and getting out the vote.
So I suspect that when it comes to political campaigning, Biden will have a distinct advantage. Unfortunately given the extent to which everyone has already made up their mind, the actual campaign will be less influential than in years past.
