As we enter the final week of the election, a slew of last minute polls that will emerge. This is a good time to remind ourselves that we should not put too much stock in what they say. As I said in an earlier post, pollsters have to make adjustments to the raw data and this introduces systematic uncertainties so that the actual margin of error could be about double the statistical one.
Josh Clinton has done an interesting analysis to try and get a better idea of how much these adjustments can affect the results.
He says that pollsters have to address four questions.
After poll data are collected, pollsters must assess whether they need to adjust or “weight” the data to address the very real possibility that the people who took the poll differ from those who did not. This involves answering four questions:
- Do respondents match the electorate demographically in terms of sex, age, education, race, etc.? (This was a problem in 2016.)
- Do respondents match the electorate politically after the sample is adjusted by demographic factors? (This was the problem in 2020.)
- Which respondents will vote?
- Should the pollster trust the data?
To show how the answers to these questions can affect poll results, I use a national survey conducted from October 7 – 14, 2024. The sample included 1,924 self-reported registered voters drawn from an online, high-quality panel commonly used in academic and commercial work.
He shows that depending on the kinds of adjustments that the pollsters make to that same set of raw data, that can result in large swings in the results. He concludes:
Even though many people complain about how inaccurate preelection polls can be, it is actually astounding that the polls are as accurate as they are given how many choices a pollster must make.
I’ve shown that reasonable choices about how to weight a poll can produce up to an 8-point shift in the Harris-Trump margin. That’s a larger number than the “margin of error” and the expected margin in most battleground states. In a close election like this one, a pollster’s choices can radically alter a poll’s results.
Yet another reminder that we should not be getting agitated over small changes in polls over time or small differences between polls.
I did a course of statistics 40 years ago and I still recall nothing is very straightforward.
With all the assumptions mentioned in the link, it is obvious the concern about a tenth of a per cent here and a tenth of a per cent there is a waste of time.
I would of course have preferred if the evidence pointed to a huge SSADT defeat, but we cannot have everything. I might as well hibernate until Tuesday morning and save myself hypertension.
I’m not agitated by the changes, I’m agitated by the LACK of significant changes!
The lack of significant changes can be explained by the underlying message of the article; that the polling companies, the campaigns, the media, in fact everyone who is using the polling data has an interest in showing the polls to be close.
The media love to report on a close race, a blowout is dull unless the blowout is spectacular. But even if a blowout is predicted, I suspect the media would like to report the race as close until it becomes painfully obvious that one horse is furlongs ahead of the others. Since election results are not obvious to even some of the more math-literate citizens, it’s not hard to keep reporting an election as being close even if the election is not as close as it seems.
The campaigns of both of the major parties, like close races because it attracts greater donations and it builds enthusiasm for the get-out-the-vote efforts. If one side is expecting a blowout, it will depress voter turnout, and no one can predict which campaign will be hurt more. Some people on the winning side who expect the results are foreordained may choose to cast votes for alternative parties. Some people on the losing side will stay home because “why bother to vote when it won’t count anyway?”
The pollsters, well, if they report a close result, they are more likely to get repeat business then if they report a blowout. If the pollsters report consistently that the Democrats will get 63% of the votes, +/- 4% for margin of error, the number of polls requested and paid for will drop.
I’m not suggesting that any of these groups are deliberately shifting the polling results to make the race look close. But even an unconscious bias can have an impact.
My personal belief, which is not backed by the released polls, is that Trump has realized that he will lose the popular vote and the electoral college. A few weeks ago his behavior changed from optimistic to going-through-the-motions. He isn’t paying as much attention to how his message is playing with crowds, and isn’t carefully checking the messages from his sycophants. I suspect that there are unreleased polls which were held by the Republican party which shows Trump losing the election, and while these results have been shared with Trump they have not been made public. Some of that is reflected in the polling for more local politicians. The fact that Ted Cruise, among others, is facing a very strong challenge for his seat is an indicator that overall the Republicans are going to have a bad year (which, considering the nature of their candidates, they deserve).
I don’t think Trump’s rambling, repetitive, stories; or his repeating and apparently condoning the most outlandish rumors, are entirely due to age-related decline. I think he’s allowing that to happen in order to find scapegoats when he loses. He’s setting things up so he doesn’t get the blame for losing the election, and to keep his grift going. Of course, I’m not always right. That this is more of a feeling and not supported by much in the way of facts makes the likelihood that I’m wrong much higher. I hope I live long enough to see what the historians uncover.
The only poll that really counts is the one which happens next Tuesday.
I maintain that the pollsters are going to be very wrong again because they are incapable of overcoming the sexist bias in their data. Abortion rights is driving record turnout in this election among multiple demographics, but especially in the 18-30 categories.
The latest CNN poll on the early voters in 3 swing states is strongly in favor of Harris. Segment is discussed at the three minute mark in this news video.
https://m.youtube.com/watch?v=1CbxDlwRcc4