As a person who writes about politics for a living, I am often asked for my thoughts on the subject by friends, family, or people I’ve just met. Sometimes they’re interested in my opinions (“What do you think about Tim Walz?”), sometimes they ask for predictions (“Who’s going to win?”), and sometimes they want reassurance (“Trump isn’t going to win again, is he?”). Inevitably, the release of the latest New York Times/Siena poll — which has often found results slightly more favorable to Trump and Republicans than other polls — produces panicked texts from loved ones, as the latest one did when it showed Donald Trump ahead of Kamala Harris by 48-47%.
So this would be a good time to remind ourselves of the best way to understand the hundreds of polls we’ll be seeing over the next two months, so we can maintain our calm. Here are some things to keep in mind.
There are many polls. The Times/Siena poll is one of them. It’s not worth getting worked up over. In poll averages right now, Harris leads Trump by a couple of points; for instance, 538’s average puts Harris ahead by 2.8 points, while DecisionDeskHQ puts her ahead by 3.3. But no poll gets as much attention as the Times/ Siena poll, for two interconnected reasons. The first is that like every news organization, the Times hypes its own polls, so if you read the Times, you’ll see lots of material about its poll.
And even in our era of informational chaos, the Times retains a unique agenda-setting power over the entire news media. Every TV news producer and every editor at major news sites reads the Times every day, and inevitably takes at least some of their cues from it.
So when the Times does a poll, it’s bound to get more attention than others, especially — and this is the second reason — if it shows something different than other polls do. If it shows better results for Trump, liberals will lose their minds with anxiety, sharing and spreading it around.
But two polls that show almost identical results can make us react very differently depending on where they land relative to the 50-yard line. One poll saying Harris is leading 51-49 and another saying Trump is leading 51-49 are saying basically the same thing (considering the margin of error), but there’s a psychological effect of seeing the candidate you hate ahead. So be aware of your own reaction and how it might not be completely rational.
The big picture has a margin of error. As a savvy news consumer you know about the margin of error, or what statisticians call the confidence interval. Produced via a simple formula involving the size of your sample, it tells you that while you got a particular result, the “real” number in the population is probably somewhere within a few points of that result, but we can’t know exactly where. So if your poll says Kamala Harris is at 48 percent and the margin of error is +/- 3, she’s probably somewhere between 45 and 51.
As you aggregate polls, the combined sample gets bigger and bigger and the margin of error gets smaller and smaller, which is why poll averages are more helpful than any single poll in understanding where the race stands. Roughly speaking, a poll with a sample of 1,000 respondents has a margin of error of about +/- 3, but if you combine 10 of those polls together to get a sample of 10,000, the margin of error goes down to around +/- 1.
But built into that formula is a bunch of assumptions, most importantly that every poll is a perfectly random sample of the population, which of course it isn’t. Pollsters ask questions in different ways, weight responses to account for unrepresentative samples in different ways, adjust their results based on unique models of who’s going to vote, and have different techniques for reaching people and getting them to agree to participate. So what’s important to take away is that even aggregated numbers still contain within them a lot of “error,” which is the term researchers use to refer not to mistakes but to the broad category of “everything that’s going on that we don’t know.”
But guess what: Even savvy news consumers don’t need to know exactly what the true margin between Harris and Trump is. As far as we can discern it, the truth is that the race is very close and Harris seems to be ahead by a couple of points nationally, but we can’t get much more precise than that.
Tuesday’s debate probably won’t produce much of a change in the polls, and any change it does produce will probably fade within a few days. That has always been the case: We in the media act as though any debate gaffe or zinger could transform the race, but it almost certainly won’t. It’s not completely impossible — for instance, Trump could have a meltdown as spectacular as Joe Biden did in their debate, and that could cause a real shift — but in all likelihood, a week or two from now the race will probably be close to where it is now. Debates are still meaningful, but not because they’re likely to cause a shift in the polls.
Feel free to ignore any and all predictions. This is a hard one, because it’s in our nature to want to predict the future. But if someone is telling you that as of today Harris has a 57% chance of winning, don’t even bother with them. That attempts to impose imaginary precision on an inherently unpredictable course of events. Likewise, there are a bunch of competing models that purport to tell you, based on three or five or thirteen variables like wage growth and whether we’re in a war, who’s going to win. They’re fine to read as a matter of curiosity, but they’re all imperfect; yes, if the economy is good the incumbent usually wins, but knowing that only gets you so far, especially in an unusual year like this one.
If you asked me to, I could make a plausible case that Harris is going to win: She has more money, she has more volunteers, Democrats have a better ground operation, Trump is undisciplined and will continue to undermine himself, abortion continues to be an unusually strong motivator of votes, gas prices are falling, when the Fed lowers interest rates it will make people more optimistic, et cetera, et cetera. But it would still be a guess about an unpredictable future.
Things will change. But one thing will continue to be true. When you ask yourself (or anyone else asks you) where we stand, here’s what you can say: The race is close. It will probably be close all the way until election day. Polls are interesting, but they won’t resolve that uncertainty. It’s our fate to live in a chaotic universe.
And if the anxiety gets too much, you can try this:
I am not a big polling person. I say it’s better to be leading than not but keep working. Nothing is ever for sure.
Polls aren’t reliable. Stay focused on getting out the VOTE. When Dems vote they win!
I will say too that I wore this Kamala tee out this morning and got smiles and lots of "where-did-you-get-this?" 👇
libtees-2.creator-spring.com/listing/votek
The enthusiasm surrounding this campaign is at levels NEVER seen before!
There's a possibility that Trump may lose his home state 🤣
Brilliant!