Member Newsletter
Over the weekend, Democrats or Dem-adjacent persons watching polling of the 2024 presidential election got knocked over the head with a metaphorical anvil: a batch of polls collectively showing Joe Biden 2 to 4 points behind Donald Trump. I’ve gotten a lot of questions about these polls and polling generally, ranging from the technical, to the what does it mean, to please talk me off the ledge. So I wanted to try to address them here.
First: Are these polls accurate? In an age when no one answers their cell phones let alone landlines, how do we know whether these polls are representative. Who has a landline? etc.
This is a complicated question. Without getting into deep technical details, yes, the pollsters definitely get that landlines are old news and most people don’t even answer unknown numbers on their cell phones. The same applies to text requests for political surveys. Response rates — or, rather, non-response rates — are awful. But pollsters know all of that and they’ve come up with pretty smart ways to deal with it. Without getting too far into the weeds, it comes down to increasingly sophisticated ways of modeling the electorate, using those models to weight the results, and in so doing backing out a representative sample from the data.
Overall that seems to work pretty well. Polls usually get at least close to the correct result. The wild card is that this kind of modeling involves making a lot of assumptions about the voting electorate. As response rates have gone down the roll of this kind of modeling, to get to a representative sample, has gone up. Since you’re asking, no, the assumptions aren’t just what the pollster pulled out of the air. They’re assumptions based on recent elections, demographic data, etc. The good pollsters aren’t idiots. But there continues to be an important X factor of those assumptions and shifts in voting behavior can catch them off guard. The answer is that the good pollsters mostly know what they’re doing. Mostly.
Second: Okay, fine, Josh. So do I need to believe these polls or not?
Mostly, yes. You need to believe these polls. But we need to break down what we’re talking about. People often say polling right now isn’t the same as this fall. But it’s not just that the election’s eight months from now and things can change over eight months. Public opinion just functions differently in the weeks before a national election than it does eight months before it. It becomes clearer who is and who isn’t going to vote. People answer polls differently when they’re about to have to make a choice than they do months in advance.
Because of that I always find it a little difficult to answer questions about whether you should “believe” the polls. People say, well, this is just a snapshot. So it’s accurate for right now. But maybe not later. But now there’s not a general election. So in a real way, “now” isn’t even a thing.
With that said, I would absolutely much prefer Biden to be ahead by a couple of points than behind by a couple of points. That is 100% true. The best way to look at the current situation is that right now Biden’s the President and Trump is just some guy. Currently the election is a referendum and Biden is losing. The question is whether the Biden campaign and the dynamics of the election itself are going to change the election into a choice. Which of these two guys do you prefer? I think Biden will likely win a choice election. So whether this election is a referendum on Biden, or becomes a choice election between Biden and Trump, is the big question.
There’s another issue here on believing the polls.
From 2016 until 2022, in most elections you’d have the polls and then Trump or his candidates would outperform the polls. Not by a lot. But by a non-trivial amount. Often this difference was within the margin of error. So a pollster could say, we didn’t miss anything. That’s within the margin of error. It’s right there on the product label, as it were. But when it’s always or usually in one direction, that’s not the margin of error. There’s something wrong. And when you’re aggregating together large numbers of polls, the margin of error works differently. Then in 2022, pretty much after the Dobbs decision, that seemed to flip. Democrats started doing a nudge better than the polls.
We saw that in 2022. We’ve seen it in various special elections. We saw it in 2023. We haven’t yet seen whether that applies in a general election. That’s a big caveat. General elections are different. But if you’re looking for a reason that polls might be off by a bit and in Democrats’ direction, that’s it. Personally, I do think this is the case. And it’s the prism through which I look at a lot of this. I am not relying on this. But it is part of the picture when I look at these numbers.
Third: Okay, great. But isn’t the media awfulizing a lot about Biden? Obsessing about Biden’s age?
This one is complicated and the issue of media bias is one mostly best left for a different post. But there is some truth in this. It’s very paradoxical. I’ve written many times that D.C. remains wired for the GOP for all the reasons I’ve described. But it’s also true that there’s this kind of obsessing and negativity because most reporters are demographically and often politically closer to Democrats. For many of them it is kind of a given that Trump is a nut. And thus they are prone to focus on and amplify Democrats’ agita and anxieties.
Right now, Biden’s behind. Not by a lot. But he’s behind. That’s bad. That sucks. But even within that there’s a filtering of news in place that focuses on bad news for the Democrats. Let me give you just one example from today. Just after that batch of bad polls (NYT/Siena: Trump +4; Fox: Trump +2; CBS: Trump +4; WSJ: Trump +2) there was another batch of three. Those polls were Reuters: tied; TIPP Insights: Biden +1; Morning Consult: Biden +1.
On balance, the bad-for-Biden set includes pollsters I rate a bit better than the new, better-for-Biden set. But these are all quality polls. I don’t think I would have heard a peep about these other polls if I didn’t follow every poll that comes out. It’s also true that if you averaged all seven of those polls Biden would still be behind. But it is still unquestionably true that the bad polls are getting way, way more attention than the better polls for Biden.
We also see this a lot when reporters pick out a subsample in a poll as an example of Democratic oblivion. Trump is tied with African-American voters. Fifty percent of voters totally hate Biden. It’s almost never a good idea to base reporting on these internals of a poll. I actually recently saw a headline of a Quinnipiac poll which said that 70% of voters said Biden was too old. But he was actually up in that poll by 4 points.
There are more anecdotal examples. Over the weekend I was checking in on the number crunchers I follow on Twitter and Dave Wasserman, a great elections numbers guy, noted that Trump came within 40,000 votes of winning in 2020 and now, with Biden less popular than he was, Dems were “on the brink of disaster.” After posting that he caught some grief, I think, for speaking in such hyperbolic terms, and he followed up to say, “yes, losing to a candidate whom 53% of voters believe committed serious federal crimes would qualify as a disaster for any political party.”
Many would agree that Biden losing by even a single vote will qualify as a disaster. But Wasserman is a numbers analyst, not a partisan. He wasn’t talking about the substantive impact of a Trump presidency. He was talking about one party potentially losing. I note these comments because there’s a clear and consistent tendency to talk in really, really hyperbolic terms about Biden being slightly behind eight months before the election. This is real and it affects the overall tone of election coverage a great deal.
It’s worth remembering that Trump ran behind basically for the entire 2020 election and no one ever talked like this about his campaign. It’s just built in. It’s worth being aware of.
Fourth: Okay, Josh, you’re the man. Just give us some wisdom.
My take on all this is that Biden is behind now. Not by much. But it’s been consistent since the fall. That is 100% not great. He’s being hurt by the referendum/choice issue I mentioned. He’s being hurt by the Israel-Hamas war in a way that overlaps with a more general discontent among young and disaffected members of the Democratic coalition. There is a lot that remains in flux and I remain cautiously optimistic that Biden will win, because of 1) various fundamental factors about the economy, 2) the trend of the last two years in actual election results, 3) Democratic strength with high-propensity voters and 4) at this moment there’s a grass-is-always-greener effect for a slice of voters. But as they’re reminded of who Trump is, I think some of those people will shift. Small numbers, but some.
Am I not worried? No, I’m worried a lot. I would VASTLY prefer it if Biden were running ahead. But the margins are close. There’s a lot still to happen. And the race remains very winnable. I’m cautiously optimistic for the reasons I noted above. That’s my take.