NC

Nate Cohn

Things Nate Says on Podcasts

“We are not forcing the party identification of our polls to be anything. I have no idea what the party identification of the electorate was in Texas in 2024 or 2026. If that was a D plus two electorate in our final poll or an R plus 10, we had nothing to do with that.” — Nate Cohn, GD POLITICS “I'm not sure there'd be a Times Siena poll in 2028 if the Republicans had won the Senate. That would have been the third systematic misfire overestimating Democrats out of the six elections we've done, on a scale that would have left readers feeling pretty misinformed. The costs of overestimating the Democrats would have greatly outweighed the risk of underestimating them. When you overestimate the side that wins, people don't care the same way.” — Nate Cohn, GD POLITICS “Anyone is welcome to propose a better name for synthetic past vote, especially because people seem to think it's an AI thing, which I definitely didn't anticipate. People are calling synthetic respondents AI-based, and people are confused about this. All it is is that we're correcting for a couple of things that we observe are clearly wrong in recall vote.” — Nate Cohn, GD POLITICS “We work for the New York Times. We're supposed to be an authoritative source of information, the truth. Polling can't really live up to that. So we always have to hedge. We don't know whether it's going to be right until the results come in. We've certainly been wrong before.” — Nate Cohn, GD POLITICS

Where to Find Them

Nate Cohn writes Why Is This Happening? The Chris Hayes Podcast . They have also been a guest on The Daily (NYT) (39 times) , GD POLITICS (2 times) and Plain English with Derek Thompson (2 times) .

Recently: “How To Poll A Wave, With Nate Cohn” on GD POLITICS (November 2026); “How To Poll A Wave, With Nate Cohn” on GD POLITICS (November 2026); “The Six Lessons of the 2026 Midterms” on Plain English with Derek Thompson (November 2026); “The Voters Who Flipped Congress” on The Daily (NYT) (November 2026); “A Landslide Victory for Democrats” on The Daily (NYT) (November 2026); “A Guide to Election Night 2026” on The Daily (NYT) (November 2026).

What They Said

“We are not forcing the party identification of our polls to be anything. I have no idea what the party identification of the electorate was in Texas in 2024 or 2026. If that was a D plus two electorate in our final poll or an R plus 10, we had nothing to do with that.” — Nate Cohn, GD POLITICS

Asked whether weighting on past elections leaves pollsters flat-footed when coalitions shift, Nate Cohn explains that the Times/Siena poll deliberately avoids forcing party ID to match the last election. He credits that choice with catching shifts among new voters.

GD POLITICS · 2026-11-12 Permalink → Listen →
GD POLITICS
Nate Cohn

…of our polls to be anything. I have no idea what the party identification of the electorate was in Texas in 2024 or 2026. I don't even remember what it was in our last poll, but we, if that was a D plus two electorate in our final poll or an R plus 10, we had nothing to do with that. That's just

Galen Druke

based on party identification.

Nate Cohn

I think that's a main, that's one critical difference. And another is recall vote. We use recall vote in a kind of loose way. We've branded it synthetic past vote because, you know, 20% of people are getting a different answer for our waiting purposes than what they actually reported in recall vote. But the bottom line is that this allows the recall vote in the electorate to differ and potentially significantly from the result of the last election. And I think that's also really important. If you were waiting the Hispanic vote in Texas by recall vote to be whatever your estimate was for 2024, I think you would have been way off. And we weren't.

Galen Druke

Yeah. How did you come to that conclusion? Because I think a lot of folks in the middle of the Trump era were saying, okay, we keep underestimating Republicans. We don't love the idea of waiting to recall to vote, which, you know, to lay it out for folks, I think a lot of listeners know, but is you…

Nate Cohn

So stepping back, the reason why we haven't followed the industry down this path is because when we try and use it on our election, our previous data, it doesn't help. If we, and I think that when other pollsters wait on recall vote, it does a lot more for them than it's done for us. I don't know…

Galen Druke

And so that's a change that you made between 2024 and 2026.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“I'm not sure there'd be a Times Siena poll in 2028 if the Republicans had won the Senate. That would have been the third systematic misfire overestimating Democrats out of the six elections we've done, on a scale that would have left readers feeling pretty misinformed. The costs of overestimating the Democrats would have greatly outweighed the risk of underestimating them. When you overestimate the side that wins, people don't care the same way.” — Nate Cohn, GD POLITICS

Asked what worried him most the night before the election, Nate Cohn separates the likeliest direction of a polling miss from what a miss would cost. A miss in one direction, he says, would have been far more damaging to trust in the poll.

GD POLITICS · 2026-11-12 Permalink → Listen →
GD POLITICS
Nate Cohn

…supposed to be an authoritative source of information, the truth. Polling can't really live up to that. So we sort of always have to hedge. We don't know whether it's going to be right until the results come in. Certainly have been wrong before. And going to this election, that was possible too.

Galen Druke

But like when you went to bed on election night eve, what was your biggest worry?

Nate Cohn

Well, I think that it's worth, to be honest, there are two separate worries. You know, one is like, what's the likeliest way the polls can err? And then the second one is what are the consequences if the polls err? You know, if the Republicans had won the Senate, and I didn't think that was going to happen, but if they did, like, I'm not sure there'd be a Time Sienna poll in 2028. That would have been the third systematic misfire and overestimating Democrats out of the six elections we've done and on a scale that would have left readers feeling like pretty misinformed about what's going to happen. And so, you know, the costs of overestimating the Democrats, if we had done so, would have greatly outweighed the risk of underestimating the Democrats, right? So even though I didn't really personally think that we were all that much likelier to underestimate Democrats than Republicans, like I do think that the risk that we overestimated Democrats weighed more heavily on me, given what the consequences of it would have been and what it would have meant for the way our work was perceived. You know, when you overestimate the side that wins, you know, I don't want to say that people don't care, but it's just not the same as if it seems…

Galen Druke

Yeah, I think that certainly makes sense. Overall, your polls showed Democrats winning the battleground states in the Senate by an average of five points. And that's what happened in the end. But it didn't happen in exactly the way your polls suggested. Ultimately, the Times Santa polls…

Nate Cohn

So I think we have a lot of visibility already into what went wrong for our polls in the three southern states you mentioned, South Carolina, Florida, and Texas. We have a lot of visibility into it because those states have already reported a great deal of data on the final turnout. We know in…

Galen Druke

Long time follower here.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“Anyone is welcome to propose a better name for synthetic past vote, especially because people seem to think it's an AI thing, which I definitely didn't anticipate. People are calling synthetic respondents AI-based, and people are confused about this. All it is is that we're correcting for a couple of things that we observe are clearly wrong in recall vote.” — Nate Cohn, GD POLITICS

Nate Cohn explains the Times/Siena weighting method called "synthetic past vote", which readers keep confusing with AI-generated survey respondents. It is only an adjustment to how people's reported past votes are used.

GD POLITICS · 2026-11-12 Permalink → Listen →
GD POLITICS
Nate Cohn

…in the electorate to differ and potentially significantly from the result of the last election. And I think that's also really important. If you were waiting the Hispanic vote in Texas by recall vote to be whatever your estimate was for 2024, I think you would have been way off. And we weren't.

Galen Druke

…when you're done waiting your respondents. Now, it sounds like as that has become increasingly a crutch for pollsters, you still, you're not, I mean, it sounded like actually in 2024, you were interested in relying a little bit on it, but it sounds like maybe you're more skeptical about that now.

Nate Cohn

So stepping back, the reason why we haven't followed the industry down this path is because when we try and use it on our election, our previous data, it doesn't help. If we, and I think that when other pollsters wait on recall vote, it does a lot more for them than it's done for us. I don't know if that's because we have higher quality samples on the front end or if it's because actually I kind of think that's what it has to be. I think other pollsters are just getting much more Democratic skewed samples on the front end. This is like the only thing that's getting them back in the ballpark and they're pulling the lever. That's just not what we see. You know, in 2020, our polls do not get better with recall vote. They get worse. It's in the APOR report. There's a whole section where they talk about an unnamed pollster. We're the pollster that is not helped by recall vote with the data that we provided to them for their reanalysis. Doesn't help us in 2022, doesn't help us in 2024. Nonetheless, though, after those three elections, we can see some patterns where, as you put it, recall vote maybe could have helped us a little bit. It could…

Galen Druke

And so that's a change that you made between 2024 and 2026.

Nate Cohn

Yes. And I should say, by the way, I think this is really important. I don't know if there are other public pollsters who made a change after 2024 that gave them results that they really liked for 2024. I haven't reported what our results are for 2024, but I'll just say like we would have been…

Galen Druke

Yeah, I think that's something that's so important to keep in mind. People are always like, oh, well, the polls got it wrong last time. It's like, well, a lot of pollsters, especially the most rigorous pollsters, are iterating and trying new things. Are there lessons that you've taken from the past…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“We work for the New York Times. We're supposed to be an authoritative source of information, the truth. Polling can't really live up to that. So we always have to hedge. We don't know whether it's going to be right until the results come in. We've certainly been wrong before.” — Nate Cohn, GD POLITICS

Asked which way he feared the polls might miss before the midterms, Nate Cohn says he saw evidence pointing both ways. He explains why a newsroom pollster always has to hedge.

GD POLITICS · 2026-11-12 Permalink → Listen →
GD POLITICS
Nate Cohn

…these races. Mississippi, we didn't pull, but other than that, we were prepared for Democrats to win all these, all of these little stats, like whether it was the biggest election for Democrats since Watergate and so on had ready before the election because we knew that kind of outcome was in play.

Galen Druke

…worried about, but that you thought they might be underestimating Democrats. This was speculation on my part. You know, I read your analysis, but like, did I get that correct? Were you more concerned that your polls down the stretch were missing Democrats than that they were missing Republicans?

Nate Cohn

I'm concerned about all of it. Look, this is the sixth election we've done Time CNN polls for. I've needed to count them 16, 18, 20, 22, 24. So this was the sixth. And in those elections, I've just really never, in hindsight, I've never really known whether they were going to be right or wrong. You can look at the Data and see evidence either way before the election. In this election, I thought we had pretty clear evidence both ways that we wrote about different points. The non-response element was obvious. Democrats were much likelier to respond to our polls than Republicans, which we've seen before. Sometimes that augered for a polling misfire, like it did in 2020. Other times it didn't, like in 2022. We could also see before the election that the Democratic turnout was just going to be better than our polls supposed. We saw it in Florida. We saw it in terms of the black share of the electorate in a state like South Carolina. We also saw it in our Texas poll where we had Democrats up by 12 points among registered voters, seven points among likely voters.…

Galen Druke

But like when you went to bed on election night eve, what was your biggest worry?

Nate Cohn

Well, I think that it's worth, to be honest, there are two separate worries. You know, one is like, what's the likeliest way the polls can err? And then the second one is what are the consequences if the polls err? You know, if the Republicans had won the Senate, and I didn't think that was going…

Galen Druke

Yeah, I think that certainly makes sense. Overall, your polls showed Democrats winning the battleground states in the Senate by an average of five points. And that's what happened in the end. But it didn't happen in exactly the way your polls suggested. Ultimately, the Times Santa polls…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

Collections They Appear In

Policy US Politics