The following is a lightly edited transcript of the October 16 episode of the Daily Blast podcast. Listen to it here.
Greg Sargent: This is The Daily Blast from The New Republic, produced and presented by the DSR Network. I’m your host, Greg Sargent.
The new Fox poll is absolutely disastrous for Donald Trump. It has fresh data on his dropping approval, on the widening Democratic lead in the battle for the House, and many other things. But what if the polls are wrong? Republicans are betting on this, and there’s now a lively debate about it underway in polling circles that can’t be ignored.
So we’re going to work through both the Fox poll and the debate over polling accuracy with Natalie Jackson, a polling expert who has her own firm, Centerline, and just wrote about the topic well, as a columnist for National Journal. Natalie, thanks so much for coming on.
Natalie Jackson: Thanks for having me.
Sargent: So just to establish something quickly: in your piece, you say that Democrats will win the House. You’re treating that as a foregone conclusion, not really in doubt, and that the only big question is whether Democrats have a big year or a really big year, as you put it, the latter meaning they win the Senate. Is that basically where we are?
Jackson: I think so, because it just takes so few seats to flip the House that it’s hard to imagine that not happening in this environment. Mike Johnson has to keep hope alive for Republicans, but the only way Republicans hold the House is if everything we know about this election cycle is absolutely, disastrously wrong. And that’s not happening.
Sargent: Right. You’re essentially saying that for Democrats to not have a pretty good night, at minimum, you’d need, like, a major systemic error that reaches into every corner of what we know right now, right?
Jackson: Right. We would need at least something like a 10-point error on the generic congressional ballot. We would need Trump’s approval reality to be far different than it is. It would just be a massive upset of everything we think we know.
Sargent: Right. So Fox News just dropped this poll that neatly underscores everything you’re saying here. The top lines are that he’s at a record-low approval of 36 percent, 74 percent rate the economy negatively, and Democrats lead by nine points in the generic ballot matchup, 54 to 45. If that proves correct, Natalie—and we don’t know that it will, which we’ll get into—but if it were, that would be the stuff of a wave election, right?
Jackson: Yes. This is kind of what I put in the good night. The really good night, I think, depends on the Senate polling, which I’m sure we’ll get into.
Fox’s numbers match up with mine. I ran a national poll just a couple of weeks ago. I had 35 percent approval among registered voters, eight-point Democratic advantage. So Fox is right there in line with what I’m finding and with what really everyone is. I believe the averages right now are around a nine percent Democratic advantage.
Sargent: Exactly. I mean, the Fox poll is in line with the New York Times poll, which is another gold-standard poll. It had Democrats leading by eight points in the generic ballot matchup.
And all the polling averages that you can look at, whether it’s Nate Silver or FiftyPlusOne, you find the generic ballot matchup to be at around eight or nine points, or somewhere in the middle there. I mean, that’s where we are right now, right?
Jackson: Absolutely. That is just what reality looks like. Nobody really disagrees with that.
Sargent: Well, OK, a little more from the Fox poll. It’s incredible stuff. Trump’s approval among independents is 21 percent, with 78 percent disapproving. Among Latinos, his approval is 21 percent, with 79 percent disapproving. He’s even underwater with non-college whites, which are critical to his base, and that’s at 44 percent to 55 percent.
Natalie, this is historically bad stuff. I mean, it does rival some of the worst presidential performances in the history of modern polling, does it not?
Jackson: It does. The only other president besides Trump in his first term that even comes close to this is George W. Bush when he was leaving office, really 2006 on. But he did not see numbers quite this low.
The averages were in the upper 30s, and the same with Trump’s first term. He did not really go below the upper 30s until January 6. And now we’re seeing numbers that rival, and even in a few cases go lower than, his post–January 6 numbers.
Sargent: Yeah. I mean, he is now rivaling the aftermath of January 6, and that’s just an incredible thing, considering that he essentially encouraged a violent attack on the seat of government.
But I want to point this out, too, as well. It looks like Trump’s approval is really translating directly into the generic ballot matchup as well. Democrats lead among Hispanics by 69 to 31, and among independents by 62 to 36. Democrats are in the 60s with Latinos and independents. And Democrats are doing better than usual with non-college whites. They’re at 44 percent to the GOP’s 55 percent.
Natalie, I think you can see a Democratic coalition kind of forming there. It includes winning back a ton of the Hispanics they lost to Trump in 2024, doing very well with independents, and keeping Republican margins down with blue-collar whites and rural whites as well.
That’s the recipe that they’re going to need in some of these tough states and districts, especially ones that Trump won by high single digits or even double digits. Your thoughts on all that?
Jackson: Right. The coalition that seems to be coming together is reminiscent of what we saw in 2020, when Biden won. We saw heavily Democratic Hispanics, heavily Democratic independents. It’s a really similar outlook. And absolutely keeping those margins down in the rural areas, among the non-college voters. That’s what Democrats have to do.
The issue that I would point to is they have to be able to do it long term. This is great for Democrats for 2026. But what we’ve learned in the last few election cycles is that Hispanics are really dealigning. They’re not really that attached to either party. It used to be a more reliable Democratic core.
Now, especially among the younger generations—which Hispanics are, they do tilt younger than the rest of the population—they are not attached to either party. And they are also very heavily affected, directly affected, by the economy, by virtue of where they live, the types of industries they’re in. And so this is a group that’s going to swing, and we’re seeing a really hard swing against Trump.
Sargent: I mean, it’s going to be a big challenge to try and lock down some of these demographics over the long term. So you’re absolutely right about this.
I want to get at what the Fox poll shows about the economy as well. Donald Trump’s at 22 percent to 77 percent on inflation. Twenty-two percent on what’s clearly the biggest issue on the minds of voters right now. That’s barely more than one out of five that approve.
And by the way, Fox also breaks out non-MAGA Republicans versus MAGA Republicans, and on his overall approval, he’s actually underwater with non-MAGA Republicans. That’s 47 to 53. Those are GOP voters.
And I’ve got to think that looms large in some of these Senate races where Democrats have to win some Republican voters. These are states where Trump prevailed by double digits. And the only way Democrats can get there is by winning a chunk of Republicans. So if you’ve got non-MAGA Republicans disapproving of Trump, there’s a path for them, right?
Jackson: There is a path. The approval numbers are one good indicator. Democrats really need two things to happen with these non-MAGA Republicans. They need them to stay home—some portion of them just don’t show up to vote. The other thing is they need them to be not super attached to the Republican Party.
There’s a contingent of Republicans that is very unhappy with Donald Trump. But on the other hand, they would—I’ve heard this said on cable news a couple of times—they would rather put their face in a wood chipper than vote for a Democrat. So it’s really a complicated picture, and we just don’t know what choice those non-MAGAs are going to make.
Sargent: Absolutely. Well, let’s get into your piece on how to understand the debate over whether the polls might be wrong. And again, we stress that’s always possible. The data could be wrong.
As you write, the big question is whether the polls are undercounting Republicans who aren’t that ginned up to vote, but might still end up turning out anyway and voting Republican. And maybe pollsters aren’t reaching those people.
But a lot of pollsters have taken steps to try to correct for this problem. Can you just walk us through what those steps are?
Jackson: Yes. So, brief background: In 2016, in key states, polls underestimated Trump support. Trump ended up winning. In 2020, we thought Joe Biden was going to win very easily. It turned out polls underestimated Trump even more broadly that year. Biden still won, but it took us four days to learn that after the election. It was so close.
In 2024, things were generally better. There was a slight overestimate for Harris winning. But the track record is that pollsters struggle with Trump support. It’s been better in midterms, with lower turnout different electorate.
But what we’ve done in the meantime is, we weight our data to education now. So that might seem like something fairly obvious, but it was not a practice that pollsters did before, because we didn’t have a precise number for the number of people with college degrees in the electorate.
So without a solid benchmark, we didn’t weight to it. That used to not matter, but people have polarized by education. College-educated people are more likely to be Democrats now, non-college more likely to be Republican. And so if we miss the balance on education, which we were doing, it skews toward one party or the other, and it was skewing toward Democrats.
We also started weighting more on partisanship and partisanship indicators. Some pollsters will use the prior vote. So they’ll ask, for example, who did you vote for in the 2024 election? And then they’ll make sure they have the proper number of people who voted for each candidate.
That’s tricky for a midterm, because your midterm electorate is very different. It works well when you have Trump versus Biden part two, or Trump versus Harris, who is essentially a representative of the Biden administration.
So we’ve done a lot of adjusting to the methods behind polling. There is an element that we’re always fighting the last battle, because we don’t know what is missing or what might be going on until the election happens.
Sargent: Well, OK. So looking at some of the polls that we consider to be high quality, like the Fox poll or the New York Times poll—they’re doing this weighting stuff, and they’re also trying to correct for the problem of not reaching Republicans. Those two things are related, I get that.
Can you talk about the fact that Fox and the New York Times are taking steps here to do what you’re talking about? Like, do you feel reassured? Can we look at their data and say they’re really trying to get to the right place on it?
Jackson: I believe they are. Fox and the New York Times are among the top media pollsters out there, the top any pollsters out there, honestly. They are very careful to make sure that they speak to enough Republicans. What I was talking about was adjusting the data on the back end, but those adjustments only go so far if you don’t talk to the right mix of people in the beginning.
And so one of the things that Fox and the New York Times do very well is they make sure on the front end that they have talked to a good mix of people. And those things take effort and they take money, which is why a lot of pollsters don’t do it. But these two are very reputable.
Sargent: Right. Fox and the New York Times both have the money. They are putting the resources in. It would be really, really a disaster for Fox to undercount Republicans, I think, given obvious reasons like who their audience is. And the fact that they’re constantly attacked by Donald Trump, by the way, I think that actually is a factor that could be playing in favor of accuracy.
Fox knows that they’re really on the hook big time if they screw this up. So they’re really going to put in the effort to get to those Republican voters, aren’t they?
Jackson: They are. And I have tremendous respect for Fox’s polling team, because you can imagine the internal and external pressures that they feel to show numbers that look better for Republicans than what they’re showing right now.
Sargent: I mean, Trump calls for the firing of the Fox pollsters pretty much every day at this point. Let me ask you, there’s also a scenario where Democrats could be undercounted. How would that happen?
Jackson: It’s always possible that error runs in either direction, you know? So, in our adjusting and trying to make sure that we’re getting enough Republicans, there’s always the possibility that we’re adjusting too much.
That’s kind of the thing that we don’t know: do these Republicans actually stay home? Are we going through all these hoops trying to get them to answer the surveys, and then they don’t actually vote? That would be where error would come from, making Democrats have an even better year than what we’re looking at.
Sargent: And is it possible that pollsters could somehow miss the level of energy among Democrats? Like it’s not an exact science to figure out how to screen the polling questions to try to be sure that the people who say they’ll vote are actually going to vote. Right? And so is there a scenario where you kind of undercount the number of Democrats who are actually going to show up? Is that a possibility?
Jackson: It absolutely is. We are trying to guess who’s going to vote. This is the only time that we don’t know who we need to talk to for sure. If we’re talking to registered voters, great, we know who that is. We can take a good sample of that. All American adults, great. We know who they are. We know who we need to talk to.
Likely voters don’t exist as a principle. The election hasn’t happened yet. So we are trying to cobble together several other things that predict how people vote. Some of that comes from using actual vote history. A lot of pollsters use the voter file and get their respondents’ information off of the voter file, which obviously tells you whether they voted or not the last time, the last few elections.
Some of us aren’t using voter files. So we ask a series of questions. How likely are you to vote? Some people say I’m certain to vote. Some say I’m probably going to vote. Very few will tell you straight up, I am not going to vote.
So that makes it tricky in a midterm, where turnout’s likely to max out 50, 55 percent usually. We know when we get 70, 80 percent of people saying they’re going to vote that it’s overshooting somehow. So when we try to cut that down and make it look more like the electorate, it is possible to trim too many Democrats out.
Sargent: That is just fascinating stuff. All right, just to wrap this up. What’s your best guess as to the range of outcomes in terms of polling error here? Is it your sense that the polls that we’re seeing now are pretty close to accurate?
Jackson: I think what we’re seeing from places like Fox and the New York Times are pretty accurate. I think some are kind of out there. I am less concerned about the House generic ballot. I think, you know, at most we’re looking at maybe two points, two to three points, in either direction.
Some of the Senate polling I’m a little more concerned about. Some of the specifics of each state are a little bit harder to deal with. Some of the polling that we have is not very high quality. So I do think there’s potential to see two to four points, maybe even more, in some of these Senate races off of where the polls are.
Sargent: And obviously those could flip the races in one direction or another. It looks to me as if the eight-to-nine-point generic ballot matchup kind of lines up with what the Senate polls are saying, because if you get to eight or nine points in the generic ballot matchup, you’re probably winning a bunch of those Senate seats in even Trump double-digit territory, right?
Jackson: Probably. I worry a little bit about some of the specifics in these states, and the candidates, and just the little things that can come into play in an individual environment. But overall, yes, things are very good for Democrats. And no one is arguing that Republicans are going to come out of this ahead.
Sargent: Natalie Jackson, thank you so much for all that. It’s really fascinating. Thank you so much for coming on.
Jackson: Thanks.