Chapters
Shownotes Transcript
You're a podcast listener, and this is a podcast ad. Reach great listeners like yourself with podcast advertising from Lipson Ads. Choose from hundreds of top podcasts offering host endorsements, or run a reproduced ad like this one across thousands of shows to reach your target audience with Lipson Ads. Go to LipsonAds.com now. That's L-I-B-S-Y-N-Ads.com.
Microsoft changed my name to Gallon, like a gallon of milk, my entire childhood. It stopped. I don't know who got to them, but one of the other Galens. I was just going to say, I had a colleague at Pew who was named Galen, who also did a lot of AI work, and so he may have done it for you. You know, Galens of the world unite. Hello, and welcome to the FiveThirtyEight Politics Podcast. I'm Galen Druk.
Tuesday is Election Day in Ohio, and it's a bit of a convoluted election day at that. Ohioans are voting on whether to increase the threshold to pass constitutional amendments from a simple majority to a 60% supermajority. The initiative would also expand the signature requirements for getting an amendment on the ballot in the first place.
It's a technical question, but it's attracted tens of millions in spending. And if you watched the ads on both sides, you'd think it was really about culture war issues. And that's because, by proxy, it is. In November, Ohioans will vote on whether to codify abortion rights into the state constitution.
But if Tuesday's ballot measure passes, it would in turn make that amendment less likely to pass. So we're going to take a look at where the race stands and the broader trends of ballot measures like this.
Also today, we're going to dig into the New York Times' first polling of the 2024 primary and general election. Their surveys with Siena College during the 2022 midterms earned them the distinction of the best pollster in the country in our ratings. At this point, their early data suggests that Trump is far outpacing his rivals in the primary and is tied with President Biden in general election polling. Extremely early general election polling. So what should we make of that?
Here with me to kick things off is senior elections analyst Nathaniel Rakich. Welcome to the podcast. Hey, Galen. Thanks for having me. It's good to have you. Also, leader in the show, Ruth Gelnick of The New York Times, will be joining us to talk about their latest polling data. But Nathaniel, let's dive right into Election Day.
What does the polling show in terms of how competitive tomorrow's election in Ohio is going to be? Yeah, so the polling is a little bit up in the air. And I think that reflects kind of how, as you mentioned, kind of convoluted the campaign is, how the messages are kind of all over the place, how it's kind of an obscure procedural question when you really drill into it. So there have been three polls of the race that have been released so far, and it finds
when you average them, they find that 35% are in support of issue one, which would be pro-raising the threshold to 60%. 45% are opposed and 20% aren't sure. So obviously that 20% aren't sure is pretty high and I think reflects that confusion and that
is ultimately going to be what decides the election. The thing is, of course, you have to remember with ballot measures, there is this thing called status quo bias, which is that basically when people aren't sure about ballot measures, they tend to default toward the status quo, which is usually no. So they're like, you know, if I'm not sure, I better not change things because things are going okay the way they are or whatever.
So I would imagine that, you know, between the fact that there are already 10 points more people opposing the measure than supporting it, and then there's also this 20% aren't sure, those people would have to break strongly for the measure in order for it to pass. So I personally think that issue one probably will not pass on Tuesday.
And that average does actually somewhat obscure, though. What we're seeing is that there's a lot of disagreement in the polls. And I think it's in part because these kinds of elections are very rare. In fact, this is the first August statewide election on an issue in Ohio in nearly 100 years. So how did this get on the ballot in the first place?
Yeah, so basically it got on the ballot because Republicans in the legislature saw that the abortion rights amendment that you mentioned was probably going to make the ballot in November and it now has made the ballot and they wanted to stop it essentially. They've said this, Secretary of State Frank LaRose has said that the
that this issue one in August is about stopping the election in November. And in fact, they had canceled the election in August, which was kind of about some kind of obscure, you know, local stuff. And obviously August election, people are on vacation. Like there had been thought that like, you know,
why are we holding this? It's going to be so low turnout. We should cancel it. They went ahead and canceled it. Then when they realized that this abortion ballot measure was probably going to be on the ballot in November, they reversed course from that. So yeah, the Republican legislature basically put this on in order to change the rules to stop the abortion measure. And just to be clear, this is probably obvious, but this measure would only need to pass by a simple majority, but then would
henceforth change it to needing to be a 60% supermajority. Exactly, yeah. As I mentioned, from the ads, you would think that this question had nothing to do with a procedure for amending the Constitution. So let's take a listen to the ads. Here's one first from the opposition.
Do you have a condom? Yeah. Sorry, you can't use those. What are you talking about? Who are you? I'm your Republican congressman. Now that we're in charge, we're banning birth control. It's our decision, not yours. Get out of our bedroom. I won the last election. I'm not going anywhere. I'm just gonna watch and make sure you don't do anything illegal.
As you can tell, no mention of a 60% supermajority there. Here is an ad from the supporters of Issue 1. Caught. The out-of-state groups fighting Issue 1 keep saying it's just about abortion. But it's not, and they know it. These special interest groups encourage minors to get sex change surgeries and want to trash parental consent. On social media, they brag.
Planned Parenthood knows it's not just about abortion. They said it goes hand in hand with sex change surgeries. URGE says parental consent laws are unethical and need to be abolished. The ACLU, they're against informing parents about their kids transitioning at school. It's obvious. The reason these groups are in Ohio is to encourage sex changes and cut parents out of life-changing decisions. This isn't he said, she said.
They said it themselves. Stand up for your rights as a parent. Stand up against wealthy, out-of-state special interest groups by voting yes on issue one on August 8.
Okay, so there is a lot going on here, as we mentioned, you know, focus on culture wars. How are voters interpreting this? Does support or opposition to Issue 1 basically mimic the numbers we're seeing on abortion rights overall? So it's close, but they're not perfectly aligned. So those same three polls that asked about Issue 1 also asked about the November ballot measure on abortion, which doesn't have a number yet.
But the average on the abortion measure was 57% in support, 24% opposed, and then 20% undecided. So basically, if you compare the 57% who were pro-abortion rights in that, that's higher than the 45% who oppose issue one this year, or rather this month. And so, you know, that could be that...
you know, because the issue has been muddied a little bit. You know, obviously the ad that was in support of issue one was more about like trans issues. And so it could be that people feel like this isn't just, aren't literally taking this as a proxy for abortion, but are kind of being persuaded that it's also about other things like
The transgender ad is basically arguing that because the Constitutional Amendment of November is about reproductive rights generally, that that also is going to expand to things like letting minors do gender-affirming care without parents' permission and stuff like that. So yeah, so it seems like the two are fairly well aligned, but not exactly. And then I do think, though, that all those undecided numbers are obviously going to be key.
And I wouldn't be surprised. A lot of these polls are from several weeks ago. And of course, people may have heard more about the amendment since then and maybe now are thinking of it more in terms of like, oh, this is basically a proxy vote for abortion rights. And so I'll be curious to compare the final results of issue one with the final results, obviously, of the November election, but even also the polls of the November election come like October.
And just to set the stakes here, you know, this statewide election in November on abortion rights, I think is going to attract a lot of attention. First of all, you know, there is not that much going on. Otherwise, it's 2023. It's an off year election. And this will also serve as some sort of test of how salient abortion rights are as an issue heading into a presidential election, particularly in a purple-ish but now slash red state.
It seems like the situation you're describing is one in which the expectation is that this constitutional amendment will pass in Ohio, codifying abortion rights into the state constitution. Is that fair to say based on what you've looked at?
Yeah, I think so. So, you know, we have talked, you know, Amelia's talked on the podcast, for example, like, you know, about public opinion after the Dobbs decision and stuff. And, you know, in most states, they're really only a few, like, really hardcore conservative states where a majority of voters vote.
want abortion to be illegal most of the time. So in general, right, abortion rights, at least up to a point is popular and, you know, I think would pass by popular referendum. Just look at like Michigan, for example, or even in Kansas, obviously a redder state than Ohio. We had that ballot measure basically, I think it was like a month after the Dobbs decision where that ballot measure was the opposite of what Ohio's and Michigan's are
which is that that measure would have basically banned abortion in Kansas. The status quo was that it was allowed in Ohio. Then the status quo is that it is severely restricted and this amendment would expand the rights. But yeah, obviously, I quoted that
polling average a couple of minutes ago, 57% on average are in support of the abortion amendment. Obviously, that's between 50 and 60%. So I think that you would have to say that if issue one fails and 50% or rather a majority remains the threshold that it needs to pass, you would have to say that the abortion amendment is favored.
based on everything that I just said. But if issue one passes and 60% becomes a threshold that it needs to reach, then I think you'd have to say the abortion amendment is not favored. Yeah. And just to clarify where exactly abortion stands in Ohio today in terms of legality, I'm reading from an ABC News article on the topic. Abortion is currently legal in Ohio through 22 weeks of pregnancy or most of the way through the second trimester. After the overturning of the landmark abortion
abortion case Roe v. Wade, Ohio lawmakers passed a heartbeat bill that banned doctors from performing abortions after cardiac activity was detected, which can be as early as six weeks into a pregnancy, though the earliest restrictions are currently blocked in court. So this would in some ways sort of resolve
the legal issues and ensure that a six-week abortion ban did not go into effect in Ohio. So that's what voters are, you know, potentially thinking about when they vote on, well, both
issue one on Tuesday and also ultimately the election in November. Right. And that mirrors the situation in Michigan in 2022, when there was this court fight over an old abortion ban that was on the books in Michigan. And basically the ballot measure that folks might remember in 2022 that, you know, kind of created so much interest and I think helped people like Gretchen Whitmer get reelected by handy margins. That ballot measure ended up resolving that court fight on the side of abortion rights.
But Nathaniel, Ohio isn't the only state that has somewhat recently asked voters if they want to make it harder to amend their constitutions. What's the larger trend that we're seeing? One really interesting thing is that this is, as you mentioned, part of a pattern of Republican states in particular trying to block liberal ballot measures by increasing the threshold by which they need to pass.
So at least 10 states have tried doing this since 2017. In five of those instances, including this one in Ohio, it has gotten as far as going ahead in front of the voters, but only one of them has actually passed. So this was last year in 2022 in Arizona, Proposition 132 passed, and that would have raised or did raise the threshold for passing ballot measures that raise taxes to 60%.
And that passed just 51% to 49%. And all the other ones that have been kind of broader have failed. So, for example, there was a constitutional amendment in South Dakota that was kind of similar to this one. It was targeted at defeating a Medicaid expansion ballot measure that was going on the ballot later that year.
But in June 2022, Republicans put this amendment on the ballot to raise the threshold for the Medicaid expansion to 60%. That failed. It only got 33% of the vote, the rule change did. And so as a consequence, Medicaid expansion only had to reach a majority of the vote, and it did. So in general, these have not been popular moves in the past. It seems like voters like the idea of kind of a simple majority being
the one that dictates although it's it's not impossible as we saw with arizona but that was kind of a specific situation with just the taxes you know that can be i think um you know obviously raising taxes is generally not popular whereas abortion rights is generally popular um so i'm not sure that should be you know like
solace for supporters of issue one. There is one state, though, that has raised the threshold overall for constitutional amendments, and that's Florida, which was done in a bipartisan manner. What was sort of like the motivation there and what impact did it end up having?
Yeah, so this was back in 2006. And that was just, you know, frankly, 2006, you know, for those of us like me who remember 2006, like it seemed like a pretty partisan time, but I think it's child's play compared to what we have today. But yeah, that was just a thing where, you know, it was kind of Florida was, you know, felt like,
their constitution was getting bogged down because there were like multiple constitutional amendments on the ballot every two years. And everybody was like, you know, our constitution, like, you know, we shouldn't just have like random constitutional amendments for like random like tax things or whatever. And so there was this push that was like, okay, let's, let's raise the threshold here to 60%. That succeeded. And it's kind of an interesting proxy, I think, for what could happen in Ohio, right? It just had,
there there are numbers about you can see over the time since 2006 how many ballot measures have failed in florida with between 50 and 60 of the vote which obviously would have passed in a majority system um and so there were nine constitutional amendments over those almost 20 years that
failed within that zone. So there was one to increase school class sizes, one to switch to a top two primary system, one to legalize medical marijuana. So some like potentially impactful things did fail. To give you kind of like a sense of the numbers, there were 13 amendments during that period that failed with less than 50% of the vote. So that is obviously things that would have failed even under a majority system. So basically like
It actually comes out to 41% of the constitutional amendments that have failed since 2006 would have passed if Florida hadn't made this reform that Ohio is considering. Those are some pretty interesting numbers. Nathaniel, thank you for doing the math there. Of course, we will be watching to see what ends up happening in Ohio on Tuesday. Before I let you go, though, I do want to check in on...
one other topic. So it's been a little less than a week since Trump's latest indictment. And last week, we talked about how voters had been thinking about potential charges before the actual indictment came down. It seemed like
Voters were taking accusations of trying to overturn an election or obstruct an official proceeding quite seriously. But now that the indictment has actually happened and maybe folks have, well, probably not had an opportunity to actually read the indictment, but at least heard somebody in the news quote it and Trump has been arraigned. Are we seeing any differences there? Like how are voters responding now?
Yeah, so we've only had one poll of the Republican primary that's come out since the indictment and also just one poll of Trump's favorability and unfavorability since the indictment. So I don't want to read too much into that. I will say, though, that those polls didn't show much of a change. The horse race poll was from Ipsos and Reuters and had Trump at basically the same level as his previous poll. And the favorability poll was from Ipsos and ABC News. And that did have Trump with a pretty terrible 30% favorable rating. But you have to look at the
previous poll from Ipsos and ABC News, and they've always had Trump as not very popular. So the previous poll had him at 31% favorable. So not a lot of changes there. But we have had a couple of other polls kind of asking generally what Americans think about it. And it seems like Americans are basically retreating into their partisan corners and their pre-existing opinions of Trump.
So according to Morning Consult, 52% of Americans approve of the indictment relating to the 2020 election, the specific indictment. There was also a poll from YouGov and CBS News that found that 51% of Americans thought that Trump did try to stay in office by illegal means.
And 54% thought that if Trump tried to overturn the election, that would be considered undermining democracy. So all those numbers kind of in the same range of like a slight majority that of course is basically right in the zone of the number of Americans who view him unfavorably, the number of Americans who disapproved of him when he was president. So it seems like perhaps this isn't, well, I think it's still early that we still only have, you know, two or three polls, but, um,
But not a lot of, you know, surprising or kind of game-changing findings yet. All right. Well, let's leave things there for today. Thanks so much for joining me, Nathaniel. Thank you, Galen. Next up, what the country's best-readed pollster is saying about the 2024 election.
You're a podcast listener, and this is a podcast ad. Reach great listeners like yourself with podcast advertising from Lipson Ads. Choose from hundreds of top podcasts offering host endorsements, or run a reproduced ad like this one across thousands of shows to reach your target audience with Lipson Ads. Go to LipsonAds.com now. That's L-I-B-S-Y-N-Ads.com.
You're a podcast listener, and this is a podcast ad. Reach great listeners like yourself with podcast advertising from Lipson Ads. Choose from hundreds of top podcasts offering host endorsements, or run a reproduced ad like this one across thousands of shows to reach your target audience with Lipson Ads. Go to LipsonAds.com now. That's L-I-B-S-Y-N-Ads.com.
Last week, the New York Times and Siena College released their first round of polling of the 2024 election. And as we discussed briefly last week, in the Republican primary, the polling shows Trump with 54% support to DeSantis' 17%, with Pence, Scott, and Haley all at 3%. And by the way, that's not far off from what our averages currently show, particularly on DeSantis and Trump.
Support for Biden ticked up amongst Democrats, with 45% saying he should be the party's nominee, compared with just 26% last year. And in very early head-to-head polling of the 2024 general election, Trump and Biden objected.
are tied at 43%. And yes, we will get into what polling 15 months before an election actually means. Here with me to discuss it all is Ruth Gelnick, the editor for news surveys at the New York Times. She's mentioned she can also just be referred to as New York Times polling guru. Here she is. Welcome to the podcast, Ruth. Thanks so much for having me. What a great introduction.
I'm glad you liked it. And this is your first time on the podcast. So truly welcome. I'm excited to have you here. Yeah, excited to be here. As you may know, I'm not sure how well you followed FiveThirtyEight throughout the years, but we don't like to harp on individual polls.
However, you at the New York Times do extensive and notably expensive polling that can get beyond the horse race. And you also have a pretty well-established track record of being accurate and iterating when you aren't. So just for the record here, according to our pollster ratings of the 2022 midterms, the final time Siena College polls differed on average from the actual results in the midterms
by 1.9 percentage points, which is the most accurate results of any pollster with at least 10 final survey results in our database going back to 1998. Okay, so now that I've gotten the blowing smoke up your ass portion of the interview, I will be asking more pressing questions later on. Here's the first question though. Why were the Times' polls so accurate in 2022?
Wow. Why were we so great? What a wonderful question to start off. What a good softball. No, I mean...
You know, I think that what you mentioned is correct, which is that we work hard on our methodology and we're willing to iterate. We're very careful in the different things we select, but we're not wedded to any one thing. And I want to give all credit to Nate Cohn because those accuracy pollster ratings predate my time. I'm actually only about a year into this job and Nate has just been doing an incredible job and continues to. So all credit to Nate Cohn for making our poll so accurate. But what I think...
sort of the underlying secret sauce is, it's really kind of two parts. I mean, one, we are maybe some of the last of the dinosaurs in that we're wedded to phone polling, but specifically phone polling off the voter file.
Because we know that it gives us access to, you know, a full range of voters that we don't always get with online polling. And the other piece is the waiting after the fact. We employ a lot of techniques that are often used by campaign pollsters, not by sort of standard public pollsters. So kind of combining the good things about standard public polling and the good things about private internal polling for campaigns, putting that together so far has been really successful for us.
To the extent that there were any inaccuracies or challenges really in recent rounds of polling, you know, we've talked extensively post-2016 and 2020 about the challenges. But if there were still challenges in 2022, how are the times in Siena College still iterating?
Yeah, I mean, that's kind of a constant juggle for all of us. I think on our part, we did an experiment before the 2022 election in Wisconsin. We kind of picked the hardest state we could to challenge ourselves. And we did an incentive experiment where we tried to pay people to take our surveys, partially because we are concerned, like everybody, about non-response bias, this idea that people who are responding to surveys...
might be fundamentally different than the people who aren't responding to surveys. So we did this experiment in Wisconsin. We tried to get at, you know, we both did an incentive, but we also did this
big, expensive, high quality mail survey. And then we did a parallel phone survey and a parallel web survey. So we also used it as an opportunity to kind of look at differences over mode. And one of the things that came out of that experiment was one, we were able to get more people with that big, high quality incentivized mail survey, but it wasn't
A huge difference. And one of the things that came out was that our phone survey is still really good. So we took some of the learning and the things that we got from that high quality web survey and we tried to put it back into our phone surveys to improve them. So we're kind of constantly doing that. And I think we'll plan more experiments to that effect. We are definitely not giving in to like just because we're good now means we'll be good long term. Yeah.
Yeah, I mean, famously, it's difficult to get people to respond to phone surveys these days. You know, the response rate has fallen from, I don't know, back in the day, the golden age of polling. What was it? Something like in the double digits, at least now. Oh, my God. It's like 30 percent. And now it's like 5 percent or 1 percent. Or less. Yeah. So when you all paid Wisconsinites $25 to respond to a mail survey, what was it like?
What was the difference there? Like, were you able to get back up into the double digits? Yes, we were able to get back up into the double digits. There's no question that we were able to get response rates up closer to that, like 20, 30 percent. That was fantastic. What we found was that the respondents weren't that different than the non-respondents. One place we did see a difference that we're still kind of exploring is it's possible that we were getting more political moderates.
through that incentivized responding, which is to say it's not that we're missing the extremes of each party in our surveys. It's possible that we're missing those kind of like moderate slash undecided, uncertain, less engaged voters that we are able to get with those bigger, uh,
more expensive surveys when we pay people. So that's one thing we're sort of like digging into a little bit more is how can we get that group and make sure we're not missing them? Because obviously they are a crucial group and can be kind of swingy in a way that matters for polls. Right, exactly. They may be the people who end up deciding the election, deciding late, you know, and so if you think about what happened in 2016, maybe the
prime piece of the electorate for determining outcomes in places exactly like Wisconsin. That's why we picked it. So we will definitely be digging more into that. One of the things I noted in Nate Cohn's write-up of lessons learned from 2022 and this experiment that you did in Wisconsin was that you're now thinking of weighting by home value, which is
is interesting because we've all been talking about weighting by education ever since the 2016 election. It seemed like there were big divides in the electorate in terms of who has a four-year degree or greater, who has a high school degree or less. And we've emphasized that weighting by education is very different from weighting by income because you can have a high school degree and own a construction company and have a pretty high income, but you may still vote more like somebody who has a high school degree with a lower income.
But now home value, that's interesting because that seems like it tends in the direction of income but isn't quite exactly the same thing as income. So what is going on there? Yeah, I mean that's something we're trying to disaggregate. And like Nate said, we're thinking about it. We're still trying to kind of figure out what to do. But it is one of those things where when you start to tease it out, you do see those small differences like education and income.
So before I came to The Times, I worked at Pew. And at Pew, we have a longstanding question we asked about, would you rather live in a big house that's further apart or a small house that's closer together? And over time, that question has become more and more partisan in that Democrats want to live in smaller houses closer together. Republicans want to live in big houses closer.
further apart. And that's sort of related to this and home value. And we start to see that home value sort of trends in this partisan direction in a way that we didn't see before. And like education, that's a split that has been building over time. Like I think you know, and you guys, your listeners have heard you guys discuss on this podcast, you can't wait by everything because there is sort of a buy-in variance trade-off when you start adding waiting variables. So if you want to wait on something...
You have to think very carefully about it. And that's why we didn't wait by education before. We didn't wait by home value before. But as we see those splits start to happen, we wonder, like, is this something we should add to our waiting protocol? Is it worth the tradeoff? Which is to say, what exactly does home value predict in terms of political support?
Well, I think it's exactly as you might expect. Like, you know, I think that people with higher home values are more likely to be Democrats and people with lower home values are more likely to be Republicans. And so we're starting to see that sort of like split in a really interesting way.
And even within the Republican primary, people with higher home values, it seemed like we're much more supportive of Ron DeSantis than Donald Trump. Yes. And we're seeing that play out. I mean, the sort of tradeoff is that and income, because we're seeing, you know, that play out in the Republican Party. For example, in our Iowa poll of Republicans, we saw that
DeSantis did much better with high income Republicans, even more so than high education Republicans. So again, we're kind of still looking at those tradeoffs, but there is definitely that split within the Republican Party. All right, let's dig into the results here. So you released your polling data in waves, first looking at the Republican primary and Trump standing nationally, then looking at the Democratic primary, if you could sort of call it that.
and Biden standing nationally, then head-to-head polling between Trump and Biden, and finally polling in the first state to weigh in, which is, of course, Iowa. So we'll get to all of that. But let's start with probably the most news-grabbing result, but also the result that we all might be the most skeptical of. So in that head-to-head polling, Donald Trump and Joe Biden were tied at 43%, a full 15 months before the 2024 election.
Like, first of all, sort of what's the goal here, right? The nominees are not actually fully solidified yet. Donald Trump still has to go through a Republican primary. I mean, what were you hoping to gather from a head-to-head polling, you know, in August?
Yeah. I mean, you'll notice that when we released it, we actually released that a little bit later, mostly because we think it's maybe the least interesting thing, as you implied, because it's August, because these are registered voters, not likely voters. Like, what does it really mean? Nobody knows. We might as well get a benchmark on the race, but it's not all that interesting.
But I think the goal is sort of twofold. I mean, one, like I said, get a benchmark on the race. We want to look back at this in a year or in less than a year because in a year it'll be the election. But look back and say, how have things changed? How have people moved? But also, I mean, you know, we need to know as we're looking through this data, who are people who are saying they're Biden voters? Who are people who are saying they're Trump voters? So to some extent, it's sort of match people with their preferences, right?
But also, we also know there's this big group of people who aren't picking either. I think it's something like 14% in our poll. That's the group we're paying attention to. And so knowing what size that group is now, knowing how that changes, knowing the composition of that group,
do we think those people are going to vote or not? Like now's a good time to lay down that benchmark. I think it's hard to say what the 43 to 43 means right now, because again, it's so early and these are registered voters, not likely voters. Like people aren't paying attention outside of like you, me and the 538 crew, people aren't paying attention. But it's kind of worth laying down a benchmark. Well,
Well, maybe even more significant is that there are probably a lot of people who are paying attention who just look at the poll and say, oh, wow, it's tied, but like don't necessarily have all of the caveats sort of equipped to think critically about the polling.
You know, one thing that the polling showed was that to the extent that Biden is performing weaker than he performed in 2020. So obviously a 43-43 tie is worse than Biden fared in 2020 when he won the national vote by four and a half points. The Times' analysis shows that there's some softening with Hispanic voters, Black voters, particularly Black men, voters of color in general without a college degree.
Now, when you look into that, it's like, you know, maybe a change of five points. The sample size is relatively small, especially when you look into those crosstabs. What should we make of that? Like, do we have enough data to say, like, yes, truly Biden is doing worse amongst Hispanic voters than he was doing in 2020? Or is this a, hey, we don't really know, we have to pull, we have to oversample Latino voters or Black men or whatever you may have?
Yeah. So I'll give you kind of like a, like some of both answer, which is to say, I think that the data on black men, for example, our sample sizes are far too small and we would really need to do an oversample to, to know for sure that anything has really changed.
And the same answer for Hispanics, but I actually think our sample size with Hispanics is not that small. I mean, it is small to be clear, but I think some of that change with Hispanic voters we're seeing in other people's polling. And as you guys know, like looking across multiple polls is always better. And that's something that we are seeing in other people's polls. And so I think there is something potentially real there, that sort of small softening with Hispanic voters. We would need to do an oversample to really feel confident and also look at like,
the different subgroups of Hispanic voters and see what's really happening there. The other thing, you know, to keep in mind is we conducted this survey in Spanish and English as we do all of our surveys. But for example, when we did an oversample of Hispanic voters last cycle, we
We focused a lot more on making sure we had different dialects and things like that, that we try to do now, but it's not our top priority. And so to the extent that we can really dive in on that group, I think we get a better answer. But I do think some of that softening with Hispanic voters is real. With Black voters and Black men, I do think it's like the sample sizes are so small and there's no reason to believe that's necessarily happening, partially because
There was a lot of sort of hand-wringing about the change among Black men last cycle, and then a lot of the data has actually not borne that out. And if anything, this sort of like small movement towards Trump in 2020 among Black men was not really there in 2022, and like the Pew validated voter stuff and things like that. But I do think we'll need more data, but I do think there might be something real to the sort of shift among Hispanic voters. What I will say that we also saw in the data, though, for Biden was...
more strength among suburban voters and among women and sort of building back some of that. So I think some of that might cancel each other out, which is why he ends up in that slightly lower position, but definitely something to keep digging on. Yeah, no, I mean, that's definitely noteworthy because...
I've been looking regularly at polling from Equis Research, which listeners probably know is where Carlos Odio works, and he's been on the podcast many times. And what they found was, yeah, I mean, Biden did significantly worse with Hispanic voters in 2020, but their review of 2022 suggested that they had that Democrats had held pretty much steady in the midterms.
But it would be interesting, of course, if they started, you know, like falling again, especially as this more positive economic data starts to come in. And here we can start to talk about maybe Biden's standing, which is that it seems like almost every week we're getting more positive economic news. And, you know, this elusive soft landing may actually be within sight.
But Biden is still not viewed all that well amongst Americans. You know, only a 39 percent approval rating, according to your polling, that compares to our averages. Pretty, pretty similar. It's like 42 percent approval rating or 41.5 percent approval rating, if you want to be really specific. So what are his weaknesses here?
Yeah, I mean, you're absolutely right that the economic data has like helped him some, but not a lot, which is what he would typically expect. But I actually think his weaknesses here, I mean, I'm sure you guys have discussed this, but they're kind of the standard that people have been bothered with on Biden for a while. And that's his age. So maybe even like the economic data is like probably helpful, but ultimately that's not even why people don't like him. Interestingly enough. Yeah, yeah, yeah.
Yeah, yeah. Which is to say, right, like it's helpful, right? Like the economic conditions improved three or four points in the last year since we took our last survey and his approval rating sort of like directly tied to that and improved three or four points. But it really didn't help him that much because there's this decent group of Democrats that are still – they still don't like him and they kind of think that like –
these kind of non-negotiables for them, like his age, that hasn't changed. And so they don't feel that much better. So we asked the same questions this July and last July about whether Democrats wanted Biden to be the nominee. And for people who said they didn't want him to be the nominee, we asked why. And last July, a lot of people pointed to his age, but you also had this big group that
pointed to the economy and pointed to like, sort of like he hadn't accomplished anything or like what he's gotten done. Those second two groups, the economic concerns and the sort of getting things done, those groups have shrunk quite a bit. But the group who's concerned about his age, that's pretty big and they haven't really moved. So that's kind of a blocker for him with a certain portion of the Democratic Party. It's not obviously enough to really make a difference. And
Biden being the nominee. But, you know, there's a holdout. Of course, the polling sort of looks broadly. It's using the scientific method to try to get a real solid numbers of support. And then you also went back and called some of the folks who responded to your poll. And the quotes that you all got were really good and seemed to like capture things that are very clearly in the water. So one of the quotes was, I'm sorry, but both of them to me are too old.
Joe Biden, to me, seems less mentally capable age-wise, but Trump is just evil. He's done horrible things.
And so, you know, a very sort of like, honestly, put this woman on CNN. Like, you know, that's probably better political analysis than you hear most of the time. And I've seen this in other polling actually as well, where you ask, you know, if you don't support the, you know, Biden or Trump, why don't you support them? The number one thing for Biden is age or incompetence. And the number one thing for Trump is dangerous. Right.
And so it's like people view Biden as old and incompetent and people view Trump as dangerous. And then when it comes down to like, which are you picking when you think about those two things? You know, well, in this poll, it was even. It was 43-43. But it seemed like you did seem more upside for Biden, actually, than you did for Trump amongst the, you suggested, 14% of Americans who aren't decided. Yeah. I mean, it's...
It feels like you said, early and hard to say, but within that group, there's definitely more upside for Biden. There's definitely more people for Biden to grab out of that group. The caveat being it's also a group that historically votes at like tremendously low levels, right? Non-voters are overwhelmingly democratic.
So there's reason to believe a large share of that group won't vote, but the group is more favorable towards Biden, both like they view him more favorably, but also it's just a group that has historically voted more Democratic to the extent that they've historically voted. You know, it's cliche, but it's like it's a turnout game. And if Biden is able to turn out those voters that are currently on the fence,
then that could be the ballgame for him. All right, let's turn our attention now to the Republican primary, which is probably the most sort of immediately relevant polling here, but I've also saved it for last because we talk about the Republican primary all the time. So based on what the Republican electorate told you about their preferences and what you learned from Iowa, where voters have had more exposure to the candidates than anywhere else,
Is Trump's grip on the nomination firmer or looser than you expected going in?
Boy, that's a good question. I mean, firmer than I expected going in for sure. He was dominant in the national polls, 50% to DeSantis' 17%. But like, hedgy answer, he actually, it was looser in Iowa than I would have expected. Trump was still dominant and he obviously did very, very well in Iowa. But it wasn't nearly as strong as I would have expected, right? Like nationally-
Trump had 54% in our poll, DeSantis 17. When I looked across every single demographic group, every single opinion, Trump wasn't winning or DeSantis wasn't winning a single one of them. Trump was dominant in every single group. That wasn't the case in Iowa. So like rich, poor, men, women.
Yeah. College educated, not college educated. People who wanted to focus on woke issue, Republicans who wanted to focus on woke issues, Republicans who wanted to stay far away from woke issues, Republicans who were pro-abortion, Republicans who were anti-abortion. I mean, it was every single group. So he was definitely more dominant than I expected, even looking at the 538 average and expecting him to be as high as he was.
The dominance across groups was truly surprising, even groups that DeSantis probably should be doing better with. In Iowa, there were some groups where DeSantis was winning. And so that was like Trump's grip was a little bit less. And that was kind of interesting, which is to say in Iowa,
College educated Republicans and upper income Republicans. DeSantis was actually winning with those groups, which aren't sort of key constituencies for Republicans, but they're important. And so he was doing better than expected in Iowa.
Yeah, just to put some numbers to that, Trump was at 44% in Iowa, DeSantis at 20%, Scott at 9%, which is triple his national polling number of 3%, Ramaswamy at 5%, Haley at 4%, and then on down from there. What did we learn about Tim Scott from the national versus Iowa polling numbers?
Because there was the biggest differential there. So who is he performing well with in Iowa that may be suggestive of the kind of coalition he could try to put together if he's going to be competitive? And again, huge asterisk here. Like, obviously, Trump is doing extremely well in all of the national polls. That's exactly what I was going to say. To be clear, we're talking about like, you know, on the margins. But
Scott is, you know, is doing better in Iowa, as you said, in Iowa's retail politics. And so I think that there's something to that. People who have met him, people who've gotten to know him, they tend to like him. One of the problems actually for Scott and for DeSantis is they're kind of pulling from the same group of people, the more educated Republicans, higher income Republicans, who
It's interesting. We ended up not asking this, but in our questionnaire development, we had some questions about sort of like your vision for the Republican Party. Do you want this kind of more positive, optimistic version of the Republican Party, which is what Scott is sort of selling? Or do you kind of see things as a little more like, you know, in the Trump world of, you know, everything. Existential fight for the soul of the nation. Right, exactly.
Exactly. And we didn't end up asking questions exactly like that. But to the extent that there were sort of positive, optimistic, forward-looking things, those are the groups that Tim Scott is doing better with. He's doing better with Republicans who sort of want to see things in a more positive, optimistic way. His problem, of course, is there's a ceiling with that group. That's not the majority of the Republican Party. The majority of the Republican Party sort of does see it right now as this kind of us versus them Trump view of the world. So...
He's certainly doing better than expected. What did we learn about religious voters? Of course, Iowa has a large evangelical population, larger than New Hampshire, which will be the first primary state. Are DeSantis or Scott doing better or worse with religious voters than Trump? So they're both doing worse than him, but they're both doing better with religious voters than they are nationally, which is to say Trump is still winning religious voters in Iowa, which is something I think that, you know,
we all thought might not happen in 2016. Religious voters sort of were on the fence about him this early in the cycle. Right now, Trump is still winning handily. And when Paul honestly gave...
And helped Ted Cruz win Iowa in 2016. Right, exactly, exactly. But now Trump is winning evangelical voters. He's winning them by a smaller margin. DeSantis is doing better with that group in Iowa than he is nationally. And Scott is doing slightly better, but these are such small numbers that it's hard to say. Making the case for Trump going on to win the Republican nomination is—
is easy based on all of the data that we've already suggested. If you had to assemble the evidence to argue that Trump won't ultimately take the nomination, what data points would you point to in the polling that you all just published? Oh, what a good question. Wow. I'm blushing. Yeah, it's a really good question. It's hard to make the case
because he's so dominant. And I think part of the problem is largely that the remaining group is split. You know, you've got this big group of MAGA Republicans that are solidly for Trump. You've got this kind of like never Trumpy group that we all know is kind of sitting there around the, you know, sort of 20% ish of the Republican Party, 20, 25%. And then you've got the persuadable group. Now that that never Trump group is kind of split between these different candidates.
You know, I think it's funny. DeSantis has the best case for winning in Iowa, but in some ways, Chris Christie actually has a pretty good, you know, sort of thing going in New Hampshire. So again, it's hard because I think things are so split in this group that's not for Trump. So it really is hard to make the case for somebody else to win.
to really win. And all I can think about is, you know, if you look at your, the 538 aggregate graphic from 2016, you see each candidate having their little spike. And so it does feel like the path forward is anybody having a moment, they might be sort of a fleeting moment and then they go back down because they are still pulling from the same group. So...
Right. There was the like Ben Carson moment. There was the Carly Fiorina moment where she like ended up making the debate stage or the prime debate stage. Yeah, I totally know what you're talking about. So we talked about the actual numbers that you assigned to those groups on last week's podcast. So it was 37% is the MAGA crowd, 37% is the persuadables, and it was 24% or something like that was the, you know, anti-Trump, never Trump crowd.
Crowd. Amongst those persuadables, who looks like, who are they choosing between? Is it just DeSantis Trump is kind of that dynamic? Or are they also interested in Scott, Haley, Pence, whatever it may be?
Honestly, it's mostly DeSantis Trump and DeSantis is actually pretty warm with that group. I think there is a share that's pulling from, you know, like I think Scott is arguably doing second best with that group. They're more positive towards him. So, but I really do think DeSantis is, is best positioned to
pull away that group. And it's kind of interesting, like I said, that group is pretty warm towards DeSantis. I mean, Republicans in general are pretty warm towards DeSantis, but that group is pretty warm towards him. And one thing that I've noticed recently is DeSantis is trying on sort of different hats to try to see what part of that group he can peel off.
And so the most recent version of that is criticizing sort of Trump on the grounds of some of the indictments and outright saying that he thinks that Donald Trump didn't win the 2020 election. So what I'm really interested in is, you know, you've got this group of people who's persuadable. They're warm towards DeSantis and Trump, but a large share of them think that Donald Trump won the election. And so if you're DeSantis, are you really peeling away that group by saying that?
Yeah. I mean, you've really illustrated there just how challenging it is for anyone who's not Trump, because it seemed like early on DeSantis's strongest coalition was going to be amongst this more highly educated, higher income crowd of Republican voters.
But in the process of, you know, post-2022, some of the laws that he signed in Florida, the six-week abortion ban, kind of the way he was talking a lot about, you know, wokeism. I mean, even Trump, I think, started, like, making fun of him for overusing the word woke. But he sort of leaned hard into, harder into some of the culture war stuff. And it seemed like we were experiencing this almost like Ted Crucification of Ron DeSantis. Yeah.
And then he started falling in the polls. So maybe he's trying to reinvigorate his standing with the more highly educated, higher income Republicans. And obviously, that's a risk, too. When he had this broad appeal in Florida because of the economy was booming and he refused to shut the state down, well, after initially shutting it down,
Like, those are the kinds of things that he doesn't have to go head to head with Trump on. You know, like he can just claim all Republicans plus independents and probably some Democrats, too, and bring them along for the ride of like, you know, we want to go enjoy the beaches. Screw COVID. But that's not what this race has turned out to hinge on. No. And it's funny. We did kind of try to directly test that idea of wokeism versus Trump's view on on.
the world. And it's not a perfect representation, but we sort of tested this idea of a candidate who's focused on woke issues and woke fights versus sort of things that Trump is focused on, like crime and the border. And Republicans overwhelmingly selected crime and the border. And even DeSantis voters selected crime and the border, like selected against crime.
having these like woke fights. And we asked it a different way also in the context of businesses, sort of letting businesses choose what they focus on or, you know, kind of having these kind of like woke fights and not, not having businesses, um,
Like punishing businesses who prioritize diversity, ESG, etc. And like in both cases, the Republican electorate just doesn't have an appetite for that in a forced choice. And so, like you said, in a vacuum, DeSantis can have these fights and it's all well and good. But in the end, these are choices. And when Republicans were given that choice, they chose against it.
Right, right. When you can do the, like, porcanolos dose, fantastic. Like, you can have somebody, like, who can be your anti-work warrior and also, you know, fight for law and order. But when the distinction is, like, Trump talks more about the law and order stuff and DeSantis seems, you know...
I think in one of your write-ups of the poll, you cited a time when Ron DeSantis had used the word woke like nine times in 15 seconds. Something ridiculous. Okay, so I read through all of your polling, and these are the questions that I had. But you, as a polling expert yourself, probably saw some things there.
in all the data that stuck out to you that I haven't asked you about yet. Is there anything you want to like point to as like, hey, this is actually interesting that's getting undercovered in the race so far, either the Republican primary or, you know, even the general election? Yeah. I mean, you know, because we write our own stories about things, we like to suck all the meat off the bone. So if there's something interesting, we definitely put it in an article. But, you know, I think, um,
It's funny because it's going to sound very frivolous, but we did ask these questions, these head-to-head traits that I think are really interesting about Trump and DeSantis. Yeah. And we asked them nationally and we asked them in Iowa. And I think one of the really interesting things is, so DeSantis has kind of got this bad rap in national media as not being likable. And in Iowa, where it's like retail politics, everybody's met the candidate a hundred times. They actually thought he was more likable than they did nationally. Yeah.
So I thought that was really interesting that like maybe DeSantis like one-on-one is actually better than people think. I haven't met the guy, so I don't know, but obviously that's like,
something that is said about him often. And the other thing that was different there was the share saying he was electable. Again, that's kind of like the argument he's been making is that he's more electable than Trump. Nationally, that didn't really resonate with Republican voters. But in Iowa, that actually did resonate with Republican voters. So I thought that was really interesting. And then the kind of like frivolous fun part about it is we asked which candidate was more fun. And no single question across the survey tracked more closely with
their planned vote. So like Trump got nationally, Trump got like 54% saying he was more fun. DeSantis got 17% saying he's more fun. The really interesting thing is we had like nearly a third of Republicans saying neither or like sort of refusing to say which was more fun. So maybe that was some kind of like bellwether. There's 30% of Republican voters out there are still trying to find their fun in all of it. Right.
That's really interesting. Have you asked that question in other contexts, like in the national head-to-head sort of general election polling, if people think like Biden or Trump is more fun?
You know, we haven't. And now I kind of want to. I mean, it came out in the context of the Republican primary. As we were debating questions, we were thinking like, what is it that people like about Trump? And he's kind of like this entertainer. And so we toyed with this idea of entertainer. But it's really easy to say that he's more of an entertainer because DeSantis is not. And then we kind of drilled down to this idea of like, people say his rallies are fun. They're having fun with him. Is that something that people latch on to as an idea? And we wanted to test it against other sort of more standard things like
You know, this longstanding likable. Do you want to have a beer with him? Is he more electable? Like, does he care about people like me? Like, these are things that have been done for a long time. But I would like to test fun in a national head to head. I suspect that Biden voters don't have that impression of their candidate, but I could be proven wrong.
Yeah, I mean, to give another quote that speaks to that, another Biden voter giving the reason for why they would vote for Biden. Ultimately, Biden is not going to harm the country as much as I believe Trump would. Yeah, the quotes are pretty outstanding. And all credit to Cam Baker, who's done a lot of those calls. There was one in the Trump story, something about underpants that I feel like got a lot of national play. Yeah.
Well, okay, so this, I feel like reputable pollsters probably are usually not going to wade into, but I'm curious, would you ever ask a question about appearance, image? Is this candidate good-looking? Do they have, I mean, obviously our current field is not a paragon of conventional hotness or whatever you want to say. What's the good-looking choice? Right, like...
I mean, I don't know. There are certainly gradations of like hotness in politics, right? And we've been getting all this focus on like, oh, RFK is bench pressing and like Francis Suarez is like thinking about cutting an ad shirtless, you know? And the conventional wisdom has always been that in a television era, like the actual presentation of the candidates is,
But in a world where we don't have like a conventionally hot candidate as like one of the leading candidates, I think is that like, can I say that? Is that like nonpartisan, like unbiased to say? Like I think probably most people don't think that Biden or Trump is like conventionally attractive. You might be hearing from some people who feel differently, but I support that as an
I think it's reasonable to say. I mean, they're, you know, all of our grandfathers, so. Yeah. Is there any history of, like, polling who people think is attractive and sort of aligning that with how they're thinking of voting or no? You know, I'm sure if I looked in the Roper database, you could find some, like, kind of questionable polling from the mid-century that got into that. Not anything in the recent era, partially because, like you said, probably conventional pollsters wouldn't touch it. And it's really, really hard to ask about it. I mean, I don't even know what you're asking about. And, like, obviously, you know,
The gender of the respondent plays a part in that.
But I think that's where sort of this idea of like likability initially came from and like want to have a beer with that person. There was some like men want to be them, women want to sleep with them idea that obviously was like common in the like sort of like early JFK. But then Clinton, like that's something Democrats have kind of always been chasing and Republicans less so. Sorry, Republicans. But I can't imagine that's something you'd ask about. Come on. What an awful thing to say about George W. Bush. Yeah.
Okay, so my final question here is equally vulgar. Are you allowed to talk about how much you spent on these surveys? Sure. I don't nobody's told me to not talk about it. Okay, how much? How much does it cost? I want people to know because like, I think people take this kind of data for granted. But like, this is extremely costly. How much did it cost?
Yeah. So, boy, you're putting me on the spot and I don't have the numbers in front of me because we were really going back and forth a lot. And remember that we share these costs with Siena College, who very graciously, you know, we work together and they take on some of those costs. But to do this kind of national poll with a Republican oversample and this Iowa survey, we ended up around, I want to say, $150,000 to $175,000. So...
Take it or leave it. That's a lot of money. Actually, that's not as expensive as I thought it was going to be, but you're going to be conducting lots of polling like this. So your budget over the election is going to be like,
Yeah. Millions and millions of dollars or something like that. Yes, yes, yes. I mean, millions. We've very generously gotten already a lot of money, but it will end up being millions of dollars. Thank you to the New York Times company for investing in polling, if anybody's listening. No, but it's insanely expensive. And because, you know, at the top of the interview, you mentioned, you know, we do this gold standard polling. It's really expensive, right? Like we are facing a lot of pressure to do this.
from the rest or a large portion of the polling world that's doing cheaper online polling. And it may not be as high quality, but like sometimes it feels like maybe it's good enough and we're spending a lot of money. And so we get some of that pressure. I think, you know, we feel like this is worth it. And obviously you guys called us the most accurate pollster, which we appreciate. But I mean, we didn't that wasn't we didn't just pull that out of thin air. That's a rigorous finding, Ruth.
Yeah, right, right, right, right. I think, you know, you did a lot of testing on that. So, you know, we do really accurate polling. It's just really expensive. And as response rates go deeply down, that's something we contend with. One thing that we did differently here and we're kind of continuing to experiment with is testing.
This is like super wonky, but when we're doing our sampling, we do what's called PPS sampling, which is probability proportion to size. So we're trying to like sample a little bit to try to get areas that we think have higher response rates. I mean, we're still sampling nationally and everybody still has a chance of selection, but we're kind of trying to lean towards phone numbers that we know are higher quality.
quality, right? Because on the voter file, they list the quality of different phone numbers. So we know if a phone number is high quality or low quality. And we're trying to sample, we build models, Nate builds models that we try to sort of guess people's likelihood to respond. So we're trying to do some of that to ease our cost a little bit. But we know that the trade-off of that, of course, is that, you know, then you're selecting different types of people. And so there's some bias built in. So we try to not do that
you know, full throttle, but we're trying to build some of that in to help our pricing model because it's just so expensive. And I think, you know, I wasn't at the times in 2020 or 2016, but I think that our polling has gone up something like three times more expensive than it was in that previous cycle. So, you know. Yeah. It's costly. Well,
On my behalf, I'll say, you know, I'm glad you guys are doing this and spending all that money so I can, you know, learn more about the country that I live in.
But Ruth, this was fun. We'll have to have you back on as you continue to do polling throughout the election cycle. So thank you so much. Yeah, thanks for having me. Ruth Gelnick is the editor for News Surveys at The New York Times, also just known as Polling Guru for The Times. My name is Galen Drew. Tony Chow is in the control room. You can get in touch by emailing us at podcasts at 538.com. You can also, of course, tweet at us with any questions or comments.
If you're a fan of the show, leave us a rating or review in the Apple Podcast Store or tell someone about us. Thanks for listening and we will see you soon.
