Andy Craig: Welcome to Zooming In at The UnPopulist. I’m Andy Craig.
“Populism” is one of those words that gets used constantly, but it’s rarely pinned down. Everyone agrees we’re living in a populist era, but almost no one agrees on what that means or how you’d measure it. For the past two years, ISMA, the Institute for the Study of Modern Authoritarianism, publisher of The UnPopulist, has been trying to do exactly that: a national survey built around specific, testable definitions of populist attitudes, first fielded on the eve of the 2024 election, and now repeated in June of 2026.
The second round results just came back, and the headline finding is a bit of a surprise. Trump’s hardcore populist base has shrunk by about a third.
Trump's Hardcore Populist Base Has Declined By a Third, ISMA's Survey Finds
The populist base anchoring President Donald Trump’s political movement has declined substantially over the past 23 months, according to a late-June survey conducted by the Institute for the Study of Modern Authoritarianism, publisher of The UnPopulist
To walk through what the survey found and what it tells us, I’m joined by the man who designed and ran it, Tom Shull, editor at large of The UnPopulist and ISMA’s director of survey research.
Andy Craig: Tom, to start, I want to dig into how this survey is unique and what we were looking at with our definition of populism, because people use that word loosely. We talk about populist politicians and populist attitudes in a broad sense, but this was really digging in on a consistent definition for how we can measure the core of illiberal, or broadly speaking, populism. So what definition did we use, and what did we find out of that?
Tom Shull: The definition in this area of academic research, on populism, is hotly contested. In fact, I think our model, which is an abstraction of various summaries of that literature, probably would find people in the academic community pointing at parts of it and saying, “that’s not necessary,” and pointing at other parts and going, “but this is good.” So my hope is that, by and large, anyone who’s studying in that field would look at these things, maybe find some degree of agreement or disagreement, but they would understand and recognize these all as being within the realm of populist attitudes that are typically discussed.
We look at four attitudes. The first of those is feeling some sort of unusually strong personal attachment to the candidate—the leader of the party, the candidate for president—and his agenda. The second is essentially a feeling of anti-elite sentiment, typically towards government elites, though it can be others, such as business elites, cultural elites, and so on. Another is a sense of antipathy towards a social group, something that we call a social outgroup. This means a negative feeling towards that group generally, as well as a sense that they are somehow advantaged by government elites within society, and, finally, that they are not really able to be true members of the national community. The fourth attitude is an attachment to the leader of the party, or the presidential candidate, that is strong enough that they’re willing to see that person exceed the legal limits on their presidential authority, their executive authority, in order to accomplish their goals, the movement’s goals. There’s a concern for liberal norms that’s built into our fourth attitude.
Basically what we do is test for those attitudes in the survey and then say that anyone who has at least three of them would count as being part of a populist coalition.
Craig: This is the second time we’ve run this survey with that metric. Compared to last time—shortly before the election; this time we’re looking back at people who say they voted for Trump—your top finding was that among Trump 2024 voters, that measure of populism has declined as a relative share. It was about a third, and now it’s about a quarter of Trump voters. What do you think is driving that decline? And within those different metrics you’re looking at, what’s moved the most?
Shull: You’re right, what we did was look at the people who in August of 2024, in the run-up to the election, said that they were going to vote for Donald Trump. At that time, we measured populism along the lines of what I just described. And when we looked at the percentage of populists that we found amongst those who said they would vote for him, it was about 36%.
Now, we went back into the field very recently. We were in the field from June 22 to June 29. We had a nationwide survey, a thousand respondents, and we asked people who they had voted for in 2024. Those who said they voted for Donald Trump were also then assessed on these various populist attitudes. And this time we found that only 25% actually registered as populist by our methodology. So this decline among Donald Trump’s 2024 voters was from 36% to 25%—as you say, almost a third.
“[T]he Trump populists are folks who, in a variety of ways, are particularly attached to that movement. They're attached to his political mission. And that attachment tends to weather pretty well when you've got some storms. But what we see amongst those people with that stronger attachment is that even they have shifted their attitudes towards some things that were, in the past, part of their agreed set of attitudes towards Donald Trump and his presidency. … [T]here was a strong feeling by 2024 that [this stronger populist sentiment amongst Americans] was just an enduring part of the political landscape in the United States. … What I think I would dispute, and what I think the numbers here show, is that it isn't completely immune to change. It is not an ever-fixed mark.” — Tom Shull
We saw declines in three of the four areas that we measure. There was some decline in terms of personal attachment to the president and/or his agenda. That decline was on the order of 22% in August 2024 down to about 17%, a little over four percentage points.
Then we also saw a decline in the social outgroup area—the percentage of people who had chosen a social outgroup of some kind and said that yes, they had those three attitudes of negative feeling, a sense that they were advantaged by government elites, and a sense that they couldn’t be true members of the national community. That also had declined somewhat, from about 54% down to 49%—so, a decline about the size of what we saw on the question of personal attachment.
But the biggest decline came in that fourth and final attitude, and that was a willingness to see President Trump using his power outside the bounds of the law in order to accomplish his goals. We measured that in a couple of ways, but we saw drops in both [sets] of the questions that we asked. And it amounted to the largest decrease that we saw, which was from 63% in August 2024 all the way down to 48% in June of 2026. So it actually went from a strong majority to less than a majority, and that was a surprise.
The other thing I would say is that we divided all the data up from August 2024 and June 2026 and did our very best to dissect just exactly which of the declines—whether it was in personal attachment, in social outgroups, or this issue of less support for his use of extralegal authority—really had the biggest impact. It’s impossible to tell in a sense, because you’ve got two different groups of people. It’s not the same set of voters taking the survey. But as a practical matter, it looked pretty darn clear what happened: a number of people that we qualified as populists dropped out of the populist category, in particular because of that last attitude, that willingness to support the president in the use of extralegal authority.
So yes, all of the declines that we saw in the populist attitudes affected the number of people we categorized as populists. But as a practical matter, probably the biggest effect came from this question of support for use of executive authority outside the bounds of the Constitution or law.
Craig: That illuminates part of what we’ve been seeing in the news. Because of course, Trump came in at the start of 2025 and there was a big spree of things that we would say are abuses of power—really pushing hard on, and in many cases flat-out breaking, the legal limits on the presidency and executive power. And since then, there’s been a lot of tracking of his approval ratings, which have declined. So in some sense his overall support has declined.
But what this really highlights is that within his support base, such as it is, there’s been a decline in exactly that: in that degree of support for “my guy can bend the rules or break the rules.” And I think part of that is maybe people learning the lesson that they got what they wished for and didn’t like it as much as they thought—or at least the ways Trump has been doing it in practice, with things like launching the Iran war or the controversies over immigration enforcement. There was a bit of people recoiling from it, even among people who had voted for Trump, who are generally speaking Republicans or Republican leaners.
Shull: I think that’s exactly right. We do see that decline amongst Trump’s supporters. It is perhaps not shocking in one sense. Many of those decisions have not polled that well. So when the question comes up—would you expect a decline in that support for his use of extralegal authority, or bending the rules as you put it, and in this case sometimes breaking them—it is not a shock that we see some decline overall amongst Trump supporters.
What is particularly interesting here is that the Trump populists are folks who, in a variety of ways, are particularly attached to that movement. They’re attached to his political mission. And that attachment tends to weather pretty well when you’ve got some storms. But what we see amongst those people with that stronger attachment is that even they have shifted their attitudes towards some things that were, in the past, part of their agreed set of attitudes towards Donald Trump and his presidency.
When you go back to 2015, 2016, and you see the rise of what appears to be this stronger populist sentiment amongst Americans—there was, I think, a strong feeling by 2024 that it was just an enduring part of the political landscape in the United States. Because here Donald Trump had done something that is very, very unlikely: he got three nominations to run for president from the same party, and he won two of those elections. And it just seemed like it’s deeply embedded.
“We asked whether you approved or disapproved of the actions of the people who entered the Capitol during the counting of votes on Jan. 6. We asked that in August 2024, and we asked it again in the most recent survey in June. And what we found is that essentially we got about the same level of support, which was very little. … [F]or all of that argument that these people were mistreated and misunderstood, you just don't see movement in support. … There's just not support there, except amongst the Trump populists.” — Tom Shull
Honestly, I would tend to agree that there is probably going to be persistent populist sentiment moving forward. I wouldn’t dispute that. What I think I would dispute, and what I think the numbers here show, is that it isn’t completely immune to change. It is not an ever-fixed mark. It is something that can be changed by circumstances and, as you say, experience—watching what happens. Because ultimately, if you want to protect certain political results, if you’re looking to achieve certain political goals, and you see decisions that aren’t leading you in the direction that you’d originally hoped for, it’s quite possible and quite likely that you will start to change your mind, even if you have a strong loyalty to that project.
Craig: So on the other side of the aisle, we’re in the middle of the primary season for the upcoming midterms, and there’s been a lot of drama on the Democratic side—factional fighting between the more left-wing and the more centrist wing of the party, lots of high-profile primary fights. But one thing that I thought was interesting: we saw all this movement among Trump voters on the populism score, but among Harris voters, among Democrats, which is broadly the whole left-of-center chunk of the survey, it was flat. It was 15% before the election and it’s about 15% now.
Does that tell us something? When we speak of this, people will often conflate the more progressive wing of the party, all the way out to the democratic socialist, DSA-type candidates, as more populist. But does this tell us that maybe that’s a bit orthogonal to this question—that those ideological splits don’t necessarily map onto our definition of populism?
Shull: Yes, I would agree. I don’t know that these things are a dead match. Now, without knowing for sure—especially because we have yet to do some important data cuts there—I believe that there will be some overlap.
Part of the reason I say that is, when you look at Trump populists, as we have [in 2024], and compare that to the number of people who say that they are very conservative—which is what you do hear from a lot of strong Trump supporters: “I’m very conservative”—or if you ask, “Do you consider yourself a MAGA Republican?,” there is overlap, but it is on the order of 40 to 60%. So while it’s a statistically significant connection, and you’re not going to deny that there’s some strong overlap, you do see definite differences. And those differences are distinct enough that we feel looking at this question of populism per se, through the lens that we’ve used, actually does add to your understanding of what you’re seeing.
We did not ask people, for instance, whether they were members of Democratic Socialists of America—though obviously moving forward we would be doing that. But if I had to project, I think that we probably would see a correlation, but not a one-to-one mapping by any means.
Craig: So one interesting finding, zeroing in on that antipathy-to-outgroups prong of our populism filter: there was an overall decline in that. And to be clear, you asked about a whole wide range of groups—all the different religious minorities, lots of different groups. But there was a noted increase among Trump populists in antipathy to Muslims and to transgender people. Obviously those are two things that, to a degree, have been in the news. Does this show that the party’s messaging, Trump’s own messaging, still has some ability to move his base on these things?
Shull: Yes, I think it does. It’s difficult in a sense, because we did not know in advance exactly what groups might show an increase and what groups might show a decrease, and we did not ask follow-on questions about what drove the change in people’s views. But there is no doubt that when you look specifically at Trump populists, they do follow the news carefully, they’re more inclined to vote, and they have strong views. All of these things suggest that they’re paying close attention. That’s no doubt true of the non-populist contingent that supports Donald Trump as well, but there are differences between those two groups that are really quite pronounced on a number of important questions that we ask.
Do I think that part of this is the emphasis the administration has put on, say, enforcement of immigration laws and questions about the possibility of Muslim terrorism and things of this nature? I would say that no doubt had an impact. There’s obviously a growing resentment on the question of transgender athletes, and that may well have increased even entirely absent actions taken by the Trump administration.
Craig: One thing you mentioned there that I think is worth touching on: some people might say this is the populist era. We hear that all the time—that we’re living in a time of rising and very influential populism. But by our filter here, they’re still a minority. Most people are not populist in this sense. They are not holding animus against outgroups, supporting abuse of power, or having this parasocial relationship with their leader.
But one thing about populists is that they are more highly engaged. They are relatively more informed. We saw pretty consistently that they were less likely to say they don’t know, or have no opinion, on various questions. For example, in the piece I just wrote about electoral reform and the party system, they were pretty consistently likelier to have an opinion in both directions on those questions, because there were fewer in the middle answering unsure.
“Among [Trump supporters] we rate as populist, 51% say that they approve of Trump potentially running for a third term, and only 29% said they disapprove. So it’s not an overwhelming supermajority, but it’s a pretty big, healthy chunk of those voters saying that they approve of this pretty radical, crazy idea that he keeps floating. But among Trump nonpopulists, it was pretty much reversed: 57% of the nonpopulists who voted for Trump last time said that they disapprove of the idea of seeking a third term, and only about a quarter, 26%, said they approve. So I think that’s a pretty strong indication that this filter we’re using … is getting at real divides that do track onto other issues.” — Andy Craig
But also on propensity to vote—I thought this was really dramatic. Among Trump populists, you found 88% plan to vote in the midterms, and 83% of Harris populists, versus only 50% of other voters, which is around what we see typically for midterms turnout. It’s usually around 50%, give or take. So do we see that populists, because of their higher engagement, and because they’re in that sense more opinionated or more informed, depending on which gloss you want to put on it, punch above their weight in terms of political impact?
Shull: That certainly was my conclusion after just our August 2024 survey. As I mentioned earlier, they were more inclined to say that they were going to vote, and they expressed that more strongly too. They also were more inclined to follow the news. If you tested them for their political views on various issues, you saw that they took the more absolutist position in general, or at least came closer to it than other voters.
Now, there’s a huge debate about whether or not that truly affects your propensity to vote, and whether or not you therefore dominate the primary elections. I would grant that debate is not settled. But in the midst of this, I think your actual average political practitioner who looks at survey research would look at those groups that we’ve deemed as populists and say: those guys are probably going to the polls. And either that’s bad for me or that’s good for me, but that would be their strong instinct. These are folks who are going to the polls.
In this most recent survey, we asked not just, “Will you probably vote?” or, “Are you likely to vote?” We gave “I will definitely vote” as one of the possibilities. And this was because we know that people will say they’ll probably vote and then they don’t vote. But “definitely vote” is upping the ante. Those percentages that you just gave out—about 88% of Trump populists, 83% of Harris populists, and 50% of everybody else—that was on the question of “I will definitely vote.” And that willingness to say “definitely vote” is a pretty good indicator.
So I would say that they do punch above their weight. I think they do tend to have an outsized influence. And all of that can be independent of the ballot box, because there’s a social atmosphere, a political atmosphere, that’s generated—especially in an era of social media—by the people who are most involved in politics. These are the things that they’re most likely to talk about. If they’re on TV, they’re going to bring it up. If they’re on social media, they’re going to share that particular post. They drive the conversation in many ways independent of the actual ballot box. But I think you’re also looking at them punching above their weight at the ballot.
Craig: One of the things that we’ve seen from Trump—it’s come and gone in waves, but recently he’s revisiting it more and more—is telling his jokes-that-aren’t-jokes about running in 2028, about breaking the term limits and trying to seek a third term. And one of the results that I thought most confirmed that we’re onto something real with this measure of populism is that there was a huge split here among Trump populists versus Trump voters who are not populists.
Among those we rate as populist, 51% say that they approve of Trump potentially running for a third term, and only 29% said they disapprove. So it’s not an overwhelming supermajority, but it’s a pretty big, healthy chunk of those voters saying that they approve of this pretty radical, crazy idea that he keeps floating. But among Trump nonpopulists, it was pretty much reversed: 57% of the nonpopulists who voted for Trump last time said that they disapprove of the idea of seeking a third term, and only about a quarter, 26%, said they approve. So I think that’s a pretty strong indication that this filter we’re using—that you came up with, and that is unique in the literature for trying to measure populism—is getting at real divides that do track onto other issues.
We also saw that when you ask people their 2028 presidential choice, other than Trump running again, the Trump populists are a lot keener on Vance. Vance came out ahead across the board, with Rubio in second, but we saw Vance do quite a bit better with populists, and in particular with people who had the strong personal attachment to Trump among the populism prongs. So does that feel like a little bit of a vindication that this is not noise?
Shull: I think even just with the August 2024 survey, there were places where the splits and the things that we were seeing were so in line with what one might expect—with the occasional surprise that, on reflection, was also quite plausible—that it felt like we were already there in August of 2024. But I will grant that you just don’t know. Maybe that was simply the luck of contingent circumstances at the time. And having this survey come out and seeing the kinds of splits that we did was something that did lead us to feel that we had something giving us a different leverage on what was going on, how the attitudes were being determined, what you might be able to predict.
So, as you say, you have a 51-to-29 favorability rating from Trump populists on the idea of his running for a third term. Then if you look at his non-populist supporters—and remember, these non-populist supporters are now 75% of the people who voted for him, so they’re a very large share, arguably a supermajority—they are against this: 57% of them disapprove, and only 26% approve. If you go outside of that, to non-Trump voters, now you’re up in the 70s—I’m looking here, 79% of them are opposed to this. So if there’s going to be a drive for a third term, it’s going to take a lot of work to get there, because the public support really isn’t there.
You see the same thing, by the way, on the question of satisfaction. Most surveys start with the question—and ours asks it right near the beginning—”do you find yourself generally satisfied or dissatisfied with what’s going on in the country right now?” If you look, 74% of the Trump populists are satisfied, but they are a complete outlier. You go to Trump non-populists, who, despite their partisan attachment to the Republicans and to Donald Trump, are split on this question: 42% dissatisfied, 40% satisfied. So they’re actually leaning dissatisfied when their party is dominating Congress and holds the presidency. And then when you get to dissatisfaction elsewhere, amongst non-Trump supporters—non-Trump voters, I should say—it’s 76% dissatisfaction. So once again, you see that the Trump populists are definite outliers.
“We were in the field from June 22 to June 29. We had a nationwide survey, a thousand respondents, and we asked people who they had voted for in 2024. Those who said they voted for Donald Trump were also then assessed on these various populist attitudes. And this time we found that only 25% actually registered as populist by our methodology. So this decline among Donald Trump’s 2024 voters was from 36% to 25%. ... We saw declines in three of the four areas that we measure. ... But the biggest decline came in that fourth and final attitude, and that was a willingness to see President Trump using his power outside the bounds of the law in order to accomplish his goals.” — Tom Shull
And by the way, I will throw one more in on that, Andy: the question of Jan. 6. We asked whether you approved or disapproved of the actions of the people who entered the Capitol during the counting of votes on Jan. 6. We asked that in August 2024, and we asked it again in the most recent survey in June. And what we found is that essentially we got about the same level of support, which was very little. If anything, it had gone down just a little bit.
I thought that was fascinating, simply because there’s been a hard push in the Trump administration to look at that group of people and say they were mistreated, they were not treated well. I would say if I were in the Trump strategic group, or the Trump White House, looking at the survey numbers, I’d be a little frustrated—because for all of that argument that these people were mistreated and misunderstood, you just don’t see movement in support. If anything, it tailed off just a bit.
The only people that we’re finding who say they’re in favor of it—and it’s just a plurality—are the Trump populists, who are running about 36% approval, 28% disapproval. When you get outside that, it completely reverses: Trump non-populists are at about 20% approval, 47% disapproval. And by the time you get out to non-Trump voters, it’s like [5]% approval, [70]% disapproval—and that disapproval is very high in the sense that people are saying they strongly disapprove, not just that they somewhat disapprove. For the people who aren’t accounted for, they’re just saying they don’t have an opinion. There’s just not support there, except amongst the Trump populists.
Craig: As we’ve touched on, this is the first slice of findings, but this was a very large project. This is not a normal horse-race poll—I can speak to that, and you’ve of course worked heavily on it. There are dozens and dozens of questions, and we’re going to be continuing to roll some of these findings out at The UnPopulist. So I just wanted to ask you: What more do we have to look forward to? What are some questions you still want to dig in on more?
Shull: Some of that’s going to depend a little on what we find. Do I see some areas that I find interesting? Yes, I do. You stole one of them, Andy: you looked at the question of exactly what kinds of electoral reforms might find favor amongst the voting public, and I thought you did a real nice job of treating that question—whether it was proportional representation, ranked-choice voting, Top Two. So for folks who have not looked at the UnPopulist essay on that by Andy Craig, I would absolutely encourage you to do that.
For me, one of the things I’m very interested in is looking at the question of populism amongst the 2024 Harris voters. We never saw, and still do not see, a strong personal attachment to her. So an awful lot of what we are seeing in populist attitudes there has more to do with anti-elite sentiment—which, perhaps not surprisingly, has actually gone up a bit in the past two years. And support for extralegal uses of power from an executive, even if it is Kamala Harris—which is how we pitched it: “What if it were Kamala Harris? Are you comfortable with this?”—that declined, and that followed the pattern that we expected. We expect that when there is a Republican in power, particularly the one who defeated Kamala Harris, there are going to be a lot of folks who might otherwise support it who are looking at the results and saying, “maybe I’m not in favor of that after all.”
So we will look at that question—see if there might be potential, if there were different kinds of leaders involved that evoked more passion, if there’s a possible increase ahead. One of the things that we do try to do in this survey is provide an early warning detector of when populism might take off.
But aside from that, we saw some important shifts in the demographics of the Trump base [between 2016 and 2024]. There were increases in minority voting for him, amongst Hispanics and amongst Black voters. I’d like to look at that within the context of our model and see where that comes out. And there are other straightforward questions that we haven’t touched on here in this interview, or in the essays so far, and some of those have to do with attitudes towards America. Here we were [in the field] just before the 250th birthday of the nation, and so we asked questions about pride in being American, whether Americans are freer now than they once were, and whether we’ve made moral progress or not. I think all of those, from what I have seen so far, could produce some very interesting work and essays.
Craig: Excellent. Tom, thank you very much for all the work you’ve done on this. This is really—I want to emphasize to people—a massive undertaking. And ISMA has provided, I think, something that’s a genuine contribution to understanding populism. Because, as you touched on, there are a lot of definitions floating around about populism, but it’s been really difficult to disentangle: are you actually measuring populism, or are you just measuring something that’s coded in a way that captures more of the right? And I think this really shows that you can get at something real: even though there is more of it on the right, we found that there are some populists on the left, and that the share of populists on the right can move independent of other variables.
I’m really looking forward to seeing more of this as we roll it out. And hopefully, at some point down the road, we’ll have another cycle of this. So Tom, thanks for joining us. I encourage everyone to stay tuned to The UnPopulist for more of our populism survey results.
Shull: Thank you, Andy. I appreciate the chance to talk.
© The UnPopulist, 2026
Follow us on Bluesky, Threads, YouTube, TikTok, Facebook, Instagram, and X.
We welcome your reactions and replies. Please adhere to our comments policy.
