Subscribe here: Apple Podcasts | Spotify | YouTube | Overcast | Pocket Casts
Across 11 different democracies, politicians share a shockingly pessimistic view: They believe that their voters are uninformed, unreasonable, and short-sighted. In a paper recently published by the American Political Science Review, the University of Calgary political scientist Jack Lucas and his co-authors surveyed 12,000 citizens and conducted face-to-face interviews with nearly 1,000 elected officials. In this wide-ranging study of countries including Australia, Denmark, Germany, and Canada, the researchers find that elected officials and voters diverge wildly. Unlike politicians, voters believe themselves to be policy-oriented, politically knowledgeable, and engaged.
On today’s episode of Good on Paper, I speak with Lucas about how this research reveals a striking divide between democratically elected officials and the voters who put them into office. Even in countries with extremely varied political systems and demographics, this difference persists, suggesting that something fundamental to democracies—at least modern ones—is creating this division. Tellingly, Lucas’s research shows that senior politicians tend to be more cynical about voters than junior ones, suggesting that expecting the worst from voters doesn’t carry an electoral penalty.
Despite this, Lucas is an optimist:
“I’m more on the ‘voters aren’t that dumb’ side of the spectrum,” Lucas tells me, acknowledging this puts him in the minority among political scientists. “There is pretty good evidence that at least on issues that voters care a lot about, they’re thinking carefully about policy.”
The following is a transcript of the episode:
Jerusalem Demsas: When Donald Trump said, “I could stand in the middle of Fifth Avenue and shoot somebody, and I wouldn’t lose any voters,” he was ostensibly making a joke about the loyalty of his followers. But another way to read it is that it is a revealing comment on how he thinks about voter psychology—that they care little about personal virtue and are with him for his other qualities.
Deciphering voter psychology can feel like an insurmountable exercise. If you talk to individual people about why they voted for their chosen candidate, their answers range from reasonable to incomprehensible.
In a new paper, the University of Calgary political scientist Jack Lucas and his co-authors investigate this through survey evidence. Their interests lie in both how voters see themselves and, crucially, how politicians see their voters.
In face-to-face interviews with almost a thousand elected politicians across 11 countries, as well as survey data from over 12,000 citizens, the paper seeks to map how voters think across several key dimensions. Unfortunately, none of that data includes the United States, for reasons we’ll get into, but it’s still relevant to understanding the American electorate.
[Music]
Demsas: The paper seeks to understand whether voters vote based on policy or identity. Are they short-term or long-term oriented? Do they tend to be single-issue or multi-issue voters? Are they pocketbook oriented, or do they vote based on their perception of the national economy? Do they make decisions based on past performance and behavior, or on what they expect to see happen?
My name’s Jerusalem Demsas. I’m a staff writer at The Atlantic, and this is Good on Paper, a policy show that questions what we really know about popular narratives. To tease out what we know about these questions, I asked Jack to come on the show and give me a better sense of voter psychology.
Jack, welcome to the show.
Jack Lucas: Thank you. Thanks for having me.
Demsas: So a recurring question on this show, and unfortunately in real life, is: Are voters dumb?
What’s your take? Are voters dumb?
Lucas: Well, this is the age-old debate in theories of voting behavior. There are a number of competing theories, as you know, and one of the things we’re looking into in this paper is to try and understand where politicians stand on this question. But if you ask me where I stand on this question, I guess I’m more on the “voters aren’t that dumb” side of the spectrum.
Demsas: It’s a bold claim to make right now. (Laughs.)
Lucas: Yes, I know. (Laughs.) It’s tough times for those of us who believe policy voting is a thing.
But I think that arguments that come out of the democratic realist kind of tradition and the work of Christopher Achen and Larry Bartels in their book, Democracy for Realists, really poses an important challenge to any theories of policy voting that we have to grapple with, and there’s a lot in there that’s, I think, quite persuasive.
But I also think that when we look at the wider picture of examples, where we can really tease apart whether people are voting on the basis of policy versus whether they’re voting on the basis of things like social identity or irrational retrospection, there is pretty good evidence that, at least on issues that voters care a lot about, they’re thinking carefully about policy.
So I don’t want to go all the way down the road to say, you know, I’m hard-core spatial voting, and nothing else ever. But there’s, I think, some pretty good evidence to suggest that voters aren’t perhaps as dumb as they seem, if the only thing you look at is the, kind of, democratic-realist theory of voting and voting behavior.
Demsas: And can you define “spatial voting” for us?
Lucas: Yes, absolutely. We’re right, straight into the jargon. I apologize.
Demsas: No, you’re good.
Lucas: The idea of spatial voting is that all of us carry around in our heads a bundle of policy preferences. You can summarize that bundle of policy preferences with something called an “ideal point,” but this is really just kind of where you live on the left–right spectrum.
And when you vote in an election, spatial voting theory suggests that you look around at the candidates or the parties; you identify the one who is closest, whose policy promises are closest to your own ideal point, your preferred bundle of policies; and you support that candidate or party. That’s sort of the basic logic of spatial voting. And it’s prospective, so people are looking forward at the candidates’ and the parties’ policy commitments, and they’re voting on that basis of future promises. And it’s policy oriented because you’re not voting on the basis of how well things are going in your life or your social identity or any other kinds of things like that. You’re really voting on the basis of your policy preferences. So that’s sort of the theoretical arguments at the heart of spatial voting.
Demsas: So you’ve sort of alluded to this, but there’s a variety of axes on which you could be more of, like, a democratic optimist versus a democratic realist. Though, in my head, maybe realists, I would classify them as pessimists. And there’s whether you’re looking forward (you’re prospective) or you’re looking backwards (you’re retrospective). Are you actually voting on a specific issue, like a single-issue voter? Are you a multi-issue voter? There’s just, like, lots of different ways that you can evaluate whether you’re kind of fitting into one of these two camps.
You’re kind of an outlier, I would say, in political science. Maybe I’m wrong here, but I feel like most political scientists are more on the realist, pessimist side. Or am I reading that wrong?
Lucas: No, I think that’s probably right. I mean, all of the debates that we’re exploring in this paper, we selected because they are genuine debates in political science. So you can find examples—and I’m not talking about obscure arguments that nobody ever looks at—but serious, well-cited arguments on both sides of all of these debates.
So there’s certainly active debates in political science on all of these dimensions. But I think you’re right that, at the moment, you know, we haven’t done a survey of political scientists, so we’ve always talked about how we should, so that we can actually—
Demsas: Oh, yeah. That’d be great.
Lucas: —tell people what proportion of political scientists hold these different theoretical positions. I think that my guess is that, as with everything, most people are kind of in the center. And so what you’re going to see if you talk to political scientists is varying degrees of acceptance of the core tenets of democratic realism or spatial voting or whatever.
But a lot of people will say, as I do, this slightly more hedging kind of argument that there are circumstances in which we would expect to see policy voting matter a lot. And there are a lot of circumstances where voters don’t have sufficient information on candidates, or the election issues aren’t especially salient, or whatever might be happening in that particular election, where people are just gonna fall back on social identities, partisan identities, ingrained rational or irrational retrospection about the past. So you’re going to see different examples of this depending on the election.
So part of the debate after Achen and Bartels’s book, Democracy for Realists, has been—I think much of the debate has been—not so much about, like, Let’s show that they’re completely wrong, but Let’s try to understand the conditions under which the kinds of phenomena they observe in that book happen, and how common are these things?
And so, you know, Anthony Fowler, for example, has a paper where he uses some experimental designs to try to understand how often we would expect people to just vote with their party, regardless of the party’s policy commitments. And he finds it’s, like, maybe a little less than a third of the time. And that’s not nothing. I mean, that’s an enormous proportion.
Demsas: Wait—how does he identify that? How does that paper go?
Lucas: It’s based on a conjoint experiment, and the conjoint experiment adds additional policy items. So you know the candidates’ parties. The way these conjoints work is you get two hypothetical candidates, and you’re asked, basically, which one you prefer.
And so not only does the partisanship of the candidates vary, but the policy commitments of the candidates also varies, and the number of policy commitments varies. So what he shows is that as you add more policy information to the experiment, people are less likely to just vote straight party selection.
And also, as you add more kind of counterintuitive policy commitments on the part of the candidate—like, you have a Republican who is, I don’t know, supportive of income-tax increases on the rich or something—that in those circumstances, you also see that people move away from those candidates. So they’re incorporating the policy information when they can, and about, I think, 29 or 30 percent in that study of the respondents just stick with the party candidate regardless of the policy commitments of the candidate.
So, you know, that’s a large number. The conclusion I draw from that is definitely not Okay, well, Achen and Bartels’s Democracy for Realists is wrong. But it’s also informative to try and understand, like, how often would it be the case that people would just vote for a partisan candidate, even if that partisan candidate doesn’t share some of their core policy commitments?
And obviously, most of the time, these things are observationally equivalent. The Republican candidate is also more conservative on policy, and so it’s really hard to disentangle what is driving the choice of the voter to select that candidate. But there are these cases, both experimentally and also just some weird events that happen in the course of history, where you can try and disentangle that. And when you do, you see that there is some evidence to suggest that voters are able to incorporate their policy commitments into their voting.
But to come back around to your original question: Yes, I think if I were to preregister a hypothesis on this, my guess is that the democratic realists are the majority among political scientists who do voting behavior right now.
Demsas: Turning to your paper, you’re interested in evaluating two things: one, how voters see themselves, and secondly, how politicians see voters. To this end, you designed a pretty massive survey. Can you walk me through what you and your co-authors did? What was the survey? Who did you talk to?
Lucas: Sure. So we wanted to understand on all of these debates that we’ve been talking about where politicians and members of the general public stand, rather than political scientists. So we have theories—political scientists have theories—about voters and voting behavior, and we spend much of our career arguing with each other about those theories. But especially when you watch politicians do their work, or you listen to what they say in speeches, or you even just pay attention to what they say in their memoirs about elections, they, too, seem to have theories of how voters behave and how elections work. And so we wanted to try and measure those among politicians and then also measure them among citizens to have some point of comparison for the politician’s theory.
So what we did was we wrote up eight questions that we hope capture eight of the more important theoretical debates in voting behavior, without kind of technical language or jargon or anything like that, and we just asked politicians and the members of the general public in 11 countries where they position themselves on each of those debates.
So we end up with data from just short of a thousand politicians, as well as data from about 12,000 citizens in the same country. So we can compare the citizens and the politicians within countries. And this is all part of a project that’s led by Stefaan Walgrave at the University of Antwerp called POLPOP, which stands for Politicians and Public Opinion.
Demsas: And what kinds of politicians are we talking about here? I mean, I know these are very different systems. You’re looking at Australia, Belgium, Canada, Czechia, Denmark, Germany, Israel, Netherlands, Portugal, Sweden, Switzerland. These are very different political systems. But are these, like, local? Are these federal? I mean, what kind of mix are you looking for?
Lucas: In this POLPOP project, it’s all national and regional politicians. So these are politicians who are in the central legislature in their country, as well as in some cases, some federal countries like Germany or Canada, where we have really important subnational systems, and we interview provincial or state politicians as well.
So these are kind of top-level politicians, if you think about the ladder of ambition in politics. We’ve done earlier studies with local politicians, but there are no local politicians in this study.
Demsas: As you said, the questions are pretty straightforward. Like, for example, you just ask straight out, “Some say that voters are impatient and think about the short term when they vote. Others say that voters tend to focus on the long term. Where would you position yourself in this debate?” Zero to 10, with zero being short-term.
It’s very, very clear to the politician or the voter—you know, the general public that you’re surveying—what you are actually asking them to do. Why did you choose to have this kind of straightforward question versus, like, a question that could get at this answer, perhaps, by asking like, “Would you prefer a candidate whose policies you think would create long-term economic growth or one who would help address immediate cost-of-living concerns?” What was the benefit of the former?
Lucas: I think that you’re already onto the basic idea here, which is we just wanted it to be maximally clear. These are, in some cases, sort of complicated social-science debates, and we wanted to say, here, Look—there’s two poles in this particular debate. Here’s one pole; here’s the other. Where would you position yourself in that debate? So there are a number of ways to think about measuring this. And we’re thinking about ways to try and do some of those measures in the future. But because this was the first time that anyone had set out to measure these things among politicians in the kind of way that we do, we wanted to be able to make a clear case about what it was that we were measuring.
And so asking sort of straightforward questions, having people position themselves in those debates seemed to us the best way to get started with this sort of work. And this, Some say X; others say Y. Where would you position yourself? is a really common structure for survey questions, and so that was something we thought would feel familiar to the people who are responding, which might also help with some understanding and clarity in the responses.
Demsas: Cool. So let’s get to the findings. What do politicians think of voters, and what do voters think of themselves?
Lucas: So politicians, we say, are democratic realists. That is, they tend to subscribe to theories of voting behavior in which voters are not particularly well informed. They focus on short-term considerations. They focus on single issues rather than many issues. They blame politicians for things that are outside the politicians’ control. All of the sort of key ingredients of that book, Democracy for Realists, seem to be reflected in many politicians’ views.
And we find that three-quarters of politicians across all of our countries fall into that theoretical type. And that contrasts very starkly with the members of the general public, who are much more evenly divided between the democratic realists and the other group, which we call “democratic optimists.” And these are a much more optimistic theory of a policy-oriented, well-informed, long-term-focused kinds of voters.
Demsas: I thought this was really interesting. Most variation that you find is within country rather than cross-country, which is not something I expected. So essentially, it’s just in two of 22 cases that the country someone is from explains more than 10 percent of the variation in their position. To make that maybe even more clear: In the rest, 90 percent of the variation had nothing to do with whether you were Canadian or Portuguese or whatever. Does that surprise you?
Lucas: Yes, absolutely it does. I think it’s really remarkable that in countries as diverse as the ones we’re studying—some have proportional-representation systems, some have single-member plurality systems, some have compulsory voting, and some have mandatory—I mean, it’s just really diverse party systems and institutions. And we see really similar patterns among both the politicians and the comparison between the politicians and the general public across all of these countries. And yes, that was a surprise.
Demsas: But there were two cases where it did explain more than 10 percent of the variation. What were those countries, and what happened there?
Lucas: The two areas where we do see meaningful cross-country variation are the question about whether voters focus on a party’s policy commitments or on the leader’s characteristics and competence—the characteristics of the party leader. And we see there is some cross-country variation on that one for sure. It does make sense because some electoral systems and party systems create more personalized politics than other systems. And so we do see that’s reflected in the results.
And the other difference is on policy versus identity. This is really the question that’s kind of at the heart of a lot of debates about voting behavior these days—whether voters are oriented toward policy commitments that candidates and parties make, or if they vote based on deeply held social identities. And we see that a pretty substantial amount of the variance in responses on that question is explained by cross-country variation but, again, only among the politicians.
So it seems that there is something to be said for the experience of particular kinds of institutional structures or party systems in the development of the theories that you have about voters, especially on these two debates. But on the others, it’s much more consistent across countries.
Demsas: One way to read your results, right—the fact that there’s such a difference between how voters view themselves and, you know, I guess many people would view it as, like, voters have a more rosy picture about themselves and what they’re doing than politicians do—is that voters are lying about themselves. And I think particularly in this moment, where there are a lot of people who are calling voters stupid for one reason or another, it can be tempting to just assume that voters are full of it, right? Like, that they’re just wanting to sound nicer than they are.
But there’s another study that pairs quite nicely with this that just came out recently in PNAS by a friend of the show, David Broockman and his frequent co-author Josh Kalla, about whether political practitioners have good instincts. So what they do is they basically measure the effects of a bunch of messages, real ones that politicians have either tweeted out or, like, the Heritage Foundation, for example, has put in documents, or other partisan outfits have put out about, you know, marijuana or guns or whatever. And then they ask political practitioners and lay people to predict which ones were effective. They find that political practitioners and lay people both performed barely better than chance at predicting persuasive effects. And you don’t really see much difference in those two groups.
This is obviously a different sort of study. They’re looking at different questions here. But there’s something here about our expectations that political professionals really understand the public better than the public understands itself. Like, maybe it’s the case that voters are more correct. So where do you fall in this? Like, I know you already said that you are more of a democratic optimist, but why would politicians be wrong about how voter publics are thinking about voting?
Lucas: Yeah, that’s a great question. One of the things I love about the work that David Broockman and his co-authors, Luke Hewitt and the others, have been doing on persuasion is that, in a way, it reveals how important politicians’ theories of the world really are. Because if it’s extremely difficult to know, for example, what kind of message is going to be persuasive in advance—in the absence of an experiment, kind of A/B testing persuasiveness, it’s really hard to know in advance—that opens a lot of room for people to rely on their intuitive working theories of the world to make decisions about what they’re going to do and how they’re going to behave, and in that case, what kind of advertisements to run, but also how to campaign or how to represent their constituents more generally.
So I think it’s just really difficult to learn about voting behavior from direct experience when you’re a politician. So there’s a lot of noise mixed in with the signal. It looks, based on our evidence, like politicians tend to be surrounded by democratic realists, and so it’s a little less likely that they’re going to update starkly on what seems like policy information. If you’re surrounded by people who have a different theory, you’re more likely to sort of retain that theory.
And just in general, it’s very easy to interpret election results in a variety of ways. So you can imagine a scenario where a party runs an election campaign which is really focused on social identity, in-group appeals, anti-outgroup kind of sentiment, etcetera, and that party loses. And sure, a politician in that party could say, Well, I’m gonna update my theory a little bit and say, maybe, social identities aren’t as important as I thought. Or they might say, Well, we didn’t go far enough. We need to double down. We didn’t really fully commit. I mean, we hear this kind of thing in politics from campaign strategists and politicians and so forth sometimes.
So the feedback mechanism is really noisy, and the way that politicians learn about voting behavior is sometimes—you know, you’re getting conflicting signals sometimes, and it’s really hard to know. So I wouldn’t necessarily say that they’re always wrong, but you can imagine a mechanism where politicians, even though they have an incentive to know true information about how voter behavior really works, could hold theories that aren’t actually accurate.
Demsas: So if voters were right about their own views of themselves, then wouldn’t that sort of imply that they were able to get good information about what elected officials are trying to do? We had this episode a few months ago on the show with Hunter Rendleman, who had a paper about the earned-income tax credit and looking at, in the United States, whether voters were actually responsive to governors expanding this social-welfare benefit in their state. And she is kind of an outlier in this field—well, I wouldn’t say outlier. She’s in the minority in the field of believing that voters, especially at the state level, are able to see and respond effectively to policies that may benefit them, that they say that they like.
So like, I guess my question to you is, like, why don’t we see more of these results in political science then? Why don’t we see voters responding positively to policies that they say that they want, that they like? Why don’t we see politicians recognizing this? Like, doesn’t this kind of push against a lot of our intuitions about how democracy is not really working right now?
Lucas: Yeah, I think it does. I will say, though, that we do have, even now, quite a bit of evidence that things like clear-eyed retrospection and policy commitments among parties still may matter. I mean, we have some evidence, for example, that more extreme candidates who run for office do worse than more moderate candidates coming from the same parties, and that suggests that voters are incorporating the information about the policy positions of these candidates into their choices.
The other thing to say is that I don’t think even the most fiercely committed spatial-voting theorists today would probably not want to claim that voters have the same level of importance or attach the same level of salience to every policy issue. And on a lot of issues, it does appear to be true that voters haven’t thought about it much. A party takes a position on that issue, and they just adopt the position of their preferred party.
But there can be an important difference between policy issues in general and the issues that you really do care about or think about a little more regularly. And we have some pretty good evidence that a lot of people—not all, but a lot of people—do have one or two issues that they care about with particular strength. And if their party abandoned their preferred position on those issues, there’s some reason to think that the voters would notice.
So I think you’re right that the citizens’ responses in the survey reflect a level of optimism that doesn’t seem to be manifested in the kinds of voting behavior that we’re seeing in elections right now. And so recent elections, I think, maybe lead you to think that there’s something persuasive about democratic realism. Let’s put it that way.
Demsas: Well, I guess to steelman it, right, because the case that some people will make about Donald Trump, for instance, who I think is maybe the elephant in the room in every conversation right now, is that, you know, Donald Trump did attempt to moderate on key political promises. When he was running for office, he disavowed—well, he distanced himself, at least—he distanced himself from the most extreme parts of the pro-life movement. How reasonable you find that, given that he appointed members of the Supreme Court who were responsible for the end of Roe v. Wade, you know, your mileage might vary there.
But you know, he did distance himself verbally during the campaign. And there are reports that people believed he was interested in expanding IVF access and that people found that promise to be credible. On Medicaid and Social Security, he’s made multiple comments in public that he doesn’t want to touch. Of course, this was when he was running. Now that DOGE is happening, it’s unclear whether those promises will remain true.
But there is a story you could tell where Donald Trump really attempted on key issues to the electorate to appear more moderate. And there was—I mean, speaking of spatial voting—a survey where they looked at how the average voter saw themselves on sort of a left–right spectrum. And they saw themselves much closer to Donald Trump than to Kamala Harris. And to me, that indicates that, you know, regardless of whether people find some of his other policy positions and commitments to liberal democracy to be disqualifying, the voting public clearly saw many of his policy commitments as closer to where they were.
Lucas: Yes. Well, let me first say, as you well know, there’s quite an industry of research on the correlates of vote choice for Donald Trump these days, and I don’t have a particular comparative advantage on that as a person who mostly studies Canadian politics.
But a couple of pieces to put into the mix here that are absolutely happening: I mean, one, the point you made about policy voting, I think, is true. You see much more attention to cultural policy issues, in general, in recent elections and particular attention to immigration. And on those issues, there was some kind of policy alignment that Donald Trump was able to create, I think, that’s really important.
But the big thing I think I would point out is just this long-run trend toward personalization in politics. This is captured by our “party versus leaders” kind of theoretical debate in the paper—that people seem to be voting more and more on the basis of a particular leader’s character and competence in the eyes of voters, and that’s separate from their specific policy commitments. So even though we still see some evidence of spatial voting these days, it’s also true that when you look at many countries, including the United States, there’s very strong evidence of leader-based voting, where people are voting for somebody because they like that person, they like their character, or whatever.
And they’re adopting the policy positions of that leader kind of after having decided that they really like them. So it’s, If this party leader has this view, then maybe I should have that view as well. And that’s not something that we would expect from spatial-voting theory, and it is something that seems to be happening in many countries, and so I think that’s definitely true.
Demsas: After the break: the foreign national election that is somehow about Trump.
[Break]
Demsas: Well, let’s bring you to Canada, which remains a separate country from the United States. (Laughs.)
In Canada right now, the Conservative Party was slated to win quite handily. But things have changed, both after Trudeau resigned but notably after Trump was elected. And in particular, the kind of bump comes in January, when he is about to come into power and the threat of tariffs was becoming more real. How does that drastic shift towards the Liberals fit within these theories of voting behavior? I feel like there are a lot of stories you could tell.
Lucas: There are a lot of stories you could tell, and I suppose the story is still in the process of being written.
But one of the things that has happened in Canada since Pierre Poilievre became the leader of the Conservative Party of Canada is that we’ve had a radical shift in the kinds of issues that we’re talking about. And so whereas this time last year, we might have been talking about carbon taxes and the excesses of woke politics or whatever, now the focus is trade, economic health, economic survival, etcetera.
And so we have an incoming leader of the Liberal Party who’s a former governor of two national banks, who is, I think most would agree, very competent on economic-policy issues. And that’s occurring at a moment when economic policy is just more salient than it has been in a long time. And there’s some really nice examples in the literature where researchers have seen that for parties, either an issue becomes extremely salient unexpectedly, or parties surprisingly change positions on an issue.
And that gives you an opportunity to see what effect it has on voting. So a policy-voting theorist would expect in Canada that, because trade and the economy have become so salient, voters are going to—maybe a voter who would have otherwise supported the Conservatives, for example, but they trust Mark Carney’s Liberals to deal best with the trade issue—may shift their vote in the direction of the Liberals if it’s about that particular policy being important to them.
Demsas: That raises, for me, a question about how stable you would expect findings about people’s self-assessments and politicians’ assessments of voter publics to be on these political theories. Like, do you think you’ve captured something lasting or just a snapshot of 2022 and 2023, when you were doing these surveys?
Lucas: There is evidence that politicians’ theories of voting behavior do evolve over time. There’s a study by John Kingdon many years ago, which compared candidates who had recently won an election and candidates who had recently lost an election. And as you may imagine, the candidates who lost were a little more cynical about voters’ capacities.
So certainly, politicians learn from experience. They become more or less optimistic, I think, based on how things seem to be going. But we have been asking these same questions of local politicians in Canada each year, going back a number of years now, and we see that there is some meaningful stability to the responses they give.
So we find that there’s some correlation year over year in their responses, in the way that you would expect if their responses are tapping into some meaningful position on the debate. So they’re certainly not perfectly correlated; there’s plenty of measurement error in these questions. But there’s a relationship over time that suggests that politicians have somewhat stable views on this.
Now, among the general public, our working theory of what citizens are doing when they answer these questions is essentially accessing top-of-mind cultural narratives about how democracy works. We don’t want to make the claim that when members of the general public encounter these questions, they think, Well, I’m glad you asked, because I’ve been spending the last eight months thinking about prospective versus retrospective voting, and here’s my position.
It’s more like, What is the most common top-of-mind, culturally available or culturally acceptable narrative about how democracy works? I think that’s a reasonable inference about what those general-public responses reflect. And so the important thing for our purposes is to say, If it’s true that this is your kind of default, culturally accessible story about voters and voting behavior and elections, then it’s remarkable that politicians have developed a very different kind of theory that doesn’t reflect those highly accessible kinds of cultural narratives.
Demsas: Yeah. I wonder, too, about which way the causal arrow goes, right? Like, is it that working in politics makes politicians more likely to have these democratic-realist views of voters? Or that there’s a selection effect going on whereby democratic realists are more likely to become politicians? Does your study give us any insight into that?
Lucas: Yeah, I love this question. I’m sort of obsessed by it, actually. How politicians come to develop these theories, I think, is super interesting. So we have some clues about this, but really nothing more than clues in the paper as it stands.
One thing we can say is that more senior politicians are a little more likely to be democratic realists than more junior politicians. So it’s not a huge difference. It’s, like, 60 percent versus 68 percent when you compare the junior to the senior. But there’s a difference there, which suggests that politicians learn; they become more democratic realist in their theories over time as they move through their political careers. So they’re not arriving on the scene kind of fully formed, hardcore democratic realists, it looks like from the data.
The other sort of clue that we have is that we can compare members of the general public whose demographic characteristics most closely resemble politicians—because, of course, politicians are not a random draw from the public. They tend to be older, better educated, more wealthy, and so forth. So we can look at citizens who match those characteristics of politicians and compare them to the politicians. And we still see there’s an enormous gap between citizens who are kind of like politicians in their demographics and politicians.
And so that suggests there is something like the selection effect you describe going on, where people who have these kinds of views about politics—maybe because they’ve become involved in politics, and they’ve learned these theories from others, more senior people who are involved in politics, early in their career, before they even jump into politics themselves as a candidate—but something is going on to select people who are disproportionately likely to be democratic realists in the first place. And then it seems like the political experience is reinforcing those views over time.
Demsas: So I guess another story I could tell is that the reason why older, more experienced politicians are morally democratic realists is: Maybe cynical people kind of stick it out longer in politics.
Lucas: Could be.
Demsas: And I guess the reason why I am pondering that story is because if you abstract away from politics, if you’re just talking generally about someone with expertise in a field or in a job, you’d expect that over time, that you get better at doing the job, right? Even if you’re just, like, learning from your environment. You’re, like, picking things up generally. But also just in politics, it’s not exactly like the private sector, but there is, like, a win-loss condition that’s extremely impactful, both to your financial interests but also your identity, like desires in life and your self-conception as a successful person. And so you’d expect people to get better over time the longer they stay in there, or at least better at winning elections, which I would think requires understanding how to get voters to vote for you. So I don’t know how plausible you find that story.
Lucas: It’s true that there’s a certain species of very senior politician that has become quite grizzled and cynical about voting and voting behavior. We heard from them because we did these interviews face-to-face with the politicians. So as they work through the survey, because they’re politicians and they generally like to talk, they would sometimes provide a little color commentary as they’re working through the questions. And I certainly remember some very senior politicians in Canada saying, Well, you know, I’m sorry to say it, but voters are not that knowledgeable. You know, that kind of “grizzled old veteran” view of politics.
But yeah, I mean, I think that you would expect there to be some kind of learning mechanism. The tricky thing is, and the thing that we have to ponder here is how involved politicians are in different countries in making decisions about the campaigns they run and the consequences of those campaigns.
So we think it really matters what politicians’ theories of voting and elections are for the choices they make about how they behave as representatives and how they communicate with constituents and maybe how they campaign. But how they campaign is quite variable. Depending on if you’re looking at, like, a closed-list PR system in Portugal or a single-member plurality system in Canada, you have kind of varying levels of control over how you present yourself to voters.
So I think you would expect that the machinery of campaigns would improve over time. But that may not be reflected very directly in the theories that individual politicians hold about voters and voting behavior. And the other twist on this, going back to the paper you mentioned and more generally the research agenda that David Broockman and his co-authors are doing on persuasion: Using the swayable experiment data, they show that the factors that predict whether an advertisement in their case is persuasive are quite context dependent. They change over time. So what predicts persuasiveness in one election looks different in the next election. And so it may be harder to learn. There may be less of a sort of evolutionary convergence toward the optimal election strategy than you expect, because each election has dynamics that are so different from the previous one.
Demsas: So one of the axes I definitely want to ask you about is on sociotropic versus egocentric voting, because I think that’s been a big point of debate in the American context. To define things, sociotropic means, like, kind of voting in the national interest, versus egocentric, in self-interest, sometimes referred to, I think, as pocketbook voting.
So my friend and colleague Derek Thompson had this fun piece titled “‘Everything Is Terrible, but I’m Fine.’” We’ll link it in the show notes. And he points to a couple of indicators where people say that their own personal situation, you know, they’re doing pretty well. Like, on finances, they look at consumer sentiment. But if you ask about the nation as a whole, they think we’re, like, being fed to the dogs. You know what I mean? And it’s, like, this massive divergence between people’s self-assessments and even their local-community assessments and the national context.
You know, we’ve seen this, also, in crime, as well, when you ask people about how much crime is happening, but also about public schools. Like, people usually like their public schools, but they’ll say, like, Public schooling in the country is just absolutely trash. So lots of poli-sci literature shows that voters are sociotropic in their voting, which feels pretty weird to me. I would assume people would be self-interested.
And there was a 2017 study in ASPR by Andrew Healy, Mikael Persson, and Erik Snowberg, which showed that voters do focus on their own pocketbook as well as national assessments of how the country is going, which is a more moderate view. I guess, like most things, people are doing both of these all the time. But what do you find in your survey? How are people thinking about these questions?
Lucas: Politicians overwhelmingly think that voters are egocentric. And citizens are a little more divided, but they tend to think that voters are more sociotropic. And if there’s any one of these debates where one could I most strongly claim that the politicians have it wrong, this is probably the one that you’d pick. I mean, this is actually one where we got a little bit of pushback when we were presenting the paper to some colleagues, and they said, Well, you call these debates, but we all know that it’s sociotropic. So it’s not really a debate, is it? But as you say, there’s the “Digging Into the Pocketbook” paper by Healy et al. And it uses really high-quality Swedish registry data, where you know survey respondents’ actual tax records and their personal financial situation, and it suggests that the pocketbook voting does happen.
There’s some other evidence like this, but I do think that the consensus for a long time—and it was unexpected at first but has grown over time to be, I think, accepted as consensus—is that sociotropic voting is most of what people are doing when they engage in retrospective thinking about: How well are things going? And on that basis, will I reward or punish the incumbent? So on this one, I think you could make a pretty good case that the politicians’ general consensus is not the same as the political scientists’ general consensus.
Demsas: I mean this “sociotropic versus pocketbook” really fits into what happened with the Biden administration. I mean, I hear from Biden officials all the time that they’re like, Well, if you look at people’s personal financial situation, they were doing much better when you look at real wages, or all these, like, debates. But then, of course, you know, a lot about the American economy was great in the last couple years in terms of growth, but inflation as a macro story was really high.
And it’s funny—even as people have understood that inflation was a big part of this election, they still talk about it in terms of their pocketbooks, right? They’re still like, Oh, the reason why inflation was so harmful to the Biden administration’s reelection efforts and, of course, Harris’s election effort was because people had to pay too much for eggs or for milk or for whatever. But it seems like: No. People actually really do care about the national vibe of these things. And I mean, to me, that’s actually a very optimistic view of my fellow countrymen, that, like, Even if I’m doing okay—you know, my wages are pretty good; I got a house in 2021 with a 2 percent mortgage rate—like, I still care that inflation’s doing really badly across the whole country.
Lucas: Yeah. Well, I’m all for your optimistic take on your fellow countrymen. I mean, why not? But I suppose the democratic-realist response, the sort of slight devil’s advocate response here would be that the sociotropic retrospection depends much more clearly on a perception of how well things are going that may be quite a bit less clearly linked to reality. So you kind of know what your income is and how it’s changing and probably know how much your groceries are costing week by week, but to have a clear perception of how well the economy has done over the last four years on a number of indicators is more challenging, and so more susceptible to elite framing and so forth. So I suppose the-glass-half-empty story would be: It’s just much more malleable.
Demsas: I was looking at the response rate from your data collection from the politicians, and it’s highly variable. I was really surprised. I don’t know if you read anything into it. In Canada, where you are: 12 percent response rate, and that’s the lowest of any country you guys reach out to in this study. And in Belgium, the response rate is 85 percent. Does that mean anything to you about the politicians you have or the system? (Laughs.)
Lucas: What it means to me is that my colleagues on this project love to tease me about our low response rate in Canada. So thank you for bringing that up. This is sort of an enduring mystery of elite research that nobody really quite understands.
So there’s two pieces to this. One is that the Belgian team are a bunch of wizards when it comes to attracting and recruiting politicians to participate in their studies. They’re just tireless and really amazing at what they do, and so part of their high response rate is definitely down to their hard work and the credibility that they’ve built over the years with the politicians in that country.
But we also see systematic differences in response rates across countries. When you do surveys of politicians, the response rates in North America are just lower, no matter how much you beg, what strategies you employ. They’re just a lot lower, and nobody seems to know—I mean, even within Europe, there’s a lot of variation—nobody seems to know exactly why. It’s not entirely a function of how many requests they receive. Because we see some very large countries, where the politicians are getting a lot of requests, have low response rates, and some big countries have high response rates. It’s a bit of a mystery, to be honest with you.
But yes, the response rates are quite variable across countries. And even though we have a large number of politicians in total, it is certainly true that more of those politicians come from some countries than from others.
Demsas: Do you worry about selection effects in countries with low response rates? Like, you’re getting responses from people who are especially motivated to engage in this research versus one where you’re getting kind of broad representation?
Lucas: Yeah, absolutely. This is something that we worry about, and it’s especially the case because this isn’t a random sample of politicians. I mean, I suppose there’s really no random samples out there anymore. But this is certainly an opt-in study. We invite them to participate. We send invitations to all of the politicians, and those who choose to participate, they do. And the other ones don’t. So yes, we have to think about this a lot.
And so one of the things you can do is just collect as much background information on all of the politicians in the country that you can, and then you can compare the people who participate in the survey to the full population of politicians to get some sense of how different they are. And what we find is that on most of the observables that we collect, the politicians who participate in our study look a lot like the broader population of politicians. So that’s true on age, gender, on the party they come from, the ideological flavor of the party they come from. On many of those things, we don’t see any particular reason to worry.
The one thing I would flag as a difference between our sample and the population is that the participants in our study do tend to be a little bit more junior in their careers. So the amount of time they’ve had in office is a little less, and it’s not surprising that’s the case, because if you’re a cabinet minister or a prime minister or a party leader or something who’s been around a long time, you probably have less time to spend with us answering questions about politics. So—
Demsas: But that could actually cut both ways, interestingly. Because one way you could read that is like, Oh, are you missing out on people with the most information about voters? But junior people are the ones who just had to win potentially more difficult elections, because they don’t have, you know, the incumbency advantage or the party connections in parliamentary systems. And so, you know, they maybe are actually more closely aligned with what voters are looking for right then.
Lucas: Yeah, that’s exactly the logic I think you have to use to think this through, is if you do have a slight underrepresentation from some group, like, what evidence is there that the people in that group would be different in their responses to the people, when compared to the people who did respond? So we can look at the smaller number of very senior politicians who are participants in our study, and we can say, Is there any evidence that our findings are really particularly influenced by a slight underrepresentation among those folks? And that’s a way that you can kind of test this out.
And we find that the people who don’t participate in these surveys, partly because of how many politicians do participate, they would have to be so dramatically, kind of implausibly different in their views from the people who do participate to change the results. That gives us some confidence. But yes, you’re right. I mean, it could be an advantage. Sometimes you want overrepresentation from certain kinds of participants in order to get their views a little more clearly.
Demsas: So can I ask why you guys don’t have America in your sample? Just because it’s an American show.
Lucas: Yeah. This is a good question. It’s a little bit above my pay grade, but I think that if I can speculate a little or put words in Stefaan Walgrave’s mouth a little bit: I mentioned how response rates in Canada are quite low relative to Belgium or Switzerland and so forth. The response rates in studies of national politicians, especially, in the United States are even lower.
Demsas: Brutal.
Lucas: So the prospect of getting members of Congress to sit down with us for 30 or 40, 45 minutes and participate in this study, I think, are really low. Now, we could still explore maybe state-level politicians or state-level politicians in very big states. There are alternatives here, but for the moment, I mean, this project started in Europe, and it seems to be continuing to expand. So it may be that down the road, the United States is included in future rounds.
I will tell you, though, that we have collected data on these questions in the United States, separately from the POLPOP project. There are local-politician responses. It’s not the national politicians, but we still do have some information about where American politicians stand on these questions. So if you’re curious about the American politicians, I can tell you a little something about them. Although, they’re not exactly equivalent to the others in this study.
So we have data from about, I think, 580 American local politicians. These are mayors and councilors from municipalities above about 1,000 population. And what we found is that two-thirds of the American local politicians are democratic realists. So what that tells me is two things: First, the majority of American local politicians are democratic realists, just like every other country that we study, and second, in Canada, we can compare local politicians to federal and provincial politicians. And we find the local politicians tend to be a little bit less likely to be democratic realists. And so if we extrapolate that logic to the United States, my guess is that you’re probably seeing proportions that are similar to Canada, like 80–85 percent of national politicians espousing these democratic-realist positions. That’s a guess, but it’s an educated guess based on the data that we have from American local politicians.
We also see some differences across party. For the most part, two things: The American local politicians are more democratic realist in character than the Canadian local politicians, and second, there’s not that much difference between the two parties. So only on the question of identity versus policy do we see meaningful, statistically distinguishable differences between the parties, where the Democrats are more identity-oriented theorists, and the Republicans are more policy oriented. So Democrats are a little more likely to think that voters are identity oriented, that they kind of make their voting decisions based on deeply held social identities, when compared to Republicans. So that’s an interesting—
Demsas: When did you do this survey? What year?
Lucas: It was 2022 when we did this survey of American local politicians.
Demsas: So I guess that would indicate that Democratic politicians—well, like, if we assume that voters are correct about why they vote and voters in the United States are similar to the voters you’re seeing in your sample of other Western democracies in your newer paper, that would indicate to me that Democrats are worse than Republicans at assessing voters’ behavior. Is that what you’re getting?
Lucas: It could be. It could be that Democratic voters really are more oriented to social identity. I mean, there’s this long-standing argument that the Democratic Party is more of a coalition of groups, and the Republican Party has been, at least historically, focused on particular kinds of policy commitments. I don’t feel like I’m in a position to adjudicate the persuasiveness of that argument. But it is out there that maybe the factors that predict voting are a little bit different among Democratic voters than Republican voters. So that’s possible.
And it’s also possible that just by virtue of being socialized into a particular political party, you learn slightly different theories from your mentors and campaign strategists and campaign managers and so on, and that there’s just, like, a slightly different subculture in those two different parties.
Demsas: Always our last question is: What is something that you once thought was a good idea but ended up being only good on paper?
Lucas: Well, I’m embarrassed to admit, I have always had a bit of a weakness for the life-hacks genre of online writing, which means I’ve tried many, many things that turned out to be a bad idea. But let me give you a non-life-hack example.
So when I started my job at the university, I was so excited to have gotten an academic position, and I felt like I really wanted to give this new role the respect that it deserved. So I came up with this romantic idea that I was gonna write all of my lectures word for word, and there are gonna be these beautiful lectures that had allusions to poetry and history and so forth.
And I can tell you that the look on my students’ faces when I showed up in a classroom in 2015, and they realized I was gonna be lecturing at them like it was a sort of Oxford College in 1875, they were not thrilled. So I quickly realized I needed to adopt a teaching style that was somewhat more appropriate for the century I was living in.
Demsas: Well, thank you so much, Jack. This was a fantastic conversation.
Lucas: Thanks so much for having me. It’s a real thrill to be part of the podcast.
Demsas: Good on Paper is produced by Rosie Hughes. It was edited by Dave Shaw and fact-checked by Ena Alvarado. Rob Smierciak composed our theme music and engineered this episode. Claudine Ebeid is the executive producer of Atlantic audio. Andrea Valdez is our managing editor.
And hey, if you like what you’re hearing, please leave us a rating and review on Apple Podcasts.
I’m Jerusalem Demsas, and we’ll see you next week.
