How Elites Misread Public Opinion


The recent rise of populist movements in the United States has made it clear that Americans think elites yield too much influence over the country’s political life. On the right, this belief has fueled the rise of figures like Donald Trump, who came to power decrying the influence of “liberal elites” and promising to transfer political power to average Americans. On the left, anti-elite sentiments have crystallized behind progressive figures like Bernie Sanders, whose critique of the top one percent of economic elites has earned him a national following.

Sure, the left and right frame the problems with and solutions to elite influence in radically different ways, but at their core, these movements share a common presumption: To understand what’s gone wrong with American democracy, you have to understand how elites think.

But how, exactly, can we get inside the heads of America’s political elite?

It turns out that political scientists are trying to do just that. As the political scientists Joshua Kertzer and Jonathan Renshon detail in a new paper in the Annual Review of Political Science, the rise of populist movements has coincided with a rise in the use of “elite experiments” — or randomized studies using elite subjects — in academia as a way to study the decision-making processes of political leaders.

Ultimately, Kertzer and Renshon told me in our conversation, the findings of these studies challenge some of our basic assumptions about how democracies should function: For one, elites, even when they try to act on public opinion, often have no idea what the public actually wants. At the same time, these studies suggest that studying the defects in elite decision-making may be the first step to correcting them — since at least in some cases, elites are still responsive to public opinion.

“We acknowledge that political elites are so important to understanding how our democracy functions, so understanding ways to improve their decisions should lead us to understand how to improve the policymaking process more generally,” Kertzer said.

The following transcript has been edited for clarity and length.

POLITICO: Political scientists have traditionally used one set of methods to study mass opinion and another set to study elite opinion. Could you explain those different methods and why they developed?

Joshua Kertzer: When we typically think about experiments in the social sciences, we think about them as a method that researchers use to study ordinary citizens.

Those kinds of methods traditionally weren’t used on political elites, because it’s really hard to haul political elites into the lab. These folks are busy, they don’t want to be poked and prodded by a bunch of researchers, and they’re hard to access. Because of these challenges, people who studied political elites tended to use a very different set of methods. They would read biographies and memoirs, for example, or they would write case studies. You would observe how elites behaved, but you wouldn’t really intervene directly. Maybe you would occasionally interview elites.

But recently, we’ve started to see political scientists use the same kinds of methods to study both populations at the same time. So what if you embed experiments in surveys, and rather than just surveying members of the mass public, you also survey members of Congress and their congressional staff? What if, instead of just doing a laboratory experiment on ordinary citizens, you also take advantage of high-level military leaders? These technological changes and these methodological changes in the social sciences mean that now we can actually use the same tools to study both groups at the same time, which I think is really exciting.

POLITICO: A baseline challenge for the methodology that you’re describing is defining who counts as an elite in the first place. What are some of the definitions that political scientists use to try to decide who qualifies as an elite?

Jonathan Renshon: “Elites” gets sort of bandied about, and there’s not a lot of agreement about a definition.

There are three basic ways of thinking about what an elite might be. One simple model is an occupational model, in which elites are people whose institutional roles give them influence over public policy. Elites in this model would be people who are elected representatives or other people whose institutional authority allows them to do certain things that other people just can’t do.

Another way of thinking about elites is to use a compositional model, in which political elites are a distinctive class characterized by certain socio-economic characteristics — so people who are high in political knowledge, for example, or people who are likely to vote or likely to demonstrate or likely to be involved in politics.

A final way of thinking about elites is that there’s something about cognition that’s important for defining what a political elite is. Under this definition, political elites are characterized by knowing certain things and having certain experiences and expertise. For example, if you’re interested in decision-making about conflict, you would care about elites who have had some experience thinking about conflict in real life — political conflict, international conflict, whatever.

The complicated thing is these different definitions are not entirely distinct. In lots of cases, people are thinking about elites using different but overlapping versions of these models.

POLITICO: A curious element of anti-elite political discourse is that many of the people decrying elites in the abstract themselves belong to elite institutions — members of Congress, for instance, or presidential candidates. Do these different definitions of eliteness help explain this apparent contradiction in populist political rhetoric at all?

Kertzer: Certainly there are people in politics whose political brands are based on them being outside the elite class. Famously, Donald Trump is an anti-elite figure while also owning gold toilets — but anyone who has a gold toilet would probably still count as elite in most compositional models. So there’s a way in which, even if people don’t self-identify as elites, it’s still the case that political scientists would think of them as elites — just elites with a different kind of political brand.

POLITICO: So what do elite experiments tell us about the role of elites in American politics?

Kertzer: Most of our theories of politics in democracies emphasize this notion that there’s public opinion: Our political leaders are paying attention to public opinion, and they’re forming policy as a result. But one thing that these elite experiments are really good at showing is that oftentimes elites are actually misinformed about what the public wants — that elites are fundamentally misreading the room. So in the context of American politics, there are experiments where you go to state legislators, and you tell them more about what their constituents actually think about given issues, and you see if it affects their political behavior. Similarly, in a foreign policy context, there are researchers who have done lots of great studies where they correct misperceptions about what the public wants and see how that affects elite decision-making.

There are a bunch of regularities that folks doing this kind of research have noticed. For example, the public is much more likely to believe that global warming exists, that climate change is real and that it is being caused by human beings than elites think the public believes. Similarly, elites tend to have this assumption that the public is much more isolationist in foreign policy than the public actually is.

Renshon: Another way of thinking about this is that the types of studies that we’re interested in doing are ideally suited to examining whether elites are actually responsive to the public at all, which many theories of accountability and democracy suggest they ought to be. So as an example, some of our colleagues participated in a large, multi-instrument study in Israel a few years ago, and one of the studies that came out of that was about the responsiveness of elected Israeli members of the Knesset to public opinion, in which they randomized information about how supportive the public was about war. They found that consistent with a lot of theories about how public elected politicians ought to act, they were quite responsive in terms of their judgments about going to war based on what they learned about the public’s preferences.

POLITICO: In cases where researchers corrected elites’ misperceptions about public opinion, did they find that elites ultimately wanted to follow public opinion?

Kertzer: Predictably, elites hold differing views on the role of public opinion: For example, work on the U.S. presidency suggests significant variation in the extent to which American presidents believe that public support is required for policy initiatives. Leaders may also believe they’re able to shape public opinion themselves. The challenge, though, is that the bully pulpit has its limits, and you still need to understand where your audience is at in order to bring them on board.

POLITICO: How do social scientists explain instances in which elites misjudge or misread mass opinion?

Kertzer: There are a bunch of different mechanisms that political scientists offer to explain this, one of which has to do with differential contact with constituents. So if you are a politician, how are you forming your judgment about what your constituents think? Maybe your office is looking at the number of tweets you’re getting on given issues, or the number of letters that your office is receiving, or the number of phone calls. Oftentimes it’s the case that people who are on one side of an issue will be much more mobilized than those on the other side, so elites are going to be hearing more from one side than the other, even though if you look at the actual distribution of opinion, it doesn’t look like that at all.

Another mechanism has to do with the role of the media. Oftentimes there’s a template for what news stories look like: You describe an issue, and you talk about one side, and then you talk about the other side, and you treat these things as being equivalent. But when you do that, what ends up happening is that, even if 80 percent of folks are on one side and 20 percent of folks are on the other, there’s a way in which you start to think that the 20 percent is actually much bigger than the 20 percent, because you see both sides presented in this kind of even-handed manner that doesn’t necessarily represent the actual state of the issue.

Think also about gun control, which we know is much more popular with the public than elites often think it is. One of the things that we’re seeing recently in American politics is an effort by activist groups to correct misperceptions about what the public does think to change the nature of the conversation in Washington.

Renshon: As a slight counterpoint, I’ll just note that across a number of studies, Josh and I have found some interesting ways in which there is considerable convergence in how political elites and the general public think about matters relating to war and peace. So just as an example, we found that both Israeli members of parliament and Israeli citizens converge in seeing democracies as less likely to be highly resolved in crises compared to authoritarian countries. And so while there are these interesting cases where elites may misperceive or misunderstand what the public believes, there are other really important ways in which they think about issues related to war and peace in a relatively consistent manner.

POLITICO: What’s stopping elites and elected officials from accessing the public opinion polling that social scientists use to understand what the public wants?

Renshon: Absolutely nothing is stopping elites from using the same public opinion data that academics or the public has access to, and yet we still see compelling evidence that elites misread public opinion, either because of stereotypes they hold about the public, over-weighting their own preferences, or unequal exposure to particular constituencies or special interests. As we saw in the 2020 presidential election campaign, it’s also not unusual for politicians to discount or dismiss public opinion polls when they don’t like the results. In a larger sense, this is not surprising: There are many domains in which access to more or more accurate information doesn’t necessarily reduce the tendency for bias to creep into our judgments.

POLITICO: Why might there be convergence on issues like war and peace but not on issues like gun control or climate change?

Kertzer: It’s worth making a distinction between questions of representation versus substantive properties. There’s a lot of research that’s come out in the past decade that shows that it’s not the case that elites are less susceptible to psychological biases than ordinary citizens — it’s not the case that elites are necessarily cognitively superior to ordinary citizens. Oftentimes they think very similar things. The question is whether elites have accurate perceptions of what the public thinks. The two groups can actually want the exact same thing, but elites end up assuming that the public wants something different than what it actually wants.

The other thing that’s worth mentioning is that there used to be this model of democracy that made the argument that what you want to do in a democracy is to elect the best leaders possible — people who think better than us and who make better decisions — and that that’s how we’ll end up having a better political system. But what you see when you start fielding these elite studies is that the cognitive architecture of political elites and ordinary citizens is actually pretty similar. So then this raises questions about the things that we actually look for when we elect people to higher office. What are the things we actually look for when we appoint people to these high-level bureaucratic positions?

POLITICO: Are there institutional or political reforms that would give political elites a more accurate understanding of public opinion?

Renshon: Most of the work on correcting misperceptions tends to work on correcting biases or fact-checking information, for example, on social media. I don’t know of any work that applies that framework to thinking about how you might correct the misperceptions of elites — although that’s sort of an obvious area where those two research interests might converge. We’re still at the level of identifying those misperceptions, where the next stage would be thinking about fixes or corrections.

Kertzer: I think the broader question is how do political elites learn about the public’s wishes. You need to understand that in order to really understand how to change things.

Similarly, think about ways in which citizens communicate their wishes to elected officials. There are all sorts of other channels through which citizens communicate. Protests, for example, are another way that elites come to understand the wishes of the mass public, and there’s lots of interesting work in American politics showing that one of the reasons why many social movements are successful is precisely because they are able to signal the public’s wishes and the intensity of its preference to elites through organized collective action. So there are all sorts of other mechanisms apart from just constituency representation questions which also are worth thinking about in this context.

But I really liked Jonathan’s comment about how, when you think about correcting misperceptions, it’s often about fact-checking or about combating fake news. There’s a way in which we sort of presumed that these misperceptions only apply to ordinary citizens and that they don’t apply to our political leaders as well. But it’s really a more ubiquitous phenomenon.

POLITICO: What’s the ultimate application of the findings of these elite experiments?

Kertzer: Within political discourse, there really is a growing sense that if you want to understand how politics gets made in the 21st century, you need to understand the role of political elites. What’s nice about elite surveys and experiments is that it’s a way to really get inside the heads of political elites, which allows us to better understand how they make decisions. And then this leads us to other questions — for instance, if there are ways that we can improve their decision-making so that we end up with better decisions and thus better policy. That’s ultimately why we do this kind of research in the first place — because we acknowledge that political elites are so important to understanding how our democracy functions, so understanding ways to improve their decisions should lead us to understand how to improve the policymaking process more generally.