“Ordinary People Are Really Not All That Extreme”: A Conversation with Matthew Levendusky

By Greg Berman

Newsweek reports that a “Plurality of Americans Believe ‘Civil War’ Likely in Next Decade.” The Atlantic documents “The Doom Spiral of Pernicious Polarization.” The Guardian declares that “The US Has Entered a New Phase of Political Violence.” The American political scene is rife with gloomy headlines and predictions of disaster. 

In this context, the work of Matthew Levendusky, a professor of political science at the University of Pennsylvania, is a breath of fresh air. Levendusky is the author of a series of books about partisanship in the United States, including The Partisan Sort, How Partisan Media Polarize America, and Our Common Bonds: Using What Americans Share to Overcome the Partisan Divide. 

While Levendusky is no Pollyana—writing that “animosity and ill will between the parties have become the norm,”—he does believe it is possible to lower the nation’s collective political temperature. “By working on correcting the misperceptions that people have about the other party, you can reduce that level of animosity that they have,” Levendusky said in an interview with the Niskanen Center.

Levendusky has studied a number of different strategies designed to bridge political differences and has identified nearly two dozen ways to reduce polarization and improve attitudes toward the other party. Among them, showing videos to partisans in which people bond after having a political disagreement, or having partisans read news stories highlighting how people on both sides of the political aisle are tired of polarization. 

Levendusky recently discussed his findings with Greg Berman, distinguished fellow of practice at The Harry Frank Guggenheim Foundation. This transcript of their conversation has been edited for length and clarity.

Greg Berman: I’ve read a lot about how Americans have basically sorted themselves into two separate bubbles—red America and blue America. In contrast, you have written that most of us still have very good friends on the opposing political team. To what extent have Americans really sorted into separate worlds, and to what extent is that construct overhyped? 

Matthew Levendusky: I think it’s important to understand the difference between “mostly” and “completely.” If the question is: Are there important differences between the parties? Or: Do members of the two parties value different things? The answer is “Sure.” But does that mean that there’s no overlapping communication between the two sides, that people don’t know anyone from the other party? The answer to that is clearly “No.” 

It’s also important to understand the difference between a survey response and reality—they are not the same thing, and people often confuse them. A survey response is a very artificial scenario with very low cost. That doesn’t mean surveys aren’t valuable—I use them a lot. But you have to think about the difference between what someone says and what someone means. People will tell you all sorts of things in surveys. It’s not that they’re lying, but their answers have to be understood within a particular context. 

Most people are still closer to the middle than the extremes on most issues.

There’s a lot of skepticism right now about surveys. In addition to the usual complaints that the wording of questions can drive certain answers, the consistent underestimation of support for Donald Trump has led some people to argue that pollsters are no longer able to construct truly representative samples. How fair do you think that critique is? 

Modestly fair. There are low-trust individuals who are less likely to be polled. That has probably always been true. The difference is now that maps onto politics in a more meaningful way. If being low-trust doesn’t tell you much about someone’s politics, it doesn’t really matter that much if you miss low-trust people in a political poll. But these days, being low-trust is correlated with a certain kind of politics. That’s the problem for pollsters now. 

Echoing the political scientist Morris Fiorina, my instinct is that the American public hasn’t moved much ideologically since I was a kid, with the exception of a handful of issues like gay marriage and marijuana legalization. Is that an accurate read of the polling? 

Most people are still closer to the middle than the extremes on most issues. Ordinary people are really not all that extreme. Now, you will see significant differences among the donor class and among the kinds of people who are very involved in politics. They tend to look more extreme on one side or the other. 

Look at the recent kerfuffle over book bans. A lot of that was driven by a relatively small number of actors who were deeply involved in politics already. In the school district where I grew up, it was being driven by the daughter-in-law of the state senator from the district. So not exactly a randomly selected person from the district, right? I think if you ask most people, they wouldn’t have felt super strongly about the issue. They might have an opinion, but it is a weak one. I think we see this dynamic over and over again with relatively small groups of people who care intensely about an issue.

What are the biggest misconceptions that Democrats have of Republicans and vice versa? 

Because we tend to interact more with people who are like us than unlike us, we have a particularly bad understanding of people who aren’t like us. Birds of a feather flock together—that’s just true in lots of things. I’m a middle-aged, White, college-educated person, so a lot of the people I know are middle-aged, college-educated White people.

The people we interact with a lot tend to be like us along many dimensions, including politically. If I don’t know a lot of people in the opposite party, I’m going to draw inferences from what I see on social media, or from what I see talked about in the mass media, or from the political elites that I see in campaigns. 

The people that they really dislike are the sorts of people who make everything political—the stereotypical family member who wants to argue with everyone at Thanksgiving dinner. That's the sort of person most people object to politically. 

In general, when people think about those on the other side, they tend to think of them as being more extreme than an ordinary person would be from the other party. I published a paper a few years ago with a couple of collaborators, and we asked people, “How do you feel about the Democrats or the Republicans?” When we did that, the kind of person they’re bringing to mind is not someone who is very typical of the other party. They’re bringing to mind someone who’s quite ideologically extreme and who’s very politically engaged. But if you ask people about someone who looks more like the average partisan—someone who’s not very politically interested and not all about political extremes—they’re basically fine with them. The people that they really dislike are the sorts of people who make everything political—the stereotypical family member who wants to argue with everyone at Thanksgiving dinner. That’s the sort of person most people object to politically. 

The two professional worlds that I know best are the nonprofit sector and academia. In recent years, both nonprofits and universities have been attacked for being almost uniformly left wing. I fear that there is some truth to this critique. Do you share my worry?

I share it, but I think it goes to a deeper issue. One of the most deleterious things that has changed in the United States has been the growing educational polarization divide. This predates Trump. To be honest, there’s not a super clear, compelling explanation for why this has happened. I worry about creating a cycle where access to better opportunities becomes centered in a world that is seen as believing in one side of our political divide, but not the other. Having said that, I’ll believe that Republicans are truly skeptical of universities when they stop trying to have their kids come to institutions like mine. Plenty of Republicans are still fighting very hard to have their kids attend Penn and Stanford and Harvard, and all the rest. So, there’s definitely some political double-speak there. 

It is important to try to get people to communicate better so that they can come away with a deeper understanding of why people believe what they believe. 

In Our Common Bonds, you argue that it is, in fact, possible to change our political climate. How do we do that?  

To be clear: it’s not some magic wand that you just get to wave. You know, on the internet, when you see things like “One weird trick to a flat belly”? Well, there is no one weird trick that we can use to reduce partisanship. It’s a deep problem. It’s less about magic wands and more about moving things on the margin. 

I think a big part of addressing the problem is learning to communicate with those who have opposing views in ways that enhance understanding. It is hard enough for us to have knowledge of ourselves. But we really tend not to do a good job of understanding the rationales of people who are unlike us. It is important to try to get people to communicate better so that they can come away with a deeper understanding of why people believe what they believe. And that means teaching people about civility, about listening, and about asking good questions, with the idea that you’re setting up conversations to be constructive rather than destructive.

Now, sometimes what surfaces in conversations between partisans can be common ground, a sense of “Oh, I didn’t realize that we actually agreed on some things.” But a lot of times the goal is simply to better understand what someone else is saying. The kinds of people who usually want to come to these kinds of conversations are generally pretty political. They tend to be interested in persuading people to come to their side. My pitch to them is always that they will never persuade someone if they don’t understand their perspective and their point of view. I’m part of a working group with many other people across lots of universities and nonprofits trying to think about how to test this at scale. 

Talk to me about the challenge of scale, because one of the things that makes me skeptical about bringing people together for conversations across political differences is that I think it’s hard to scale up. I also worry that any positive effects from these conversations may wear off very quickly as people return to their normal lives.

Are those reasonable concerns to have about these kinds of things? Sure. Again, there’s no magic wand here. These sorts of exchanges shouldn’t be a one-off thing that you do. It’s about building up a set of skills. It’s about rethinking how you approach learning about other people.  

Do you think there is a connection between polarization and political violence? 

Define violence and define polarization. 

I guess I would define polarization as affective polarization. And political violence is using force to try to achieve political goals. 

In a book I published in 2024 with a bunch of coauthors called Partisan Hostility and American Democracy, we asked some questions about support for politically motivated violence. And the relationship with animosity was basically flat. One caution here is that it is really hard to gauge support for violence in a survey. 

The problem with political violence isn't that there's a significant percentage of the public who think violence is acceptable ... The problem is really the seeming justification for political violence that comes from elite actors.

Help me parse the academic debate about public support for political violence. I know that a number of professors argue that we’ve recently seen a significant increase in the percentage of Americans who support the use of violence. But then I look at the work that the Polarization Research Lab and others have done which suggests the levels of support for political violence have been, and remain, very low.

So this is a case where the way in which you ask the question matters a lot. What people are envisioning when you ask them about violence varies from person to person. This is also a case where the inattentiveness of survey respondents matters. The modal person taking a survey these days is doing it on their phone—they’re flipping through and clicking buttons. And the truth of the matter is that a lot of people aren’t paying super close attention. So you’ll get some people who are just inattentively clicking, “Yes” to things. To be clear, it’s not that the support for political violence is zero. It’s that the true support, if properly measured, is probably pretty low.

The problem with political violence isn’t that there’s a significant percentage of the public who think violence is acceptable and that if we put the issue to a vote, we might lose. To be honest, in some ways, I think this whole thing about surveys is sort of a distraction. The problem is really the seeming justification for political violence that comes from elite actors. I think the way violence actually spreads is that a small fringe of people who are willing to engage in violence get messages from elites that this kind of behavior is OK.

You have recently published some research about cancel culture. What did you find? 

Does it exist? Yes. But most ordinary people are not really subject to it. I think the prevalence is low. Most of the examples that you can find tend to be relatively high-profile people. But for me, the more interesting aspect of cancelling is that this is a way of signaling your bona fides to your group. So I think the way to understand these dynamics when they happen is that this is less about genuine beliefs and more a kind of performative action that is designed to signal who your group is. 

Did you find any difference between Democrats and Republicans when it comes to engaging in cancellation? 

If there are differences, they’re relatively minor. 

What is your take on the current fight over gerrymandering? Are the Democrats justified in responding in kind to the Republican gerrymander in Texas? Or is this a case of the Democrats trying to save democracy by engaging in deeply anti-democratic actions? 

I think there is a danger of tit for tat. Partisan hardball is just a bad equilibrium. The fight over congressional districts only matters because of the even balance of power in the country. The parties are so evenly divided that small things can tip things one way or the other. If we were to go back to the 1960s or 1970s, five or six seats aren’t going to make a difference in terms of control of Congress. Frances Lee’s book Insecure Majorities talks about this problem, that part of the reason why our political fights have become so intense is that very small things on the margin now matter a lot. 

The fight over gerrymandering also seems to me like a prime example of the nationalization of politics. I’m wondering how you think of the relationship between hyperpartisanship and nationalization. Which comes first?

I think that they’re pretty endogenous to one another. I don’t really have a strong opinion on which one moved first. I think in some ways, they kind of moved concurrently. The sorts of things that used to promote interest in local politics have just slowly eroded—a sense of place, local media, local elites. Our sense of community is just weaker on lots of different levels.  

And at the same time, the whole idea of what it means to be a member of a party has also changed. If you want a good example of this, look at Congressman Mike Flood of Nebraska. He was interviewed on The Daily about town halls. And at one point, he says something like, “Well, you know, I just have to pick and choose my battles with President Trump. I can’t be fighting with him every time my constituents are upset about some issue. He’s the leader of our party, and I have to go along with what he says.” I’m paraphrasing, but that’s kind of an amazing perspective. I think if you played that clip for Alexander Hamilton or James Madison or John Jay, they’d be pretty horrified by that sort of sentiment, which suggests that we have a separation of parties, not a separation of powers.

The effect of social media isn't zero, but I have not seen any compelling evidence that this is really the thing that drives polarization. I think it is more a case of amplifying tendencies that people already have. 

I’ve recently been reading about the movement, which I think started in academia, towards what is known as “responsible parties.” This is going back more than a half century. There was a deliberate attempt by both Republicans and Democrats to increase ideological polarization so that voters would have a clearer set of options when they went to the polls.

You mean the 1950 American Political Science Association report? 

Yes.

I think that may be a case of be careful of what you wish for. There was a movement among political scientists to do precisely that back then. But I’m not sure how important it ultimately was to the parties. Political scientists, like everyone else, we like to center ourselves. But I think the evidence is pretty thin that Dwight Eisenhower and LBJ carefully pored over that report and used that as the basis of what they should do. I think a bunch of other things intervened to generate the kind of partisanship that we see today. I think ultimately it’s about the incentives that political actors face. 

One thing that hasn’t come up so much in our conversation is social media. I know that correlation is not causation, but for me it seemed like things really started going off the rails around 2014–2015, right around the time when people got smartphones and started to be on social media all day long. 

Does social media contribute to the problem? Sure. That said, there are a lot of studies on this and, at least so far, it’s really hard to generate the kind of empirical evidence you need to show social media having big effects on the way that people think. Going back more than a decade, a team of researchers looked at Facebook data and tried to see whether the algorithm moves people one way or the other. And the answer was, “Yeah, a little bit,” but the much bigger thing is just people’s preferences. The effect of social media isn’t zero, but I have not seen any compelling evidence that this is really the thing that drives polarization. I think it is more a case of amplifying tendencies that people already have. 

If I were to appoint you the czar of American political science, and you had the ability to direct all of the research in the field of polarization in the years to come, how would you direct your peers? 

In true academic fashion, I want to reject the premise of the question. The older I get, the more convinced I am that you need to let people just do their thing. It’s not always obvious what the best and most rewarding thing is going to be. These are just hard problems. They don’t have easy solutions.

---

This interview is part of  The Polarization Project, a series about political division from Greg Berman, distinguished fellow of practice at The Harry Frank Guggenheim Foundation.

Berman is co-editor of Vital City, an urban policy journal, and co-author of the book Gradual: The Case for Incremental Change in a Radical Age. In 2022, Berman led At the Crossroads, a series of interviews for HFG, which addressed the rise in crime in New York City following the pandemic.

This interview also appears in The Fulcrum through a copublishing agreement.