In this “Conversation with Authors,” we spoke with APSR authors Anthony Fowler and Lynn Vavreck about their open access article (coauthored with Seth Hill, Jeffrey Lewis, Chris Tausanovitch, and Christopher Warshaw), “Moderates.”

APSR: The first question is two-fold. First, where did you get the inspiration for this paper? Second, how did this group of co-authors come together to work on this project?

Lynn Vavreck: Each of us [co-authors] has always been interested in the connection between people’s ideas about policy or issues and how that connects to their political behavior, including vote choice. The idea for the paper is a result of that mutual curiosity and a crazy idea I had to bring people together in an attempt to write a paper in a weekend. Of course, it took 2 and a half years, but we did write the paper because of that weekend.

I got that idea from a close friend at UCLA, a psychologist who always talks about this meeting that her group of social psychologists does every year in North Carolina. I thought it sounded like a good medium for moving knowledge forward, one where we could get a small group together and bring “your data set” and “my crazy idea” together. We went to Newport Beach, California with the authors of this paper and two other colleagues. We started talking about ideas we were interested in, the data that we had, and this is where we ended up.

Anthony Fowler: It was very fun to be a part of that process, gathering with that group of people brainstorming ideas for a paper. It was organic in the sense that no one came into that weekend with a firm idea of what the end-product would be. We spent the weekend debating different ideas and this paper is the result. To the extent that there is a commonality among a group of co-authors, I think all of us are interested in political behavior but also skeptical of some of the strong claims coming out of the political behavior community. We hear claims such as “voters are idiots” or “everything is about partisanship” and so forth. During that weekend, we discussed these claims and their related literature, why we’re skeptical of some of those claims, and how we, with the data available to us, might try to get some traction on these problems. One big puzzle was “why do people seem to give a mix of liberal and conservative answers in survey questions? Is it because they’re idiots who don’t have an ideology, or could they happen to be truly moderate?” We spent some time figuring out how to test this, what methods to use, and that’s where the collaboration of the group really came in.

Lynn Vavreck: The skepticism Anthony mentioned comes from each of us wanting to know more. There are always bounds to these claims, especially the broad claims that Anthony mentioned. We each are drawn to questions like, “Under what conditions? Exactly when, precisely, might that claim be true? Are the conclusions of these claims based on limitations of the data available to us? Are they a limitation of the way that the actual elections, and the actual candidates, are behaving? Could it be that we simply can’t observe the things that we’d like to know?” Our curiosity draws us to push on those claims. We also want to paint a more complete and accurate picture of the electorate.

APSR: What did each of you find the most difficult, either in researching or in writing, this topic or this paper?

Lynn Vavreck: The paper is challenging in that it’s complicated in a few unique ways. One, the method is not intuitive to most people. There are a set of people who will know it well, and there are a set of people who will understand it sort of intuitively. For most people, however, this is not an intuitive method.

Second, working through the findings is complicated, and that’s why we wanted to bring some structure and simplification to those findings by naming the three groups of moderates. We needed to discuss them in ways that are familiar in the literature. And then, of course, it’s complicated to address the method and the findings within the framework and limitations of one paper. There’s a lot packed into this paper—there are very few wasted words. Every sentence is important to understanding the method, the findings, and the implications.

Anthony Fowler: It is a complicated paper, and it is a hard one to present and explain. We can’t just show one table and say, “Here’s the result!” like we can for a lot of other papers.

This process revealed that political scientists don’t typically collect the kind of data useful for answering the questions that we wanted to answer. So even getting data that would allow us to do what we want to do is difficult. Most survey researchers don’t ask a lot of policy questions. When they do, they ask them in a fairly coarse way: “Are you pro-life or are you pro-choice?”

They often ask survey questions in a way that roughly evenly divides the population. I think the way pollsters ask questions might give this illusion of, first, a polarized America that’s evenly divided on issues, even though that’s probably not true. And second, they give this illusion that maybe there’s a large group who are inconsistent or uninformed because they’re giving liberal and conservative answers. I would argue this is an artifact of the questions researchers ask. If we asked more continuous questions, it would be easier to see all these people in the middle. During our research process we realized, one, there’s not a lot of good data to answer our questions and, two, the existing data has many limitations. Knowing this, maybe we’ll do a better job in the future coming up with better ways to measure these things.

APSR: What was the most exciting or surprising part of researching political moderates?

Anthony Fowler: Well of course getting to hang out on the beach, play volleyball, talk to smart people—all that was great!

I was genuinely surprised at how much we were able to get out of the data. Chris Tausanovitch suggested answering some of our research questions using an IRT model. I was very skeptical and pushed back on this, resulting in a productive back and forth. Chris was perhaps more optimistic than he should have been, and I was more pessimistic than I should have been, but we got to a place where we were happy with what we could convincingly glean from the data. It was exciting to be faced with what seemed like an almost intractable problem, yet with the limited data we had available, we were still able to learn a lot. This is a paper that none of us would have written ourselves, and it was only through working with this team that we made this kind of progress.

Lynn Vavreck:  Speaking as someone who’s invested a tremendous amount of time trying to advance survey research, it was exciting to see that only 6% of survey respondents look like they’re guessing at these questions. It might not be the most exciting finding in the paper, but it was going to be hugely problematic for us (and others!) if that number was in the double digits.

A second exciting moment for me came after 2 and a half years of work and getting ready to send it for review, when I said, “This paper needs a title, what are we going to call this thing?” And I just remember Anthony and Chris Warshaw looking at me and saying, “We’re calling it ‘Moderates’”. That’s what had been at the top of this paper for 2 and a half years, but it never dawned on me that it wasn’t a placeholder. But, yeah, of course that’s the title of this paper!

APSR: How might your research impact our understanding of the upcoming 2024 election, or perhaps even survey research?

Lynn Vavreck: What I want to say over and over to reporters and others who ask about this paper is: Most people in America have a mix of liberal and conservative positions on issues. There are very few people providing wholly liberal or wholly conservative answers to any questions you might ask them. This is very counterintuitive to us in a political context that feels calcified, where politics feels like it’s stuck. This is not the same thing as everybody in the electorate being either all liberal or all conservative. The paper’s biggest connection to current politics is understanding that most people have a mix of liberal and conservative positions. The policy items that candidates, parties, and voters prioritize are important: If there are only 3 things we’re fighting over, it’s possible that people do have 3 liberal or 3 conservative positions. Their overall portfolio, though, is mixed.

Anthony Fowler: The biggest puzzle in American politics today is that elected officials are so extreme while voters remain moderate. This is just one paper of many that’s contributing to that puzzle. We don’t directly study elites in this paper, but in looking at our work and the broader literature, the vast majority of Americans are to the right of your typical Democrat in Congress and to the left of your typical Republican in Congress. This is true both on average and issue by issue. Most Americans likely want a minimum wage somewhere in between, they want a tax rate that’s somewhere in between, foreign policy somewhere in between. It’s puzzling why no major political party and few politicians try to change this disconnect and actually cater to the preferences in the middle. You could look at any major election and see this disconnect play out.

If there is a big lesson that comes from this paper, I hope that parties and politicians realize how moderate the electorate is. I think they are gradually learning this lesson, but I hope they are confronted with the facts and forced to acknowledge that they are out of step with the broader electorate. This is probably true even if you’re a Democrat in San Francisco or a Republican in Alabama, for example. At the very least I hope that the public and the elites are aware of these facts. Representation isn’t working very well right now. For electoral accountability to work well, voters have to have good choices rather than merely selecting candidates on whether they’re at one extreme or the other. Is the candidate competent? Are they working hard? Are they engaging in corruption? This representation issue is one of the biggest problems I see with American politics and it’s a big puzzle for researchers. I hope that this paper contributes to that a little bit, and can give people more information so that we can begin to improve electoral accountability.

APSR: What directions might the study of moderates take moving forward, either by you or by other scholars?

Anthony Fowler: I would love to discuss elites and what’s going on with them at our next retreat. Jeff Lewis and I are working on a new project inspired by our “Moderates” paper but applied to members of Congress rather than the public. Considering the findings from both projects, “Moderates” suggests regular voters are perhaps more constrained by a single ideological dimension than we previously thought. In contrast, members of Congress are perhaps less constrained by a single dimension than previously thought. We’re largely studying cases in Congress where a member might vote ‘No’ on something that they ideologically agree with for some non-ideological reason. You see the Squad frequently casting these “protest votes”. We can use the kind of mixture model we have here for voters, but apply it to politicians and do a better job of estimating their real ideologies rather than getting a vote-based average. For example, it may appear that AOC’s coming out as a moderate Democrat because sometimes her votes are extremely liberal and sometimes she votes against the party. This is a fun project that’s come out of this work that Jeff and I wouldn’t have explored if not for working on this paper.

As we’ve discussed, I’d also like survey researchers to ask more creative questions. An example is a perfectly continuous question like, what do you think the minimum wage should be? This is rarely asked. I have a forthcoming co-authored paper using continuous questions. We also looked at party cues, exploring the extent that people shift their positions in response to the positions of elites. People can answer these questions sensibly and look relatively moderate on many questions. We get a nice unimodal distribution of responses, with almost everybody to the right of the Democratic leaders and to the left of the Republican leaders. I would love to see creative questions like these in future research.

Lynn Vavreck: I’ve become re-interested in candidates, largely because of questions like: why is there this mismatch? This is inspired by the work that we did in the final push to complete “Moderates” and work Chris and I (along with John Sides) did during 2020. An interesting thing we found in that work was that, across political party, across age, across gender, across race – the things people thought were the most important things to be voting on were the same. How does that happen? Are candidates setting that agenda? Is context playing a role? Certainly, both things are true. Figuring out who gets to choose what we’re fighting over and what that means for moderate voters is something I’d like to explore.

– Anthony Fowler, University of Chicago

– Lynn Vavreck, University of California, Los Angeles

“Moderates” is a co-authored paper published open access in APSR with additional contributions by:

– Seth J. Hill, University of California, San Diego

– Jeffrey B. Lewis, University of California, Los Angeles

– Chris Tausanovitch, University of California, Los Angeles

– Christopher Warshaw, George Washington University