The topic of this episode is, “What do we know about primary elections?”
Last week America held congressional midterm elections. Some big stories emerged, and one of them involves primary elections, those elections-before-the-general election. Earlier this year, Sen. John Cornyn of Texas lost his primary—in great part due to President Donald Trump, a member of his own party, supporting a challenger, Ken Paxton. Well, Ken Paxton lost that seat to a Democrat, and the vote was not even close. This is a high salience example of a primary election weakening a party’s ability to compete.
But the Cornyn-Paxton situation is but one anecdote, and thankfully we have social science research we can draw upon to give us better insight into whether primaries are good for parties and good for governance.
To help us think through this topic, I have with me Richard C. Barton. He is an Assistant Teaching Professor at Syracuse University’s Maxwell School. He studies how electoral systems, party dynamics, and legislative structures influence political behavior and policy outcomes. Most recently, he is the co-editor of the book, Reforming Primary Elections: Voters, Campaigns, and the Future of Congressional Politics (De Gruyter, 2026).
Kevin Kosar:
Welcome to Understanding Congress, a podcast about the first branch of government. Congress is a notoriously complex institution, and few Americans think well of it. But Congress is essential to our republic. It is a place where our pluralistic society is supposed to work out its differences and come to agreement about what our laws should be.
Rich, welcome to the podcast.
Richard Barton:
Thank you so much for having me, Kevin. I’m very excited to be here.
Kevin Kosar:
Of all the things in the world that you could have studied, or even within the realm of American political science, what got you interested in the niche topic of congressional primaries?
Richard Barton:
That’s a good question. I was a PhD student at Cornell, which is known for an American political development, policy, and historical approach to studying politics, and congressional primaries were not on my radar when I started there. I remember sitting in an American politics seminar where we covered what were sort of the pillars of the literature on congressional primaries.
The big takeaway was that primaries don’t really make much of a difference. They’re not a contributor to polarization. The primary rules don’t really matter. We’ve had primaries in congressional elections going back a century, and a constant can’t be the explanation for a variable. So to the extent that there was something new and problematic happening in our politics, the thought was that primaries aren’t a source of it.
And I’m sitting in this graduate seminar in 2015 as Bernie Sanders is making a serious run in the Democratic presidential primary and Donald Trump is taking over the Republican Party, and it just felt like this literature was missing something, or that something had changed in recent times. It piqued my curiosity. My research was focused more on congressional politics and policymaking, so I gravitated toward congressional primaries more than presidential ones. But that was the initial impetus for why I got involved.
Kevin Kosar:
I believe it. I was in graduate school far earlier than you, but I honestly don’t recall reading a single article on primaries on the syllabi of my courses on the American Congress. Primaries were just an assumed feature of the system; nothing to look at here, nothing to study. They were hiding in plain sight.
You have a new book that came out not terribly long ago, co-edited with Robert Boatright. What prompted the two of you to decide you had to do a book on this subject?
Richard Barton:
Rob has been studying this topic for a while now, and I think he is rightly known as a leading scholar of primary elections. I have become more involved over the last five to eight years. There is more work being done and more interest in the topic, and the literature is growing. But there is still a need for more studies; a lot of unanswered questions remain.
Congressional primaries are an extraordinarily important part of American politics, as we can get into more deeply in a bit. In most districts and states, the primary is the main election of consequence. And yet most of the research focuses on general elections, and, like you said, there hasn’t been as much work done on primaries.
To the extent that there has been work done, there used to be something of a consensus, which I mentioned before: there had been some studies of primaries, but they didn’t find that primaries had any significant impact. Interestingly, Rob Boatright has done the research showing the rise of primary challenges, and ideological primary challenges in particular, and he has very clearly documented this.
But I think Rob has been skeptical that those primary challenges are an important source of polarization, even as they are becoming more common. I think he has maybe come to the position that incumbent legislators are increasingly afraid of getting primaried, but that they really shouldn’t be. Primary challenges are more common, but they’re still the exception. Competitive challenges are relatively rare, and successful ones are rarer still.
I have a different take on this. So Rob and I have been collaborating, but we see the evidence a bit differently. This book project was a way to get a bunch of other smart people who are doing research on this involved—not necessarily to answer all the questions that remain about primary elections, but to focus our attention on the important questions and some of the important debates that are live in the field.
Kevin Kosar:
Kudos to the two of you for getting together to do the book, because social scientific evidence is often quite nuanced. I know there’s an appetite out there for broad-brush conclusions, but social science seldom gives that to us. So for you to join with Rob, who is a terrific scholar on elections and primaries, and to admit up front that you see things a little differently but are nonetheless able to agree to put words on pages that you can both feel satisfied with. That is excellent.
I want to briefly follow up on the idea that maybe primaries don’t matter a whole lot in affecting our politics. I remember encountering that in grad school, back in the day, when I was reading a lot of literature about the incumbency advantage. Legislators get elected, they go to Washington, and they use all the perquisites and authorities of their office to make themselves very reelectable, whether it’s using the frank to send mail to constituents or having access to the media to give speeches and craft a brand and an identity. You look at reelection rates back then, and even today, and they’re pretty darn high. These legislators get there; very often their voters like them, and they send them back again and again. So it’s easy to come away thinking that primaries are just a distraction.
But I remember years ago, my good buddy James Wallner and my professional colleague Elaine Kamarck, who is at the Brookings Institution, wrote a paper with the wonderful title “Anticipating Trouble,” which tried to square the circle. It said, in effect, “Yes, legislators get reelected all the time at a very high rate, but that’s because they’re worried about primaries.” They think primaries are very scary situations, so they spend a lot of time altering their behavior and asking the leaders of the Senate and the House of Representatives to help them get reelected, whether that’s raising money, not taking hard votes, or whatever it may be. So it’s not an either-or of whether primaries matter or people get reelected and primaries are fluff. Primaries matter, but people do get reelected, and that’s how you square it.
Let’s get into the evidence. You have this edited volume with lots of smart contributors, lots of data, and very interesting material. Having wrapped your arms around this mass of social scientific research, what are your concerns about primaries?
Richard Barton:
The place I would start is that primaries are very low-turnout, generally pretty low-salience elections. Whenever a small share of the electorate is making a decision on behalf of the constituents in a given district or state, it raises the question of whether those voters are representative of the broader district. So I think the low turnout of primaries is the first data point that should give us some concern. It suggests that there may be a lack of representativeness in these electorates that could in some way distort representation. There are a couple of questions that follow from that.
But before I get to those, think about how low turnout in primary elections is. It varies across states and districts. In closed primary states like New York, where I’m from, you have to be registered with one of the parties in order to vote in its primary, and you have to register well in advance. There are a number of administrative hurdles that make that a somewhat difficult process, so we’re talking about turnout rates below 5% in a congressional primary. In other states, it might be as high as 10% or 20%, but we’re still talking about a really small share of the electorate.
When you couple that with the fact that, in most states and districts, general elections are not particularly competitive, it is the primary election, with this very select, small group of the electorate, that ultimately decides who the representative will be. In many districts, winning the primary is what it takes to become the representative, because the general election is often a foregone conclusion.
I should note the timing. This is the way I generally talk about the problem, but we’re a couple of days after the 2026 midterm, in which a lot of districts and states that seemed like perfectly safe red seats flipped. I think that should cause all of us to rethink how we view the electoral map, party strategy, and what we think of as safe seats, or the “vanishing marginals,” as Mayhew called them. But generally, I think the point will still hold, despite what was a really interesting electoral cycle.
Kevin Kosar:
This year we did have a wave election, and those will happen on occasion. Seats that were solidly red or solidly blue did, in some instances, flip. But in an average election year, safe seats are called safe seats because they really are safe. If you win the primary, you’re good to go for the general election.
This year was a little different. Not least, you had a president who held his party so tightly to him and incentivized its members to behave essentially as mini versions of him that, when his ratings tanked (he was at 36% a few days before the election, which is historically awful public approval), a lot of people got dragged down with him.
Digging back into the evidence about primaries, one of the things I learned a few years ago when I read the book Who Wants to Run? is that the primary process is not especially attractive to a lot of would-be candidates. You have to raise money, of course. You have to deal with a lot of single-issue interest groups that will put you in a position where you can either be the warrior for their cause and get their money or watch them give the money to somebody else in the primary who is competing with you. And you have to deal with base party voters: people who are either intense partisans or single-issue voters who flock to one party or the other because that party has glommed onto their issue.
As a result, the book argued pretty persuasively that we have good candidates who are staying on the bench, and that the selection process is not good. I’d be remiss if I didn’t mention that Professor Danielle Thomsen also had a book earlier that came to a similar conclusion. Am I remembering this right?
Richard Barton:
That’s right. Andrew Hall and Danielle Thomsen have both published two great books. Their arguments are distinct, and they argue for different mechanisms, but they both fit into what I think of as a supply-side story: fewer moderate candidates are running. I think that supply-side story is an important part of the explanation for why we are getting more ideological nominees in both parties, cycle after cycle.
“Moderate” sometimes sounds a little boring or vanilla, and people rightly ask, “Are we saying that we should just have a Congress of all moderates?” That is not my take. What we are seeing, and what I think is more inherently problematic, is fewer of what I would call ideologically heterodox candidates: candidates who do not fall in line with their party on every single issue.
This is important because if you look at the distribution of the ideology of members of Congress, it looks nothing like the distribution of the electorate. Most voters are cross-pressured on different issues. They are not ideologues who fall down the line on every issue, from immigration to healthcare to foreign policy. They really have a mix of progressive, conservative, and moderate positions, and we see very few members of Congress with that kind of mix anymore. I think that’s problematic. One explanation is that candidates who have an ideological tension with their party’s base are dissuaded from running in the first place.
To get back to some of my disagreement with Rob Boatright, I think there’s a live question here. We know that incumbents worry about getting flanked by a more ideologically extreme challenger, and generally there seems to be an idea that ideological extremism provides some sort of electoral reward in primary elections. Some researchers, including Rob, Mike Cowburn, who is featured in our book, and others, think there’s smoke but no fire there—that the fear is more or less irrational. It exists in the imagination of incumbents but isn’t reflected in reality.
Others disagree. Andrew Hall has done some other work on this, and Steven Rogers has done work at the state legislative level. I also have a book of my own coming out with the University of Michigan Press. What I find, what Andy Hall has found, and what Steven Rogers has found is that there is in fact an electoral reward for being more extreme in primary elections.
And I think there’s a dynamic relationship between that supply-side story and the electoral benefits of extremism. There are a number of reasons why more moderate, more heterodox candidates are sitting out and aren’t throwing their hats in the ring to begin with. But one of those reasons is that they correctly assume they’re going to struggle to raise money and to perform well in a partisan primary.
Kevin Kosar:
Right. Before we get to the possible positives we know about primaries, I wanted to ask: how much research have you seen on legislators’ perceptions?
There has been research for decades on incumbents running scared and overestimating the actual risk they face. If incumbents perceive primaries as a kind of ideological purity test, even if they’re not quite that, that’s pretty significant. If you believe it and your behavior changes based on it, then that’s a real thing. If that’s the way you feel primaries are, that’s an important finding. Do we have more recent evidence about that than what I’ve looked at from years past? Or what’s your hunch here?
Richard Barton:
It’s a great question. I should say that in this conversation, for the sake of being interesting and highlighting a debate, I’m probably overstating the extent to which Rob, Mike Cowburn, I, and others who have studied this disagree.
Directionally, there is an advantage to being more ideological in a primary election. That can be true, and it can also be the case that incumbents who are afraid of getting primaried are overstating the benefits of being ideologically extreme. There can be a penalty for moderation, or a reward for extremism, that is more modest in reality than it is in the imagination of incumbents and other candidates.
Some good recent research on this is a great book by Sarah Anderson, Laurel Harbridge-Yong, and Dan Butler that asked state legislators whether they worry about primary challenges and, if so, what kind. They find that incumbents’ concern about getting primaried from the extreme is quite high, and that it is a motivating factor leading them to reject what they call half-loaf compromises. It leads them to engage in less bipartisan compromise.
There’s also a great chapter in the edited volume that Rob and I have by Haley Cohen, who looks at fundraising records and finds that incumbents are strategically raising money to try to deter possible challengers from getting into the race. So we know that incumbents are sweating primary challenges, particularly those from the extreme, and that it’s actually affecting their behavior.
Kevin Kosar:
Absolutely. My read of the evidence is that there are a lot of single-issue voters among primary voters, and single-issue voters don’t like half loaves. You’re either for guns or against guns. You’re either for abortion or against abortion. Any politician who has to court those voters, and to compete with people in their own party to court those voters in order to make it to the general election, is at least going to be required to flirt with them, if not outright marry them and become their partners in politics.
Richard Barton:
Can I say one other quick thing?
Kevin Kosar:
Yes, go ahead.
Richard Barton:
When I started with the diagnosis of the problem, I talked about the low turnout of primary elections. I would point out that there are two possible mechanisms through which the low turnout of primaries and their unrepresentativeness could be driving polarization or be a major contributor to it.
We’ve been talking about voters, and that’s part of it. We have some research in the edited volume that looks at the ideology of primary voters and their distinctiveness. There is a long-running debate about whether primary voters are distinct, and I’m happy to get into it more. But what I find in my chapter with Carlo Macomber from Unite America is that primary electorates are in fact quite distinct from the general electorate in those districts. And what I find in my upcoming book with the University of Michigan Press, which is called Distorting Democracy, is that the distinctiveness of primary electorates has grown quite a bit over the last decade and a half or so.
The second mechanism is what I would call interest group or donor capture. In these low-turnout races, all the candidates are from the same party. So the fundamentals we think of in general elections, like the state of the economy, how that maps onto the president’s party, and partisanship generally, are not determining factors in primary elections when all the candidates are from the same political party.
And because turnout is so low and so few voters are actually participating, it’s a much better return on investment for a big outside donor or a super PAC to get involved and invest resources in these races, where you only need to persuade or mobilize a very small share of the electorate. That money can go a lot further and be a lot more impactful. In fact, this is something I found in my research.
The impact of money in general elections is often really overstated in the imagination of the American public. We find a lot of null effects in studies of the impact of money in general elections. It looks like Elon Musk’s half a billion dollars didn’t get him very much a couple of days ago. But in primary elections, you throw a bunch of money into the race, and what you’re really trying to do is boost the name recognition of the candidates. When voters go to the ballot in a primary, they’re not looking at Ds versus Rs; they’re looking at the names of the different candidates, so that money has a lot more influence. I just wanted to highlight that it’s both voters and the potential for interest group or donor capture in these low-turnout races.
Kevin Kosar:
One hundred percent. We’ve talked about the negative aspects of primaries. Let me give you a conjoined question: what about the positives? Are primaries good in some way, or is the good so small that we should really look to alternatives to primary elections—different ways of picking candidates?
Richard Barton:
Primary elections were an alternative to conventions—to systems with the classic smoke-filled room, where party bosses decided who was going to be the nominee. In that sense, primary elections were really a democratization of the nominating process. For those who think we should have more small-d democracy, primary elections took decisions over the nomination process out of the hands of a small group of party elites and put them into the hands of voters. If that’s our baseline comparison, I think primary elections are probably better than what preceded them.
Nonetheless, I think they are problematic. As for other benefits, the way I would answer this question is to describe how primary elections are supposed to work in the theory of responsible parties, which is an important intellectual tradition in the study of American politics. Whether it’s primary elections with committed partisan voters or smoke-filled rooms of party elites, the system was supposed to work with the party having tight control over the nomination process, so that we see democracy between the parties rather than within them.
That makes sense if the parties are strategically nominating candidates who will appeal to broad segments of the electorate and who are selected to stitch together a majority coalition by really speaking to the interests, preferences, and problems of the broader American electorate. To the extent that’s what’s happening, primaries might be a perfectly fine system.
My take on what has been happening, especially over the last few decades, is that this is not what primaries have been producing. They have not been producing candidates who are optimized for general election competitiveness and who are really doing their party a service. Instead, I see candidates who are electorally weaker than a strategic party would want to select and who are hurting the party brand by taking positions that are out of step with the broader American electorate.
Kevin Kosar:
Right. As I mentioned in my introduction, Ken Paxton is just one of many examples over the last 10 or 20 years.
But I have to ask quickly about alternatives to primary systems. We don’t want to go back to the smoke-filled rooms, because that was just a corrupt way of doing things, quite frankly, and it does not suit our modern demand for transparency and broader public participation. So what do you think? Are there better ways to build this mousetrap to find good candidates?
Richard Barton:
There are. I should say that a fair number of political scientists actually do think we should put the genie back in the bottle and return to smoke-filled rooms, or some version of that.
I don’t think we should. I think we should further democratize our nominating process, and there are ways to do that that still leave an important role for political parties. There are good models for this already. There are all-candidate primary systems, like the one Louisiana traditionally had, known as the jungle primary, although Louisiana recently repealed it for congressional elections. There are different versions of this in California and Washington state, which use a top-two primary, and in Alaska, which uses a top-four primary—a nonpartisan, or all-candidate, primary.
This is something I’ve researched quite extensively, and what I’ve found is that these systems actually produce more moderate nominees and elected officials than you would have expected had those states been using a partisan primary system. That’s true if we look at Congress, and it’s true if we look at state legislatures.
I should explain what these systems are. I call them all-candidate primaries because all of the candidates, regardless of political party, run in the same first phase: what you might think of as the primary, or the winnowing phase, of the election. Then, in the Louisiana, California, and Washington model, the top two candidates advance to a head-to-head competition. In Alaska, it’s the top four candidates.
What’s really important about these elections is that all the candidates are competing and all the voters are able to participate, regardless of political party. Independent voters and moderate voters form a single voting bloc in these elections in a way they don’t when they are siphoned off into two separate partisan primaries. That’s one mechanism.
But the most important mechanism is that it creates the potential, and in 20% of instances the reality, of same-party general elections. In a blue district in California, you can have two Democrats running against each other, or in a red district you can have two Republicans. It looks a bit different in Alaska, where you can have two Democrats and a Republican running in the general election. These same-party matchups are often frustrating to voters from the party that is boxed out of the general election, but it is actually those voters’ peak political power, when they get to decide which of the two candidates from the opposing party is more moderate. What we find is that, when they’re given the opportunity, in most cases they do support and elect the more moderate candidate.
Kevin Kosar:
We’re almost out of time. You mentioned a new solo-authored book on elections. Do you have a working title for it, and what’s it about?
Richard Barton:
Thanks for asking. The title is Distorting Democracy: How Partisan Primaries Amplify Ideologues. It’s going to be published by the University of Michigan Press, probably early in 2028, just in time for our next primary cycle.
It is about a lot of what we’re discussing here today. It tries to explain why the political parties are nominating increasingly extreme candidates. I start by demonstrating that the parties are doing this, that it’s true in all different kinds of districts, and that it’s not just a feature of the supply side. In fact, when more moderate candidates run, they are less likely to win.
Then I test a few different explanations for why this has been happening over the last decade and a half to a few decades. The first theory I test is that the party establishments have lost control over the nomination process, and that as they’ve lost control, they’ve lost the ability to strategically nominate more electable candidates, which is often strongly correlated with nominating more moderate candidates. What I find is that the party establishments have indeed lost control over the nomination process, starting really with the Tea Party wave under Obama, and then, a few years later, in response to Trump, we saw the Democratic establishment lose its influence.
However, that lost influence is not clearly a source of polarization, because the party establishments themselves were not systematically preferring and throwing their weight behind more moderate candidates. If anything, for the last several decades, they had been pushing the candidate pool and the nominees further toward the ideological extremes.
The other two explanations are things we’ve talked about here. First, primary electorates have become more ideologically distinct over time; Democratic primary voters look less like the general electorate, and ditto for Republicans. I think that growing gap between the primary electorates and the general electorate has been an important source of polarization.
Second is the rise of super PACs and interest group involvement in primary elections generally. Really, since the Tea Party wave, we’ve seen a huge increase in spending, but also in grassroots interest group activity and Astroturf involvement. These ideological interest groups and PACs have strategically targeted different primary elections and changed outcomes in a way that I think has had an outsize effect, reshaping the strategic calculations of other candidates and incumbents and pushing them to be less moderate and less bipartisan.
Kevin Kosar:
I look forward to seeing that book. Today, I’ve been talking with Rich Barton, who teaches at Syracuse University’s Maxwell School and is the co-editor of the book Reforming Primary Elections: Voters, Campaigns, and the Future of Congressional Politics. Rich, thanks for joining us on the podcast.
Richard Barton:
Thank you so much, Kevin. This was great.
Kevin Kosar:
Thank you for listening to Understanding Congress, a podcast of the American Enterprise Institute. This program was produced by Jaehun Lee and hosted by Kevin Kosar. You can subscribe to Understanding Congress via Stitcher, iTunes, Google Podcasts, and TuneIn. We hope you will share this podcast with others and tell us what you think about it by posting your thoughts and questions on X and tagging @AEI. Once again, thank you for listening, and have a great day.
Stay in the know about our news and events.