Do Extremist Voters Dominate Primary Elections?

Do primaries attract more extremist voters who skew elections toward candidates like Bernie Sanders and Donald Trump? The common idea has always been that their extremism makes them more likely to vote in primaries. But political scientist Lynn Vavreck says the real story is far more complicated in her paper "On the Representativeness of Primary Electorates".

Listen on Apple Podcasts or wherever you enjoy podcasts.

Transcript:

Wioletta Dziuda:          

I'm Wioletta Dziuda.

Anthony Fowler:          

Hi, I'm Anthony Fowler.

William Howell:

I am Will Howell and we are three professors at the Harris School of Public Policy and this is Not Another Politics Podcast.

Anthony Fowler:          

This podcast is not brought to you by any mattress companies or email providers, but we are open to sponsorship.

William Howell:

Yes we are. It's just the three of us now, isn't it? We're doing as best we can. [music] The primary season is heating up and we see Bernie Sanders performing pretty well.

Tape:   

Senator Bernie Sanders waking up this morning, victorious after last night's crucial nominating contest in New Hampshire.

Anthony Fowler:          

Who has not been formally a part of the democratic party. He's a self-described democratic socialist and yet nonetheless, he's getting real traction in this primary season.

Tape:   

So these results really show deeply divided this race is becoming between progressives and moderates.

Anthony Fowler:          

He's pretty extreme. He wants free college, he wants to forgive all student debts and he wants Medicare for all. It's kind of surprising. Most Americans are not that far to the left. Why is he doing so well?

Wioletta Dziuda:          

One plausible explanation would be people who vote in primaries are very different than people who vote in the general election. People who vote in the primaries are the ones who have extreme positions and they really care about elections and they show up and they elect people that they like that are not necessarily appealing to the rest of the electorate.

Wioletta Dziuda:          

Anthony, you talk to someone who can help us think about this question.

Anthony Fowler:          

I did. I spoke to Lynn Vavreck at UCLA who has written a paper with some coauthors on exactly this question, which is how ideologically extreme are the people who show up to vote in primaries and caucuses? Let's give a listen.

Lynn Vavreck:  

My name is Lynn Vavreck and I'm a political scientist at UCLA where I hold the Marvin Hoffenberg chair in American politics and public policy.

Anthony Fowler:          

We are hoping to talk to you today about your recent study on the representativeness of primary electorates. Can you tell us a little bit about that project?

Lynn Vavreck:  

So my coauthor, John Sides, John and I had started this paper trying to figure out if the people who vote in primaries are ideologically extreme or in some other way distinctive from the people who vote in general elections for the party or the people who in some way are a part of the party following. And because it's not a mystery novel, I'll just tell you, we find that they are not distinctive. So you cannot blame political polarization in the country today merely on the primary process. That's our claim.

Anthony Fowler:          

Okay. All right, well that's good. Let's talk about the evidence and then we'll come back to that claim because that's a bold one, I think.

Lynn Vavreck:  

Almost surely stated with too much hyperbole than my coauthors would like, but I'm sure you'll walk me back.

Anthony Fowler:          

No, that's good. That's good. And that's what we'd like on the podcast anyway. We'd like the bold claims that we can bite into.

So this is surprising findings, certainly. I'm sure if you just asked political scientists or pundits or casual observers, they would all say pretty confidently, it must be the case that people who vote in primaries are more extreme than the general population and more extreme than people who vote in general elections. I guess why do we think that to begin with? Where did we get that impression from?

Lynn Vavreck:  

Well, the conventional wisdom isn't totally wrong. It's just that people make this sort of inferential leap.

So, what we show is that these primary voters, they're more interested in the campaigns than the typical party follower. But then the leap that people make is to say, "Oh, and therefore they must be the people whose preferences are more extreme." Because I think everybody, political scientists and social scientists included, conflate intensity of preference with extremity of position. And there's no reason to conflate those things.

Anthony Fowler:          

Sure, sure. I think people overestimate how many extremists are really out there. Right. I mean—

Lynn Vavreck:  

I think that's right.

Anthony Fowler:          

—most Americans are pretty moderate. And this is something we've known for a while, and I think it's still true in the recent data. You guys are showing, not only do we overestimate how many extremists are out there in general, we overestimate how many of even the very politically active people who vote in primaries and give money in politics and show up to meetings and things like that. Those people are also not all extremists.

Anthony Fowler:          

So tell me a little bit more about the methodology. So you've got a bunch of survey data, you've got people who have answered political surveys. How do we know whether they voted in primaries or not versus generals and how do we know what their ideologies are?

Lynn Vavreck:  

Yeah, so the project starts with a whole bunch of surveys. And then to that we add the administrative database across the states that reports whether someone turned out to vote in elections. And we match that to the survey responses. So we not only know whether the person says they turned out, but we know if the state thinks they turned out and then we know if they turned out in the primaries, if they turned out in the general election.

Anthony Fowler:          

So most people probably ... I mean you did say it, but most people probably ... most of our listeners don't know whether their vote is public record and so that is something we can look up and verify.

And you've done some work on this in the past, but if we had just relied on people's self-reports of whether they voted, we would often get a lot of wrong impressions about what's really going on.

Lynn Vavreck:  

Yeah. You know, people lie. I don't want to say they lie. They're aspirational in retrospect. And it's interesting, the people who will most often misrepresent whether they turned out on election day are quite frankly the people like you and me, people who have a higher level of education on average-

Anthony Fowler:          

Speak for yourself right now.

Lynn Vavreck:  

—are employed. So it's the people who know that there's social pressure to turn out and vote. Who think that the people in their surroundings would sort of make them feel bad if they don't vote. Those are the people who misrepresent whether they voted.

So yeah, that is going to give you some bias. And in fact, in this paper we run the analysis using the validated vote and then using the self-reported vote. And if you use the self-reported vote, you're going to make the conclusion that the primary electorates are more extreme, but when you actually use the validated vote, that difference goes away.

Anthony Fowler:          

So tell me more about the ideology. How do we know who's an extremist and who's a moderate?

Lynn Vavreck:  

So there are a couple of different ways that we try to show this. We have that typical question that anyone who's looked at a political survey or been a respondent in a survey has heard before. Would you describe yourself as a liberal, a moderate, a conservative? Do you think you're very liberal or somewhat liberal? So we have this self-described ideology, so we're going to look at that. Just what do you call yourself? Maybe that reflects your issue positions. Maybe it doesn't. Maybe that's more of an identity, how you see yourself regardless of what it is. If you are someone who is “going to be pushing us” into polarization, you're probably going to either identify or actually reflect your positions as being someone who is not in the middle.

And then we ask people a bunch of policy questions. And we're going to take all of those and we're going to make a scale out of them and come up with an average summary of how extreme a person is. Are they more often than not choosing things that are in the middle or are they choosing positions that are in the tails, strongly agree or strongly disagree, that kind of thing.

Anthony Fowler:          

So you report in the paper that there are not big differences between people who vote in primaries, people who vote in the general and just people who are members of the party, meaning they call themselves Democrats or Republicans or they voted in the recent presidential election for a Democrat or Republican. And that is the striking, interesting finding.

But there are some differences and the differences are on this relatively arbitrary ideology scale. So I just want to get a sense of what substantively this means. For example, democratic primary voters are, let's say 0.1 points to the left of democratic general election voters.

How big ... do you have a sense of what 0.1 means on this scale? If you had to compare it to say the difference between Bernie Sanders and Amy Klobuchar, do you have a way of giving me a rough sense of that?

Lynn Vavreck:  

Sure. I don't have a Sanders Klobuchar number at my ready. Maybe you do. Maybe that's why you asked me—

Anthony Fowler:          

No, no, no. I do not.

Lynn Vavreck:  

—but my guess is that that number is going to be bigger than any number you see in this paper.

We're trying to give people a sense of how different the democratic primary voters are from the democratic general election electorate by comparing them to the differences between well-known senators. So, the one that we have used in the paper is California senators, Barbara Boxer and Diane Feinstein. The difference between those two women who were in Congress when we wrote this paper was two and a half times as big as the difference we found between democratic primary and democratic general election voters.

And so, if you are someone who pays some attention to politics, you're probably saying, "But wait, Boxer and Feinstein, they're not that different." And that is the point. Even those two women who have really similar voting records, the daylight between them is bigger than between the primary and general election voters.

Anthony Fowler:          

I want to come back to the claim that you made, the bold claim that you made, which was about the extent to which who shows up in the primary contributes to political polarization in American politics. It seems to me that ... I mean that's a very interesting question. And if the claim is true, I think that's a really striking claim. It could very well be true.

Lynn Vavreck:  

No, you're right. It is. In fact, as soon as you said it back to me, I'm like, Whoa, wait. And then I was like, oh wait, no, I was the one who said that.

The point that I think we can start with and see if you agree with me here, is that people want to say we have to reform the primary process if we want to reign in polarization. And I think if you take our data seriously, the answer to that question is probably, sure if voters are the problem in the primary system, you could take the voters out of the nominating process. But really they would still be voting in the general election. And since there's a high fidelity ... What we're finding is a high fidelity between the voters in the primary and the voters in the general election. The candidates would still face the same incentives from the general election party electorates.

And so, if you really wanted to get this reform, you'd probably have to take the voters out of both the primary and the general election. You know what I mean?

Anthony Fowler:          

I see what you're saying. But the voters, as we talked about, are not themselves all that extreme. Right. Most-

Lynn Vavreck:  

Right.

Anthony Fowler:          

So I guess I see the argument being somewhat strange that it's the general election voters that are problem. But I guess that's the point. That's partly the point you want to make this that-

Lynn Vavreck:  

Yes. If voters are the problem, if voters are the same ideologically in both of these elections, you have to take them out of both. And that's obviously absurd.

Anthony Fowler:          

Right. Of course, of course. Yeah. We believe in democracy most of the time. Is there any last ... Is there a bold claim you would like to make at the end of this to say what's the one thing that our listeners should take away from this that they didn't know before?

Lynn Vavreck:  

When you find yourself wanting to say these people who vote in primaries they're just so different. Our data are sort of telling us that the differences between the followers of the two parties, that's a really, really big difference. Democrats and Republicans are different from one another, but the difference between the Democrats who vote in the primary and the Democrats who vote in the general election, it's nothing compared to the bigger differences that you see across the parties.

And so, 68% of the people who vote in the primary are also voting in the general election. So if you really want to make the argument that this polarization is coming from the set of people who are voting in the primaries. You've got to be saying that that extra third that make up the general election electorate, that they're really, really different because two thirds of the people are the same people. So that's the thing that I just want people to think about.

Anthony Fowler:          

Sure. Okay. No, that's great. Lynn, thank you much for doing this. It was great talking to you.

Lynn Vavreck:  

Is this the part now where you and the political scientists go talk about the paper while we're not here?

Anthony Fowler:          

Yes, that's right. We don't let you participate in that. Yes.

Wioletta Dziuda:          

So what do you think?

William Howell:

Now we have gotten Lynn out of the room.

Wioletta Dziuda:          

We can—

William Howell:

And it's just the three of us. What do we really think about this paper?

Wioletta Dziuda:          

Let’s just start it.

Anthony Fowler:          

Tell us what you think.

Wioletta Dziuda:          

Yeah, I find it surprising in light of the discussion that we had at the beginning that a lot of people think that the reason why extreme candidates seem to do well in primaries, think about Trump or Bernie Sanders, is because different people show up and vote for them. And she says, well, not so fast. At least I can rule out this as a reason why we have those extreme candidates do so well in the primary election.

So, I found it very interesting. My prior was different.

William Howell:

I, too. I was surprised. I wasn't expecting to see this.

Anthony Fowler:          

I think there are two possible things that might be going on that lead us to think that there's a bigger difference than there really is. And Lynn talks about both of them in the interview.

So one is that some people are just hobbyists, right? Some people are obsessed with politics and they watch the 24 hour news channels constantly and they're the ones tweeting and they're the ones donating to candidates and going to rallies. They love politics almost just like someone else might love basketball or something like that. But they aren't necessarily crazy extremists. They derive lots of entertainment value out of politics and those people are all voting in the primaries and they maybe aren't as extreme as you thought they were.

The other possible explanation is that people lie about ... or they lie or they're aspirational, as Lynn says, about whether or not they voted. And it turns out that the extremists are more likely to be the ones who lie and say, "Oh yes, I definitely voted in the primary."

William Howell:

When in fact they didn't.

Anthony Fowler:          

When in fact they didn't. And so both of those things combined would give a casual observer, if you just talk to all your friends and see what they're doing and see who's putting bumper stickers on their cars and who says they vote in the primary, you would think that it's the extremist voting in the primaries and it's not.

William Howell:

Yes.

Anthony Fowler:          

Can I raise a possibility? Let me just throw out a conjecture.

William Howell:

Yeah.

Anthony Fowler:          

Here's something that I think would be consistent with the data they find in the paper, but yet inconsistent with the claim that the primary voters are actually pretty moderate. So let me try this out.

Suppose every voter has one issue that is really important to them. And that's the issue on which they're going to cast their vote. So some people really care a lot about gun rights. And let's suppose, let's suppose there some Republican voters who gun rights is the issue for them. Maybe on most things, they're moderate. On minimum wage and taxes and welfare and borders and trade and all the other stuff. They're pretty moderate, but gun rights, they really care a lot about it. And suppose it's the case that for the people who care a lot about gun rights, they're extremists. For the person who is willing to determine entirely their vote based on gun rights, that's probably someone who is saying, "You can't take away my AR-15s right out of my cold, dead hands."

Anthony Fowler:          

It might be the case that when these political scientists go to measure people's ideology-

William Howell:

And you average across all these issues.

Anthony Fowler:          

Right. They don't know which issue everyone cares about a lot.

William Howell:

For sure.

Anthony Fowler:          

And so, they average across all the issues and they said, well here's someone who's moderate, but in fact they moderate on average, but on the one issue they really care about, they're an extremist.

William Howell:

And by care you mean the thing that's going to actually informed their vote choice.

Anthony Fowler:          

And so when the candidates come to face the primary electorate and think about where should I position myself, they might indeed, even though on average the primary voting population is moderate, they have a lot of pressure to make sure they cater to those 5% of people who really care about gun rights.

And maybe those people show up in primaries way more than the general population. And so maybe there is this distortionary effect in primaries, but you can't pick it up by looking at average ideology.

Wioletta Dziuda:          

But I wonder whether we need to go all the way the direction of your theory to think about what she's saying?

William Howell:

Is this a theory?

Wioletta Dziuda:          

Yes. It's a theory and Anthony's going to have a theory.

William Howell:

Really? Now he can have a theory.

Anthony Fowler:          

Thank you, Wioletta, I appreciate it.

William Howell:

It felt like more a conjecture. Like a stab in the dark.

Wioletta Dziuda:          

It's the best theory I've seen in the last six years.

William Howell:

Wow.

Wioletta Dziuda:          

So, what's your saying is let’s look at Republicans. They have certain views, so they are going to elect someone who is more Republican than the mainstream electorate, but for that we don't need even that they have the strange preferences where someone cares about guns. We just need that Republicans are different than the general electorate.

I don't think there's anything in her work that would allow us to see whether the position that Republicans hold in somehow resulting in someone who is more extreme even than the Republicans.

Anthony Fowler:          

That's right. I mean, yeah, that's right. I mean, yes. I mean, yes.

Anthony Fowler:          

A way you could explain some polarization from our current primary system, even in light of these results is that although the people showing up in primaries are not wildly different than the people who show up in the general, when they show up in primaries, they divide themselves between these two ideologues.

Wioletta Dziuda:          

Can you actually walk us through this argument again because I think this is a very important argument for us to put forward?

William Howell:

So now it's an argument.

Anthony Fowler:          

Suppose ... this is different one. I had a theory before. Now I have an argument.

William Howell:

Oh, God.

Anthony Fowler:          

So the argument would be, suppose everyone just votes sincerely. Suppose everyone just shows up and just finds the candidate that's—

William Howell:

Who's closest to—

Anthony Fowler:          

—most closely matches their true preferences. They're not thinking forward into the general election or anything like that. You divide all the Democrats over here and all the Republicans over here and let's suppose there's competition on the democratic side and they end up selecting whichever candidate is closest to their median and the Republicans pick whichever candidate is closest to their median.

You're going to get two very different candidates as a result of that process and you're going to get what looks like polarized general elections. Even though the set of people voting in the primary and the general are the same. That's the story or the argument.

William Howell:

Yes. I mean look in the data, we know that Democrats and Republicans obviously are quite distinct from one another in terms of their preferences, but to say that those differences are going to be exaggerated as a function of the primaries. It's a little odd. If the primary voters and the general election voters are indistinguishable from one another, then you would think it ought to be the case that people who then represent their views are going to step forward and try to collect the votes if in fact voters are voting in the way that you've just described.

And so, the idea that the polarization would be exaggerated by virtue of people simply voting for whomever is closest to them.

Anthony Fowler:          

Well, I don't know I would say exaggerated. I just saw you can clearly get a world with divergence between Democrats and Republicans from—

William Howell:

For sure.

Anthony Fowler:          

Yeah, that's all. That's all I said. And if there's enough separation between the median Democrat in the median Republican—

William Howell:

Then we get polarization.

Anthony Fowler:          

—you're going to get two candidates to choose from who are both actually quite far from the median general election voter.

William Howell:

Yes.

Anthony Fowler:          

One funny thing as a backdrop to all of this is that even Amy Klobuchar and Pete Buttigieg or say the more moderate choices right now the Democrats are thinking about, they are still probably, if you had to guess, way to the left of the median voter in America.

William Howell:

In America for sure.

Anthony Fowler:          

Right. And so there is still a question of why isn't there more of this strategic forward looking thinking? I mean if you're a Democrat who really wants to defeat Trump, you'd really be working hard to find a more moderate candidate.

One of the stories is out there is that it's very hard to even find a moderate who's willing to run for office because running for office is not a very pleasant thing. And the only people who want to do it are the extremists themselves. There just aren't very many moderates out there who even want to run.

And if you were going to try to reform the system in some way to reduce polarization, you'd want to work on that problem. So you'd want to think about what kind of electoral institutions and what kind of incentive system—

Anthony Fowler:          

—could actually get a reasonable moderate ... I mean, not to say the moderates are always more reasonable than extremist, but someone who's in between Mitch McConnell and Nancy Pelosi, why doesn't that person want to run for office?

William Howell:

We have this distinction between general election voters and primary voters and we see this convergence in Lynn's paper, but there are other relevant political players who play an important role early on in the candidate winnowing stage. And it would be nice to know whether or not they are more or less extreme as well.

We talked about party leaders maybe playing some kind of role. You could also think about donors, political elites who offer endorsements and that they may be tugging at candidates and pulling them to the extremes. There are these other players that people have to satisfy if they want to get the nomination.

Wioletta Dziuda:          

Yeah.

Anthony Fowler:          

It's possible or at least there's idiosyncratic other ... lots of videos, other idiosyncratic factors that allow an extreme candidate to do well.

William Howell:

Yes.

Anthony Fowler:          

Nevertheless, despite their extremism. That's also a possible explanation is that Sanders has something else going for him. He's tapping into something else that people really like. He's just a ...

William Howell:

They like his grandfatherly way.

Wioletta Dziuda:          

Yeah. I think this is something that, as an aspiring scientists scholar, I shouldn't be saying, but I think one part that's missing from this discussion is that we focus on ideology and on a policy dimension, but it seems that voters care about many other things. And perhaps this is completely rationale. Perhaps I look at what Sanders is suggesting and I look at what a Buttigieg is suggesting, Biden's suggesting. And perhaps they think that whatever Sander's suggesting is actually a little bit to the left of what I would like to have.

But I don't believe he will implement that anyway. We have institutions, he needs an agreement of Congress and I believe Congress is more moderate, but I just like him. He seems honest. He seems very charismatic.

William Howell:

Really?

Wioletta Dziuda:          

And that's it. Yes.

William Howell:

You think he's very charismatic?

Wioletta Dziuda:          

My mother-in-law just crossed you out.

Anthony Fowler:          

Without weighing too much into our views on the charisma or lack thereof a Bernie Sanders, surely we can agree that there are other non-policy things that matter a lot for which candidates we like. There's just competence and are you a corrupt person or are you an honest person? Are you going to do a good job? Are you going to hire good people or bad people? So surely all of those things matter. We're not saying they don't matter, but ideology must matter as well. And so ...

Wioletta Dziuda:          

I know but it might be the ... Okay, I know that the polarization is a little bit more common phenomenon that just presidential elections, but when we think about the last few presidential elections, it might be just a coincidence that in some of them, the candidates who are the most charismatic in some sense or appearing the most honest or had certain qualities that appeal to people were the ones who are also extreme ideologically.

Anthony Fowler:          

To this point, I have a trivia question for you guys. I just looked this up the other day. To this point that idiosyncratic non-policy factors might matter in elections. I looked up how many people voted for Bernie Sanders in the 2016 primary election and then after Hillary Clinton won the primary, how many of those same people switched parties and voted for Donald Trump in the election.

William Howell:

They swung from one ideological stream to the other.

Anthony Fowler:          

Yes. How many ... What's your guesses? So this is conditional on having supported Bernie Sanders in 2016 in the primary. How many of those people—

William Howell:

Months later.

Anthony Fowler:          

—in the general, months later, voted for Donald Trump over Hillary Clinton?

William Howell:

Extraordinarily large.

Anthony Fowler:          

It's about—

William Howell:

I mean it's going to be zero, right? That's what it should be.

Anthony Fowler:          

If ideology was the only thing that matters—

Anthony Fowler:          

It should be zero.

William Howell:

—then it's going to be greater than zero. I'm going to say it's going to be—

Wioletta Dziuda:          

It's definitely, yeah.

William Howell:

—greater than 10% but less than 20%.

Anthony Fowler:          

It's 12%. It's 12%. So there you go. So that was a good guess, which surprised me. It's in some ways it surprised me just how large it was. It may be these people are crazy or maybe they just really care about other things other than this typical ideological spectrum which we talked about.

William Howell: So this is another thing. We need Lynn to go back and she's measuring the wrong things. She's not about the ideology…

Wioletta Dziuda:          

Lynn, are you listening?

William Howell:

—they're extreme on some other thing that needs to be measured.

William Howell:

Okay, what's our bottom line on this paper? Where do you guys come out? What is the thing that you learned or the lesson that you took, having read this, that you're going to carry with you as we wade through these presidential primaries in the next few weeks and months?

Anthony Fowler:          

Crazy primary outcomes are not necessarily the result of crazy voters.

William Howell:

Seems reasonable.

Wioletta Dziuda:          

Yeah, I agree. I'm glad I saw this paper. I think I'm convinced that it's not that the voters are crazy in time. It's just something else that we still have to figure out what it is.

William Howell:

Yeah. I came away feeling vaguely kind of—

Wioletta Dziuda:          

Happy.

William Howell:

—grateful. Well, happy. Yes. I mean—

Wioletta Dziuda:          

That's the goal.

William Howell:

—thank you, primary voters. If you were to pick a delegate to go and spend all the time vetting these people, you'd want somebody who cared about the issue a whole bunch more, but that'd also shared the same basic views. And her data suggests that both those facts are true. These are people who are really interested and care about politics and they also look like the general electorate and that's not so bad. Thank you voters.

Wioletta Dziuda:          

Thank you, voters.

William Howell:

Thank you, primary voters.

Wioletta Dziuda:          

Yes, please go and vote.

William Howell:

Yeah, carry on. Carry on.

Anthony Fowler:          

Thanks for listening to Not Another Politics Podcast.

Wioletta Dziuda:          

Our show is a podcast from the Harris School of Public Policy and it's produced by Matt Hodapp. Thanks for listening.