Not Another Politics Podcast – Episode 36

We’ve been doing this podcast for over a year and we’ve covered a lot of research, but each paper is far from the final word on any topic.

On this episode, it’s time to do some updating. We’re going to take three recent papers and show how they change or deepen our understanding of prior papers we’ve covered on this podcast.

Relevant Papers:

"How Did Absentee Voting Affect The 2020 U.S. Election?"

"Partisan Social Pressure and Affective Polarization"

"Better the Devil You Know: Selective Exposure Alleviates Polarization in an Online Field Experiment"

Prior Papers:

"The Neutral Partisan Effects Of Vote-By-Mail: Evidence From County-Level Rollouts"

"(Almost) Everything in Moderation: New Evidence on Americans’ Online Media Diets"

"How Robust Is Evidence Of Partisan Perceptual Bias In Survey Responses? A New Approach For Studying Expressive Responding"

Listen on Apple Podcasts or wherever you enjoy podcasts.

Transcript

Wioletta Dziuda:

I'm Wiola Dziuda.

William Howell:

I'm Will Howell.

Anthony Fowler:

I'm Anthony Fowler, and this is Not Another Politics Podcast. We've been at this for a little while, guys. We've been doing this podcast for over a year. We've had, how many episodes is it?

William Howell:

35.

Anthony Fowler:

Okay. And so that's a lot of research that we've covered, that's a lot of topics we've covered. And I thought it might be fun for us to spend some time revisiting some of the topics that we've covered because in the time that we've been doing this, there has been new research that has come out that we should cover and we should discuss, because it's not as if each study that we discuss is the last word on the topic. Maybe our podcast is actually motivating people to go out and do even more research on these topics for all we know.

William Howell:

That's right. That's right.

Wioletta Dziuda:

Please send us your work.

William Howell:

That's what's happening. But, and truths that we were chiseling in stone, we're about to smash them. New research is coming out and saying that which we thought we knew in August of last year, in fact, we don't know.

Anthony Fowler:

Sure. Nothing is chiseled in stone in our field. And not very many things anyway. Maybe Wiola will tell us there's some great theory that's chiseled in stone that we'll never question. But for the most part, we're scientists and we care about the evidence and we care about what theory pertains to the questions of the day. And there's always going to be new innovations, and that's why we're doing this podcast.

Anthony Fowler:

One of the topics I think we should revisit is the topic of vote by mail. We interviewed Dan Thompson about his great paper.

Dan Thompson (Voice-Over):

So the project is attempting to answer a question, does voting by mail, do these universal vote by mail systems vantage one party over another or differentially turn out one party over another?

Anthony Fowler:

They find pretty small effects of vote by mail.

Dan Thompson (Voice-Over):

By sending everybody, all registered voters at ballot, The participation rate, the number of people who vote as a share of the population who's eligible to vote increases by about two percentage points.

Anthony Fowler:

And no clear partisan effect.

Dan Thompson (Voice-Over):

Despite a number of high profile people in the media and policymakers making strong claims that one party or the other would be advantaged, it appears as though neither party meaningfully performs better in elections conducted almost entirely by vote by mail.

Anthony Fowler:

And one of the things that we discussed was, well, these results are really interesting, but we don't really know what the effects are going to be in 2020. In light of the global pandemic, the effects of vote by mail might be much bigger because lots of people won't want to go out, all kinds of reasons that might be much bigger. And so there's a group at Stanford, a lot of the same authors as the previous vote by mail study, Jesse Yoder's the lead author. They have released a new paper about the effects of vote by mail in 2020, which I think is pretty interesting. And it answers a lot of the questions that we had raised and speculated about in our previous episode.

William Howell:

How do they do it? Because there's massive increases in turnout everywhere and a relaxation of restrictions on voting and new allowances that are being provided all over the country all at once. We know that turnout increased by a bunch, but how did they get a handle? I mean, it's a hard question to answer.

Anthony Fowler:

They take advantage of that fact. They take advantage of the fact that a bunch of states introduced mail elections for the first time in 2020. They previously had limited mail voting. Maybe it might've been quite rare or you might've needed an excuse to vote by mail. And then a bunch of states adopted and introduced mail elections for the first time, many states automatically mailing ballots to every registered voter. And so that's one part of the paper, is to just do this nationwide analysis and do a diff-in-diff design to see for all these states that introduced mail voting for the first time, was there a big uptick in voter turnout in those places relative to other places that other states that didn't introduce more mail elections or maybe already had mail elections?

Anthony Fowler:

And there they find not much of an effect. It turns out it didn't matter much, even though we had record turnout in 2020, it doesn't look like the increase in the turnout was no different in the places that were introducing mail elections for the first time than it was in the places that say already had mail elections. And then the other analysis that's probably the more central analysis in the paper focuses on Texas, where they have a law that essentially says you need an excuse to vote by mail normally. But if you're over 65, you don't need an excuse.

William Howell:

Kind of a hilarious law.

Anthony Fowler:

So there's an age cutoff there.

William Howell:

You don't need it... Once you're 60... Actually, I think it's a fabulous law. Once you're 65, you could say whatever the hell you want, we will give you what you want. You want to vote? Fine.

Wioletta Dziuda:

Was this motivated by the pandemic or did this law exist before?

Anthony Fowler:

I don't think so. I think the law existed before. And I actually don't know what the origin of the law is. Maybe there was some good rationale for the law. And it actually is funny because 65-

Wioletta Dziuda:

Lots of things happen at 65.

William Howell:

You're old. You just get what you want after a certain point.

Anthony Fowler:

But this is a nice opportunity for them because it allows you to compare. You can compare the 65-year-olds who could have easily voted by mail with no excuse to the 64-year-olds who, for them, it would have been odorous to vote by mail and they would have had to provide a legitimate excuse and so forth. And they have a few different designs and a few different ways they analyze the data, but there, the finding is similar, which is that, yes, 65-year-olds are much more likely to vote by mail than the 64-year-olds, but they're not that much more likely to vote overall. The 64-year-olds are more likely to vote in person. Those two things more or less counteract each other. And so it turns out that vote by mail did not significantly change voter turnout for that group of interest in Texas, the people around 65 in Texas.

Anthony Fowler:

And they similarly look at, is there a partisan effect? They can look at people who were registered Democrats, registered Republicans, is there a big difference there? And again, there's not an especially big difference. So although a lot people seem to think that vote by mail should have really increased turnout in 2020 and could have potentially really helped Biden over Trump, there's no clear evidence in this case that it did.

William Howell:

I'm shocked by that finding. I mean, it's the middle of a pandemic. We were worried about going lining up at the polls, you risk dying doing that. And the idea that these... And then we saw this unbelievable turnout. I mean, this record turnout in 2020, I mean like we've never seen before, as Trump would tell us. And that these allowances didn't facilitate the massive increase in turnout and that it had no bearing on the outcome? It didn't at the margin help out Democrats at all? When we see a political fight that pits Republicans on one side, all trying to reduce the franchise and Democrats on the other insisting, "No, no, no. Anything that makes it harder to vote is something which is," I think they expect, will cut it to their base.

Wioletta Dziuda:

I think it's also interesting in light of our discussion about mandatory voting. So when we were discussing mandatory voting, we were focusing on the cost of voting, that it's just much harder for certain people to vote. And just by decreasing this cost of voting or making voting mandatory, we could increase turnout and improve representation. And here you have this experiment where you essentially decrease the cost of voting on the one hand because you allow people to vote by mail, and increase the cost of voting on the other hand because those poor 64-year-olds, now they, as Will said, they had to stand in line and expose themselves to the virus. And we don't see a sizeable effect. So I think this really makes us rethink our theories of how people vote and whether they do respond to how costly it is for them to vote.

Anthony Fowler:

That is a general finding in a lot of this literature that a lot of these various electoral forms, slight modifications to the cost of voting don't have very big effects.

Wioletta Dziuda:

But here we have a large change in the cost of voting. If you will, a global pandemic really introduces a huge cost, I would say, and we still don't see... I think the study shows that we don't see an effect. Because you could say, "Well, perhaps at the same time, the benefit of showing up to vote increased because we had this hot high stake election." But then we should have seen the people who didn't have to bear the costs of the pandemic, the ones who are voting by mail, they responded even more strongly to this high stake elections. And we don't see that.

William Howell:

Yeah. And the comparison between the 64-year-olds and the 65-year-olds in Texas, on the one hand where we might say as well, Texas wasn't especially competitive and it's working with a margin where turnout was already reasonably high. And we don't know what it would look like among 25-year-olds or 45-year-olds, on the one hand. On the other hand, 65-year-olds were bearing the cost of turning out to vote at much higher rate than would young people because they were at a greater risk of dying. And sure enough, you then see that being allowed to vote by mail, they jump at that opportunity, the 65-year-olds do. But the 64-year-olds, undaunted, done the less turnout in person and say, "We're going to do it come what may." And that's an incredible finding.

Wioletta Dziuda:

Yeah. You pointed out something interesting, that one could say, "Well, perhaps voters in Texas don't believe that the pandemic is actually dangerous to them, that COVID-19 actually affects their health in some important way." But the fact that those that had opportunity to vote by mail jumped on this opportunity, so just that no, perhaps people indeed believed that they shouldn't be exposing themselves to the virus, but nevertheless, we see they went and they voted.

William Howell:

The 64-year-olds did, they said, "Well, I would rather vote by mail as it turns out. I would rather not be exposed to the virus. But not given that opportunity, I'm still going to show up." So are we then to conclude that all the talk about how the only reason why Biden won is because of all the special accommodations that were being made? This paper suggests otherwise. And does it then lead us to conclude that the reason why there was this massive turnout is because it was just a high stakes election?

Anthony Fowler:

Yes, I think that's absolutely right that... To lay my cards on the table, I was wrong as well. When we recorded our episode on vote by mail, I said I believe the study, I think it's really interesting, but I don't think the results would apply to 2020. I think the effects would be much bigger. So I was wrong. A lot of people looked at the very high rate of turnout in 2020 and they said, "Oh, it must be because of all this mail voting, and it must be that there are all these people who want to vote and the costs are just too high for them." And that turns out to not be right.

Wioletta Dziuda:

That's why I really like this paper, because if you think about this narrative that was in the media, that yes, it must have been that people just like vote by mail and vote by mail increases turnout and so on. This narrative made a lot of sense. If you think about it theoretically, it just makes a lot of sense. You decrease costs of voting, people show up. And there's a reason to believe that the effect will be also partisan, and the paper shows that this is not true. So I think it's a perfect example of how research can shed light on issues that we think are actually settled because they make a lot of sense.

William Howell:

All right, Anthony, make the prediction again, but vis-a-vis 2024, that is, if we get a repeat Biden versus Trump in 2024, are we going to see comparable levels of turn-in when presumably there aren't going to be the kinds of accommodations offered in 2020?

Anthony Fowler:

I don't have a clear prediction on that. I mean, it could be that the pandemic did increase turnout because of the stakes, people viewed the stakes as being so high, and maybe it was people responding to the pandemic and saying, "We need to go out and vote because the stakes are really high, regardless of the actual cost of voting or the risk of the virus when you actually go vote." So I'm not going to make a clear prediction about, are we going to see similarly record levels of turnout in 2024? But given this evidence, given now all of this overwhelming evidence, I'd be very surprised if vote by mail had a big effect in 2024, or the retraction of vote by mail had a big effect in 2024.

William Howell:

That seems right. I'd love to see some work done on what was the independent effect in this? How do you tease this out? Really, really hard to do. But what were the effects, plural, presumably, that could be heterogeneous across different segments of the population of the pandemic? On the one hand, it increases the cost of voting, right? On the one hand there's that because you expose yourself. That comes at considerable risk. On the other hand, it may raise the salience of an issue, which leads people to say, "Oh, my God, the government's failing to do something that's going to draw me out." And that may have actually stimulated a level of political engagement that otherwise might not have existed.

Wioletta Dziuda:

I really hope that this paper will make its way to the public discourse, especially when we are talking about vote by mail or various accommodations related to voting. Because yes, despite the fact that they showed that allowing people to vote by mail did not increase the turnout, we have to remember that it actually decreased the cost of voting substantially. So it doesn't change the outcome, but it actually saves people a lot of time, a lot of hardship. And I think this is something that we should emphasize when we talk about the future reforms. And for example, if Democrats support vote by mail, if they are pushing forward, they should really use this as an argument. They should say, "Look, probably we have avoided a lot of spread of COVID-19 because we allowed for by mail, and this didn't affect the result of the election." So I think now we have this extra ammunition for the camp that's supporting all these reforms that are aimed at decreasing the cost of voting.

William Howell:

So, Wiola, did you dig up a paper for us to think about?

Wioletta Dziuda:

Yes. So I really thought a lot about the discussion we had a few weeks ago about Andrew Guess's paper on people's consumption of news on Facebook.

Andrew Guess (Voice-Over):

So the idea is that if the "echo chambers" are real, you're going to get Democrats clustered on the left of this hypothetical ideological continuum of media consumption and Republicans are clustered on the right, and there's not a lot of overlap between the two. Well, I directly compute what this overlap is. And I find so in 2015 that overlap is almost two thirds, so 65% overlap. And then in 2015, it was about 50%.

Wioletta Dziuda:

So liberals and Republicans consume a lot of the same news or a lot of the same media outlets.

Andrew Guess (Voice-Over):

A surprising share of people's visits to politics or news-relevant pages in the data seem to go to large mainstream portal or aggregator type sites. So AOL.com really pops out as a huge source of traffic. Same for MSN.com, and things like Google and Yahoo.

Wioletta Dziuda:

But one question that we were left with was, what does it mean for people's political views, for polarization? Does it really matter whether we consume the same news or not? Does the news consumption shape our political views? So I found a paper that was published just this March in the American Economic Review. So it's a fresh paper. It's written by Roy Levy, who is at the Tel Aviv University. And it's called Social Media, News Consumption, and Polarization: Evidence from a Field Experiment. So Roy solicited people on Facebook to participate in a survey. During the survey, they were supposed to talk about their own political leanings, about their demographic fix and so on and so forth, their news media consumption. But at the end of the survey, they were offered a suggestion that they should subscribe to certain news outlets on Facebook.

Wioletta Dziuda:

It was presented in the following way, for example, "Oh, Wioletta, we've noticed that you have not subscribed to Fox News. If you subscribe to Fox News, you will be exposed to wider variety of information." And people who are on Facebook, they frequently get this kind of suggestion that, "Hey, we've noticed you, haven't subscribed to CNN. Please click here and you will subscribe to CNN." And if you subscribe to CNN, then Facebook will be more likely to put links to CNN articles in your news feed. So that was basically the intervention. After the survey and the suggestion to subscribe to an outlet, Roy was observing the behavior of those people on Facebook. And then he fielded a follow-up survey two months later.

William Howell:

And did he randomize?

Wioletta Dziuda:

Yeah, so he, he basically, based on the survey that you filled out, he immediately knew what channels you are subscribing to. And then he had a list of conservative and liberal channels or report, and automatically you were being given a suggestion to subscribe to four channels that you have not subscribed before. And he randomized whether you receive left-leaning or right-leaning channels.

Wioletta Dziuda:

So what are the findings? There are three actually fascinating findings. The first one is about news consumption. So it turns out that a lot of people follow the suggestion. Even though they were not incentivized to do so, they did follow the suggestion. And moreover, they ended up reading the articles from this new media outlets that they subscribed to when they got linked to those media outlets on their newsfeed. So for example, I'm a liberal, I've never subscribed to Fox News. Suddenly you recommend it to me, I subscribe, I see more Fox News articles, I end up clicking on them.

Wioletta Dziuda:

So this is an interesting finding because it's suggests that our media consumption is not solely intentional. Something as little as a suggestion affects what kind of news I consume. So that was already, I think, quite interesting, which shed some light on what we've seen in Andrew Guess's paper.

Wioletta Dziuda:

The next two findings were about how this altered media consumption changes political views. Roy does not find any effect. You would expect that if liberals are now exposed to more conservative news outlets and conservatives are exposed to more liberal news outlets they would moderate their positions. But the followup survey shows that they actually don't. They seem to hold the same views on current political events as people who were in the control group who were not exposed to new media outlets.

Wioletta Dziuda:

Finally, what he looks at is the impact of this news consumption on affective polarization. So do you know what affective polarization is. You guys know because you're both scientists.

William Howell:

Yes. It's not just how much we disagree, it's how much we hate each other. Right?

Wioletta Dziuda:

Exactly. Exactly. So what he shows is that exposure to media outlets that are not ideologically aligned with you decreases affective polarization.

William Howell:

Were you thinking the opposite? Were you thinking like the liberals who then go into Fox News and then who just start pulling their hair out and start screaming to the headings? Having been exposed to Fox News, that's the wrong image for us to have. In fact, Fox News moderates the liberals, and PBS could moderate the conservatives, if only they would watch. That's the claim?

Wioletta Dziuda:

Yeah. I think you're right, that theoretically it could have gone either way or we could have just found that there's no effect. People mindlessly click on Fox News or CNN, they maybe read only articles that are about celebrities and not necessarily about politics and there's no effect whatsoever. He finds positive, I would attach some value judgment here, a positive effect. So he finds that our polarization, affective polarization goes down.

William Howell:

So in that way, we have reason to worry about siloed channels of communication.

Anthony Fowler:

And if we buy the results from Andy Guess's paper, most people already are exposed somewhat to these opposing arguments and to these opposing outlets.

Wioletta Dziuda:

So I think all those findings together tell actually an interesting story. And here I'm just speculating about the mechanisms behind these findings, but basically they say, "Look, we seem to trust certain news sources, and if you expose us to other news sources that we don't trust so well, we are not going to change our beliefs." Let's say we trust CNN, we read Fox News suddenly, and we are not going to suddenly think that whatever they are telling us is the truth and we are not going to moderate our views. But we're going to understand what kind of information is the other side exposed to, and that's going to let us more sympathize with them, let us understand why they are supporting Trump, for example, or certain policies that we disagree with.

Anthony Fowler:

Wait. So you think that's the explanation? You think the explanation is one of sympathy? Like, "Oh, now that I'm reading Fox News, I see that the dumb conservatives can't help but have these conservative positions."

Wioletta Dziuda:

I don't know. What do you think?

Anthony Fowler:

I mean, that wasn't my first guess. I mean, my guess would be, I mean, I think I've experienced this myself. You have your own policy positions, you have your own beliefs, but then you read the perspective of somebody else and you say, "Okay, I may not agree with them. I might still answer the policy question the same way, but I understand where they're coming from. And I don't think they're a completely crazy person just because they hold a position that's different from mine."

William Howell:

So I too have had that experience, but it's usually when reading somebody who is presenting their counter argument in a thoughtful reasoned way. But so much of partisan media is deliberately inflammatory and deliberately biased, and I guess I would have thought you go over and you see the opposition and they're busy... If you're a liberal and you go over and you'd watch Fox News that's busy owning the libs and bashing the crazy lefties that you might say, "Oh, my God, these people are even worse than I thought. And they're not seriously engaging my arguments and they're not really presenting any meaningful arguments of their own," which would lead you to go in the opposite direction. And so it's in that way that I'm surprised by the finding.

Anthony Fowler:

I mean, I would guess that it does vary by outlet. I would imagine that liberals reading Breitbart would not make them all that sympathetic toward conservatives, but maybe liberals reading the National Review, they would say, "You know what? I don't agree with these writers of the National Review, but I understand where they're coming from and they're making a reasonable, good faith argument." And maybe similarly a conservative who's reading the Washington Post might say, "Okay, I understand where they're coming from." Whereas the conservative reads the especially kind of outlandish, they read something in Salon or Vice that's over the top and they say, "All right, this is why I hate these people." So I think, I don't know if that's something that's tested in the paper, but you would guess that it would vary a bit by the kind of outlets and by the tenor of the article and so forth.

Wioletta Dziuda:

I'm not really on Facebook so now I'm going to talk about something that I have no idea about. But my guess would be that people's-

William Howell:

It's never stopped you before.

Wioletta Dziuda:

Yes, exactly. My guess would be the people's news consumption on media is on the lighter side. I don't really click on links and then engage in an in-depth reading of some National Review article. I think, I would guess it's just more like those catchy titles that are short, that have a very short descriptions of the recent events and so on that people are clicking on. So again, I think your hypothesis is very interesting, Anthony, but my guess would be that it's not because we see this thoughtful explanation on the other side. I think there must be something else there. And I don't know, it might be just like, "Oh, my God, now I see what crazy stuff is going on on the other side so I'm less upset with this." I don't know, actually. When I say this it doesn't sound actually super plausible, but-

William Howell:

But no, I think you're onto something, because the explanation for the finding has to not only account for why anger or hatred dissipates, but why one's views remain unchanged. It's not that I consider the other arguments and I come around to them in any way. It's that I see the other arguments or the presentation of perspectives from the opposition and I hate the opposition a little less, but I still think what I think.

Wioletta Dziuda:

Yeah. We have to have a theory that explains those two at the same time. And again, I just want to point out that the second survey was fielded out two months after the intervention. So we are talking about, I would call it relatively short-run effects. Those effects might change in the long run. If we expose people to media that are not aligned ideologically with them for a long time, perhaps even their political attitudes are going to change. Who knows? But at least two months after, this is what we see.

Anthony Fowler:

If this active affective polarization thing is a thing, if it really is something that's important to people, you might expect that maybe there's no short-term effect on policy positions, but maybe there is some long-term effect. Maybe today I understand where conservatives are coming from as a liberal and then tomorrow maybe when they raise some new policy proposal that I've never thought about before, maybe I'm more amenable to it now.

Wioletta Dziuda:

Yeah. I think that's a good point, that the decrease in affective polarization, my leaders to understand each other's position a little bit more in the long run and actually convert a little bit more to the same position. So can I mention one other finding that's in the paper that's a little bit octagonal to what we talked about, but I think it's very interesting. Roy ends up having people who are liberal and give a suggestion to subscribe to a conservative news outlet and they do, and vice versa on the conservative side.

Wioletta Dziuda:

And what he sees, that if a person who identifies herself as liberal subscribes to a liberal outlet, Facebook indeed populates her newsfeed with news articles from this outlet quite a bit. But if a liberal person subscribes to a conservative news outlet, the Facebook does not populate the newsfeed as much in this case. So Facebook's algorithms seem to try to understand what kind of ideological leanings you have and try to bias your news consumption in that direction, which I'm guessing that wasn't news to a lot of people. It was news to me. And I think it opens up a lot of interesting topics and conversations.

William Howell:

So I thought this was a paper about citizens and news consumption, but in fact, it's a paper about Facebook.

Wioletta Dziuda:

I mean, it tells you that together with all the other findings of this paper that Facebook does contribute to some sort of polarization. Perhaps it's not ideological privatization, but it's a polarization in how we view the other side of the political spectrum. And that's possibly a negative impact on the Facebook on the world.

Anthony Fowler:

No, it's somewhat concerning, I think, to say the least just how much influence outlets like Facebook have on what kind of information we're exposed to. We have to get our information from somewhere, so of course, Facebook is providing a service for people, but at the same time, it is somewhat concerning how much influence they have often without us even knowing it. I am at least aware of this kind of thing when I browse through my newsfeed and I know if I click on this article, I'm going to get more articles like it tomorrow. And I sometimes even think about that. I want to read this article, but I don't want to see 12 of these articles in my feed tomorrow, so do I really want to click on it or not? I think it's a strange, yeah, it's a very concerning thing.

Anthony Fowler:

Have you ever wondered what goes on inside a black hole, or why time only moves in one direction? Or what is really so weird about quantum mechanics? Then you should listen to Why This Universe. On this podcast, you'll hear about the strangest and most interesting ideas in physics broken down by physicists Dan Hooper and Shalma Wegsman. If you want to learn about our universe from the quantum to the cosmic, you won't want to miss Why This Universe. Part of the University of Chicago Podcast Network.

William Howell:

Well, as it turns out, I too read a paper that's relevant for our ongoing discussions about affective polarization. And you'll recall, last summer, we had a really interesting discussion with Greg Huber at Yale about some work that he had done in which he shows that the factual misperceptions that we're accustomed to ascribing to partisanship attenuate a whole bunch when you pay people to tell you what the truth actually is.

Greg Huber (Voice-Over):

What's the unemployment rate? A Democrat say under a Democratic president, "Oh, it's 2%," and a Republican might say, "Oh, it's 10%." And the truth might be something like four or 5%. So they're both going to their corners defending their president. But now let's say I say something like, "Well, if you get it right I'll give you a dollar." Well, what happens then? And the first thing we find is that there's a fair bit of convergence between the two parties towards a common answer, not always the truth, but away from those partisan corners when we put money on the table.

William Howell:

So what's going on here? This whole literature on affective polarization suggests again that we don't just disagree with one another, but that we hate one another. And that that carries over into differences, not just with regards to policy opinions, but also with regards to issues involving who you marry, what friends you are willing to keep and what you actually think is objectively true about the state of the world. And so you'll see in surveys, big differences between what Democrats and Republicans will say about things that are objectively verifiable. And they say, "You see, we all live in different worlds. We all are just, it's Venus and Mars all the way through. Not between men and women, but between Democrats and Republicans." And what Greg came in and said, "Not so much."

Greg Huber (Voice-Over):

If the partisans differed in their beliefs about the truth, we wouldn't think putting money on the table would have any real effect on the answers that they would give.

William Howell:

A lot of what's happening here when you see this stuff is partisan cheerleading.

Greg Huber (Voice-Over):

They're not just answering the question you ask them. They're answering the question they want to answer. They're trying to express an opinion often. So one can think of an analogy where you go to a restaurant and you have a wonderful meal and the service is terrible. And the next day I call you and say, "How was the food?" And you say, "It was terrible." And you're not telling me what you thought of the food, you are telling me what you thought of the service.

William Howell:

It's that when you survey Republicans about a set of facts that are potentially damaging to prominent Republican leaders, they're going to offer one kind of answer. And if you survey Democrats, they're going to offer a different kind of answer, not because they live in different worlds and don't get what's actually happening, but that the survey's an opportunity for them to cheer on their own party. And that with some small monetary incentives, you can make that effect go away. So that's interesting as we try to make sense of this thing called affective polarization. What is it? What's behind it? Part of it might not be that we're, again, living in different worlds, it's a function of cheerleading.

William Howell:

And I read this paper, it's a working paper by Elizabeth Connors who's an assistant professor at the University of South Carolina. It's called Partisan Social Pressure and Affective Polarization, in which she suggests that there are social dynamics in play as well. That people worry about how they're perceived, they're worried about how they're perceived by their co-partisans and so they perform. And one of the ways that they perform for members of their own groups is by frankly overstating the extent to which they dislike the opposition because they think that that's going to be rewarded socially. Conversely, the Democrat who says, "Well, I don't think Trump's that bad," risks being scorned by her peer group among fellow Democrats because the lines could be, "Are you kidding me? You don't get it? What's wrong with you?" And so there's this performative aspect.

William Howell:

She does the survey and she asked the standard thermometer rating, "How warmly do you feel towards Republicans, how warmly towards Democrats?" Which is what the basis is for these measures of affective polarization. And in one condition though, she asked people, not just, "What do you think of the opposition party?" But she says, "Please answer these questions as you think a Democrat wanting to impress other Democrats would. Or alternatively, as you think a Democrat wanting to disappoint other Democrats would." And there you see just radically different responses to the thermometer ratings and much lower levels of affective polarization when people are asked in the disappoint category as opposed to the impress category.

William Howell:

It's not the experiment I think I would have run. I think I would have instead characterized this as you could tell people about the opposition views, you could trigger them to then think about the extent to which they're being watched or not by their peers as a treatment condition. Instead, she's asking people what they think other people would say, which is a curious construction. Nonetheless, she sees these huge differences, which are at least indicative of or open up the possibility that these social concerns play an important role.

Wioletta Dziuda:

But she also runs an experiment where she asks people to express their own opinions, but she primes them with a sentence, "Oh, remember, those answers will be private." And in some other treatments, she primes them by saying, "Well, this is all going to be published online." What does she find there?

William Howell:

Well, and then there's a control condition which she says nothing. And what she finds is that on average, there's really not any difference between these various conditions. But among a subset of people who score high in self-monitoring, that is, these are people when you ask them a set of questions, like, "Are you often the center of attention? Do you act a lot? How aware are you of other people's views?" What she finds is that the subset of people who are assigned to the private treatment condition, that is, they're assured that their answers are not going to be released to the public, they report significantly lower levels of affective polarization than do those people who are not offered this reassurance and who nonetheless care a lot about what other people think of them, which is again, consistent with the idea that it's not that everybody worries about these social pressures, but there's a subset who do. And those who do, you can elicit lower levels of affective polarization when you reassure them that nobody's watching.

Anthony Fowler:

I like the story. And this is more or less the story that we've told before when we discussed Shanto Iyengar's paper on the affective polarization. I more or less made the point that I thought... I mean, of course, it's fun to give this partisan response to these questions in the survey, but then when it actually comes to going out to dinner with people, you often don't even know whether or not they're Democrats or Republicans and you don't even care. And so that was my hunch, was that in fact, a lot of this is performative in some sense, and it might even be performative in surveys.

Anthony Fowler:

Given that I was surprised that there's no mean difference between these private versus public conditions, and I know it's not a super strong treatment, but given that story, you would have thought there should be a difference on average. So I was maybe surprised that there wasn't an average difference. But the self-monitoring thing is interesting, to the extent that we think we can measure which kinds of people are performative kinds of people, it's an interesting idea that I had not thought of before. I was not familiar with the self-monitoring scale before reading this paper, but it seems like an interesting way to study this question.

Wioletta Dziuda:

And the reason why she might not have found a difference in this private versus public treatment is because maybe the mechanism behind all this exaggerations that we engage in when we are answering surveys is a little bit different. So if I were to do some introspection, I would say if you asked me, "Would you be happy if your son married a Republican?" I just want to express my disapproval of a lot of what has been happening in the Republican party recently, I would say, "Absolutely not. I would be very unhappy. This would be unacceptable." But I know for a fact that in real life, if I met this sweet girl who actually happens to be Republican, my reaction will be completely different.

Wioletta Dziuda:

So I think the reason why we express our opinions in a stronger way in surveys than what they really are is because we just package a lot of emotions in those answer. We just want to get... It's an easy way for us to express certain frustrations, but we are more moderate than that in reality. So even when we were discussing the previous paper, I had this think in my mind that it would be really great if someone ran an experiment where you can actually observe people answering surveys and then acting upon the information that they were telling you about in the surveys and seeing whether actually their acts follow their expressive attitudes. And my guess is we would see much less polarization than what we see in the surveys.

William Howell:

I think this was a point that Greg made to us, that there is information in the responses that people offer to surveys. We should pay attention to what they actually say in surveys. There's something to learn. But the thing that you're learning may not be what their true deeply felt personal beliefs are. I mean, I'll say, I don't know if we should feel good about these findings or not, in that, at the end of the day, in politics, is the thing that really matters what you truly, truly deeply feel way down deep inside, or does the thing that matter that which you express and you actually come forward with? And to say, well, people are behaving in ways that publicly provide all kinds of animosity and hatreds but privately are moderate and reasonable, I don't know if that's a good news story or not.

William Howell:

It does though, productively point us towards what a solution might be. It says, well, if this is true, there's this performative quality to it. And you want to do something about to reduce affective polarization, then the intervention mean to be corrections of deep hatred and distress. It can simply be let's lower the temperature and let's see what kinds of interventions might reduce people's propensity to perform for one another.

William Howell:

Listen, that was a lot of fun. I learned a bunch from these papers. And it was fun to look back on some old shows that we've done. I hope we do this again.

Wioletta Dziuda:

Yeah. I think that's a good idea. And research moves forward and we are going to see what we can feature next.

William Howell:

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.