Is Polarization Pushing Us To Hate Each Other?

We’re constantly told that America is too divided. That we no long just oppose members of the opposite party, but actually hate them. That something is broken in American life.

We take these issues to one of the leading scholars in the world on polarization, Dr. Shanto Iyengar from Stanford University. And we focus on one of his papers, which argues that affective polarization really has gotten as bad as the experts say, and we discuss what we can do about it.

Paper: https://tinyurl.com/hrywxrl

Listen on Apple Podcasts or wherever you enjoy your podcasts.

Transcript

Anthony Fowler:

I am Anthony Fowler.

Wioletta Dziuda:

I'm Wioletta Dziuda.

William Howell:

I'm Will Howell.

Anthony Fowler:

And this is Not Another Politics Podcast. We are here today to talk about political polarization. We are told over and over again by journalists and by academics and by our friends sometimes that America is too divided.

Tape:

Increasingly we Americans occupy alternate universes.

Anthony Fowler:

That we are divided between red and blue states and red and blue voters and we hate each other and our discourse has broken down.

Tape:

There is very little common ground left, only baffling perceptions of reality.

Anthony Fowler:

And something is wrong with the American political system and maybe even is wrong with American life in general, because Democrats and Republicans just can't get along with each other and can't work together.

Tape:

Survey after survey has shown that Republicans and Democrats now view each other not simply as wrong, but as malevolent, literally a danger to the Republic.

William Howell:

This is something which has captured the attention of a whole lot of political scientists. So there's this emergent literature on affective polarization that's suggesting that polarization is not just about ideological or policy disagreement, it's rooted in social identity and emotion and it constitutes a major cleavage in the political and social lives of Americans.

Wioletta Dziuda:

So Will, you talked to someone who has fascinating research on this topic and tries to really answer the question, is this ideology or is this identity? Can you tell us about this a little bit?

William Howell:

Yeah. I had a really interesting conversation with Shanto Iyengar who is out at Stanford who has written a fair bit on this topic and he and his colleague Sean Westwood wrote an influential paper in the American Journal of Political Science that came out several years ago called "Fear and Loathing across Party Lines: New Evidence on Group Polarization." And let's hear what he has to say.

Shanto Iyengar:

The argument is that people, once they take on a group affiliation instinctively without even being aware of it, develop these patterns in group favoritism. You know, they root for their own side and they conversely begin to harbor a group animus. It's been documented on the basis of trivial group assignments, the color of t-shirt you wear, given the conflictual nature of partisanship, we would expect this ingroup/outgroup distinction to be especially pronounced.

William Howell:

It's just a thought experiment to think about what this relationship might be. Imagine that what I do is I scramble the policy commitments and ideological kind of locations of the two parties. Is the idea here that if I identify, if I have a social identity as a Republican, but what it means to be a Republican is kind of manipulated and altered, I nonetheless will stick by my commitment to the Republican social identity, of the Republican group, because I have this kind of affective commitment to it. Right?

Shanto Iyengar:

That's right.

William Howell:

Right, and so then the animus and bias, we're going to document, that you document to the extent that it is born of social identity. It's not that I hate the opposition party because I disagree with them. It's simply because they have a different social identity than mine.

Shanto Iyengar:

Yes, it's us against them. Simple as that.

William Howell:

So let's think about the evidence that you draw upon. A big strength of this paper is its effort to draw upon a whole host of experiments and you start with this implicit association test. Can you just tell us what is it and what is the problem that this approach has meant to solve?

Shanto Iyengar:

It's essentially meant to solve the problem encountered by survey researchers that when you ask people about their feelings, their sentiments, their beliefs about highly sensitive areas, people, they clam up and they will obviously deny harboring racial prejudice. And the logic here is that when people have a well developed belief or an attitude, prejudice, whatever you want to call it, they will be able to perform a task rapidly. So when you give people images of black faces and white faces and ask them to make very quick associations between the skin complexion of the face that you see on a screen and words that have positive or negative valence, the speed which with those associations are made will tell you something about-

William Howell:

The underlying beliefs.

Shanto Iyengar:

The sentiment. Exactly. And so it turns out that people are, they've more quickly associated black faces with negatively valence tones and white faces with positively valence terms implying that they have sort of a preference for whites.

William Howell:

So how did you do this vis-a-vis Democrats and Republicans and what did you find?

Shanto Iyengar:

Right, so we basically just took the exact same protocol of the IAT, but what we did was we gave people visual iconic sort of exemplars of the two parties. So the donkey and the elephant, the NRA, the Sierra Club. And then we had the same battery of positively, negatively valence terms and we developed the identical metrics so that we could actually do an apples-to-apples comparison of implicit racial bias with implicit partisan bias.

William Howell:

And what did you find?

Shanto Iyengar:

That was really quite shocking. We found that the effect size, in the case of race, the difference between whites and blacks on implicit racial bias was something like 0.2, whereas the difference between Democrats and Republicans are, the identical metric was 0.5. So it's more than double implicit partisan bias was just off the scale.

William Howell:

You then move on and you have a series of experiments that are designed to gauge people's sense of the relevance of partisanship for non-political decisions and you have this experiment wherein you ask people to select somebody for a scholarship.

Shanto Iyengar:

In this experiment, what we do is we present people with resumes and these are high school seniors who are eligible for a college scholarship. We manipulate both the grade point average, which is presumably the relevant indicator and we also manipulate extracurricular activities in the form of political affiliation. So the kid is either the president of the Young Democrats or Young Republicans. We find that the partisan cue trumps the intellectual ability cue.

William Howell:

Right? Okay. And then you move on to the these trust and dictator games.

Shanto Iyengar:

The way it works is you essentially, in the dictator version, which is a one shot game, you give someone, let's say, an initial endowment. Let's say you give this person $10 and you say, "Okay, now you're going to be playing with someone else and you're free to give this person any amount of this endowment we've just given you." Something close to 20% is donated. The trust game is more interesting because you have a reciprocity. Now you have the second player and the first player has given the instruction that whatever you donate to player two, I, the researcher here, I will multiply that amount by three and give that to player two and he has the opportunity to send some of that back to you. The observed result is there is this tendency to penalize, it's something like a 60 cent penalty maybe in the case of the trust game and maybe somewhat smaller in the case of the dictator game.

William Howell:

Right. So if I know that the person with whom I'm being partnered is of the same party, I'm more likely to give money back.

Shanto Iyengar:

Exactly. Exactly. Just knowing that this person is a Democrat or Republican makes me more willing to contribute to their welfare.

William Howell:

So let's reflect if we could for just a few minutes about kind of implications. To the extent that we're concerned about the levels of polarization, that it is making it increasingly difficult to come together as a country and get things done. What constitutes a check or a corrective?

Shanto Iyengar:

Well, if you're asking, is there a treatment that is, I think, a very difficult and the psychological nurture has a few clues. Things like, in the old days when the same question was being asked with respect to racial prejudice, one of the answers was yes, increase contact. Social contact would break down these crude racial stereotypes. That's a possibility. If people had more social interaction with people, they happened to meet their next door neighbor who happens to be a Republican and they get along just fine.

William Howell:

In the background behind that is a notion that by bringing people together they will see the folly or the idiosyncratic quality of their social identity and they will be able to kind of find common ground.

Shanto Iyengar:

Yes, yes. Or substituting ... a substitute. Yeah. Bring other forms of identity-

William Howell:

To bear.

Shanto Iyengar:

Exactly. And that way you then become suddenly the in-group in the composition of the ins and the outs and begins to switch. We need to introduce these crosscutting cleavages where people aren't always on the same team. Today, race, gender, social class and party are all in alignment. If there are some additional cleavages that can come into play, which break up this pattern, I think that would be one way of-

William Howell:

It's interesting though because it's – for a long time our discipline lamented the lack of distinction between the two parties and said, the problem that we have is that it's between Coke and Pepsi.

Shanto Iyengar:

That was irresponsible, exactly.

William Howell:

Right. And now we have, well we have real choices between the two parties at the elite level but as you point out, what is born of that is a lot of animus, a lot of anger.

Shanto Iyengar:

That's right.

William Howell:

Shanto, I want to thank you for taking, taking this time to talk to me.

Shanto Iyengar:

No, I enjoyed it. Thank you.

Anthony Fowler:

Let me ask you guys a question before we delve into the paper. You don't have to say what your political party affiliation is, but we are told by the literature that you have one and it's strong identity. It influences your behavior. Do you hate members of the other party?

William Howell:

Some of them.

Wioletta Dziuda:

Can I plead the fifth?

William Howell:

No, you can't because you don't have to say who you hate, but do you have hate in your heart?

Wioletta Dziuda:

As a Catholic, I do not like the word hate. I do not understand hate. As a famous person said, I think there is an element of dislike. Those are people who have very different positions than you and very different values and it's just hard to not to dislike them or like them less than some other people that share your values and share your point of view.

William Howell:

You have no hatred?

Anthony Fowler:

There are lots of people I dislike. I don't think I disproportionately dislike people from the other party.

William Howell:

So in this paper they make the case that a substantial number of people say, this is a real identity for me and I have all these various ways to advertise where I stand and I can figure out where you stand. And once I figured out where you stand, a whole lot of bias and discrimination, to you use their language, flows?

Wioletta Dziuda:

So Will and Anthony, you are empirical researchers. Can you talk to me a little bit about how should we think about this evidence. Is this good evidence?

Anthony Fowler:

The first thing they show are these implicit association tests, which, Will, do want to just tell us briefly how the IAT works and what it is and what it's supposed to be measuring.

William Howell:

The idea is that you can't ask people in a straightforward way about the phenomenon that interests you, but you can back out how they feel on the basis of how long it takes for them to answer a question. So things that they think go together naturally, like Democrats and good. A Democrat would respond to that question much more quickly or that association has something to say about that association. But if you're a Republican and you see those two things going together, there'll be a pause. And so they simply measure the amount of time it takes for somebody to respond to a question, to back out what kind of bias they have towards some particular group.

Anthony Fowler:

These measures are somewhat controversial among the people who study them. The actual reliability of this measure, you ask the same respondent to take the test on different days, they will have very, very different scores. There's also a question about whether or not this has any bearing on the real world. It might be the case that you are a Democrat yourself and it's easier for you to associate Democrats with good things, but does that really mean you're going to walk around the world discriminating against Republicans? I don't know.

Wioletta Dziuda:

So I have to question. If we assume that indeed we are measuring some sort of biases and let's even assume that this has some impact on our behavior outside of the lab. How should I interpret this after all, if I am a Democrat and I disagree with Republicans on policy dimension or ideology, shouldn't I have a bias then? What are we measuring kissing this test? Is it some sort of bias like racism where without any rationale, the reason I feel hostility towards the group, or is this just a measure of rational response to ideological disagreement? Do we know anything about that?

Anthony Fowler:

I don't. I don't think so. I mean, I agree with what you're getting at, which is of course if you are a Democrat, let's say you do in fact associate the democratic party and Democrats with good things and you probably associate Republicans with bad things. Not because you have some prejudice against Republicans, but you share the actual values and policy views of Democrats as opposed to Republicans.

Wioletta Dziuda:

And I want to emphasize that when they were doing this association test, it's not that they were showing pictures of people who were labeled Democrats or Republicans. They were showing logos of the parties or logos of NRA or Greenpeace. It's a natural response from a Democrats to have a negative reaction to National Rifle Association and it might be a rational response Republican to have a negative reaction towards Greenpeace.

William Howell:

Yes, yes. But the story that they want to tell is that what you're calling a natural response is not built in thoughtful reflection on the policy positions of these groups. They reflect notions of identity and bias and discrimination. It has this automatic quality that they are capturing, which is reflected in the time to response, which speaks to polarization being at its core aspect of an orientation. It's an emotional state. It's about, I hate the other team and full stop. I hate the other team because they're not my team.

Anthony Fowler:

Yet, the IAT is just measured with lots of error, which it probably is.

Wioletta Dziuda:

Or it could be that when you asked me on the survey, I'm angry and I like venting and we like having this sort of sports attitude. But when it comes to real interaction with a Republican or democratic friend, I'm actually more like a normal person who is just enjoying their personalities and not necessarily-

William Howell:

That's a little bit how I feel about this as well. I could imagine in a survey you say like, would you want to have dinner with someone from the other party and you're in the middle of the political survey you're like, "No, absolutely not. I hate those people." But then that night you go to dinner with members of the local party, you don't think anything of it.

Anthony Fowler:

I think about it. I think about it all the time. I disagree in this way.

Wioletta Dziuda:

Are you trying to make a prediction?

William Howell:

There's the willingness of one's ... your point, Anthony, is that, look, I'd be happy to go out with a Republican or a Democrat, either way it'd be fine. Like, and so these, what we're observing in these data don't carry over into the real world. But I can only imagine. I think it's decidedly true that for many people, if you're a Democrat, I don't want to spend any time with Republicans. I can't stand them. I really can't stand them because they read into all kinds of things about their values that they find deeply problematic and so your notion about like the dinner guests, Shanto has done some studies on picking marriage mates and you see that Democrats picked Democrats and Republicans pick Republicans.

Wioletta Dziuda:

Of course.

William Howell:

But I guess where I get off the bus is to say that's just because of social identity. It's there too, to the extent that you care that I'm a Democrat and that has real meaning for my values and my policy positions, then I'm not going to just sort of put those aside when I think about how I feel about other people or who I spend time with.

Wioletta Dziuda:

So I think one place, the word they are trying to get at this question is where they look at decisions of people on whom to award a fellowship. So they have, they're presented with two candidates and basically those candidates only differ in whether they were active with Democratic Party or Republican Party in the past. And what they show is that even if the candidates are the same and even in one candidate is better, if this candidate, the better candidate is from the out-group party, people award a fellowship to the person who represents the same party as they do. So do you think this goes a little bit further into direction of their interpretation that the evidence is suggestive that there seem to be some sort visceral reaction that's not justified by just policy preference?

Anthony Fowler:

I would agree with that. I think this is more interesting evidence than seeing just the IAT tests because this is closer to a real world behavior that we actually care about. You could easily imagine that, in fact, I was just in a room of people deciding which students would be getting a scholarship or a fellowship and in fact political ideology did come up as one of the things that people were talking about. And so you worry about this actually.

Wioletta Dziuda:

And you completely disregard the –

Anthony Fowler:

I certainly tried to, but I think there probably were other committee members in the room that were thinking about political ideology as one of the things that they were considering. So yes, this is much closer to a real world decision, but we wouldn't like it if it was the case that there was widespread discrimination against the other party when it came to these kinds of important decisions like which students get a scholarship.

William Howell:

I'm surprised that you're saying this because you're already lapsing into using the labels that they use, which is to call it discrimination. And one story is, is that I see that it's a Republican and I'm a Democrat and I hate him because I hate the other side and that's the story that they want to tell. But alternatively we might say as well, what it means to be a Democrat, in my view, considered judgment, is that I care about people who are disadvantaged in a deep way. And my willingness to extend benefits to somebody who doesn't share those views is not about discrimination per se.

Anthony Fowler:

I agree with that and I think that could be that indeed they are saying, well this person probably shares the values that I think are important and so I'm going to select that student.

I will also point out that in this study there is very little other information to go on and so the subjects are given a GPA and the subjects are told that they are at being asked to weigh in on these files because the committee is undecided.

What should you go on at that point? Maybe you should just flip a coin or maybe, sure, pick the person with your party because that's the one little piece of information you have about that person's values or views or something else.

Wioletta Dziuda:

So basically you're saying the stakes were so low that at the end of the day, flipping a coin would be a perfectly fine decision, so you might as well flip a biased coin.

Anthony Fowler:

It might be the case that you get a big effect here, but in an actual real world scholarship committee where these subjects had lots of more information about the students, they no longer care so much about party because once I also get to read your essay and once I get to interview you and once I look at all the courses you've taken and your letters of recommendation, then what little inference I'm drawing from your party is going to be mitigated by all of the other important things that I'm going to consider over and above that.

Wioletta Dziuda:

What about the experiments with dictatorship game and trust game?

William Howell:

Well, I think one thing, to their credit, what they find is that knowing the partisanship of the person that they're interacting with, either in this dictator game or the stress game, informs their willingness to give up money to that other person. And Democrats are more likely to give money to Democrats and they're less likely to give money or they give smaller amounts of money to Republicans. But that could be, because again, I hate Republicans because I'm a Democrat. Right? I hate him or it could be that no, Republicans are misguided and malicious and ill-informed because they've assumed all these policy positions and have these policy priorities with which I profoundly disagree.

Wioletta Dziuda:

So I must say that at this point of reading the paper, I became extremely optimistic about the situation in the United States and the political polarization. Because if you look at the numbers, yes, it is true that Republicans give a little bit more to Republicans and Democrats a little bit more to Democrats, but the numbers, the differences are extremely small. So in both games, the participants were to divide $10 and I think an average division was to give two to $3 or $4 to the other party. And the differences were somewhere between 50 cents to one dollar. So I would think if I really hate Republicans, I mean it's just so easy for me to give them zero and no one will be upset about this. So this sounded like-

William Howell:

We're okay. That's not their story, right? Their story is that, oh my God, what divides us is so much more than disagreement about the optimal policies that the state ought to be advancing. It's that these deeply rooted social identities, they're associated with parties, make it impossible for people to get along.

Anthony Fowler:

I suspect that some of what we're measuring in this paper is genuine animosity or at least strong views about the other party. Some of it I'm also concerned is just an artifact of taking a political survey.

Eitan Hersh, he did this fun exercise with FiveThirtyEight where he just fielded a survey that looks like the political surveys, but he asked, would you be upset if your son or daughter married someone who was a fan of the Yankees? Suppose you're a Red Sox fan or vice versa. And it turns out you get basically the same response as you get to the political version of that question. And it could be that people also have really strong biases about baseball, or it could be that in an artificial survey context, you just give the answer that is kind of the fun answer to give in that moment even though in reality most of us have not heard of marriages breaking up because of baseball fandom.

William Howell:

But I think one thing that they might take away from this comparison between the Red Sox and the Yankees is there too is more, it's a team sport, social identity and that you observe big effects there is consistent with the claim that partisanship at its core is about picking sides and which team you're on. And it's not about anything essential and true. And if you swapped out all the players for the Yankees and the Red Sox, you wouldn't have people switching team affiliations.

Wioletta Dziuda:

Or because they like-

Anthony Fowler:

And then I would want to know what is the evidence of that, why that, why isn't it the case-

William Howell:

For sure.

Anthony Fowler:

That people's policy views are driving their partisan choices and not the other way around. And in fact, I mean we could spend a long time talking about just that question, but almost all of the evidence we have seems to point in favor of people having genuine policy views that are driving their votes and driving their partisan preferences.

William Howell:

So where did we come out? Where do each of you come out on this paper? What do you take away from it?

Wioletta Dziuda:

I think it's not as bad as we thought it was. I think people have animosity towards each other, but not really a lot. They are still willing to give $2 to the other person, which I find extremely optimistic.

William Howell:

Yeah, you feel good about that.

Wioletta Dziuda:

Yes.

Anthony Fowler:

I don't think I take away a whole lot from this paper. I think it's probably true that Democrats and Republicans genuinely disagree with each other and that that does affect some aspects of our lives. As Wiola said, it's probably not as bad as you would think if you just read Vox all day or just listen to NPR all day long and heard all these accounts of hyper-partisan, hyper-polarized America. But I also suspect that to the extent that it looks like survey respondents are behaving in partisan ways, that's probably a little bit overblown. And if you look at people in the real world, they probably do not behave as partisan as the survey respondent in the political poll in that moment where it's kind of fun to say the partisan thing, but it may not actually influence any of your behaviors.

William Howell:

So I think that when we want to characterize distinctions between Democrats and Republicans to say that one is the party of small government and one, the other, is the party of an active government that presents a very anemic view of partisan divisions and our political life and this paper and the papers that are contributed to it, fill out our understanding that show that there is this affect of quality. I do think that there's actual anger across partisan lines.

William Howell:

I am not convinced that the anger is just born of partisan identity and that it is epi-phenomenal and it necessarily represents bias in the way, or discrimination, the way we're accustomed to talking about those things. I think it is actually rooted in something profound. In characterizing what partisanship looks like, I like this paper because it fills it out. It says is there's more in play than just policy disagreement. But in characterizing what's at the core of it, that's where I would part company.

Anthony Fowler:

Okay. I would push back a little bit on that there's more in play than just policy disagreement because I think if you want it to make that case, these are not the experiments you would do. I think you would want to see, for example, as I add in more policy information in those resumes, does the partisan thing go away or does it stay the same and so forth. So I think-

William Howell:

Yeah, but I think when they asked those feeling thermometer questions, like how do you feel about the other ones, I think that is tapping into, it's not only cognition, it's also this affect of quality. I there's an opportunity-

Anthony Fowler:

But why is it not just policy? Why couldn't I explain that result by if you asked me how do I feel about-

William Howell:

Oh, I could. At core I can't stand him because of the positions that he takes. But my point is, I can't stand him. But that's a thing. It isn't just, I disagree with him. It isn't just that I think he's misguided. It's that I can't stand him. And that's what's being expressed in my answer to that question and that that is a dynamic of our politics that we have to contend with. But absolutely, it could. You could explain the difference entirely, just on the basis of, he assumes a set of policy positions with which I disagree. And that might be overstating the other side, but to say as this paper does, it's really rooted at its core in just identity. I wouldn't go there.

Wioletta Dziuda:

So now is the time when we reveal our parties and affiliations and we see whether we'll come back for the next week's episode.

Wioletta Dziuda:

Thanks for listening to Not Another Politics Podcast.

William Howell:

Our show is a podcast from the Harris School of Public Policy and is produced by Matt Hodapp. Thank you for listening.