Ideology vs. Identity: What Drives Political Preferences?

>> Elizabeth Elder: Thank you all very much. My name is Elizabeth, and just to give a little bit more background on who I am, I got my bachelor's degree at Indiana University, go Hoosiers, and my PhD at California Berkeley, go Bears. And I'm working on bringing myself to say, go Cardinal, I'm not quite there yet, but it'll happen, I'm sure.

I've been here at Hoover for about a year now, and I'll be speaking today about a topic that is both close to my heart and close to the heart of my research, and that is, how do ordinary people make decisions about politics? So to give you an overview of what we'll be talking about today, I'll start by posing kind of a straw man folk theory of democracy.

We'll talk about challenges to that theory coming from public opinion research new and old. We'll talk about an important alternative model for how citizens engage with politics, in particular through social groups. And we'll close by talking about two particularly important social groups that structure a lot of American politics, race and gender.

And we'll have some time at the end for questions and answers. So if you did the readings for today, or maybe if you took a high school government class, you might be familiar with the folk theory of democracy. It's kind of a simple characterization of how citizens engage in the political process.

It's a simple model that looks something like this. Citizens form preferences about the kinds of public policies they would prefer. They then choose candidates to vote for that share those preferences, and then once the candidates are elected, they enact policies that citizens approve of because that's why they've chosen them to be in office.

So if I think about myself casting a ballot, the local elections here are often kind of about housing, so maybe I form an opinion about housing. I look through the list of candidates, I pick the one whose opinion is closest to mine on the issue of housing. And then once they are in office, if enough people agree with me, then they make housing based policies that I will be happy with over the course of their term.

So maybe that sounds like how you engage with politics or the people you know. But one of the most important things that you learn when you study politics professionally, as I do, is that almost no one cares about politics as much as you do, and therefore, almost no one thinks about politics in the same way as you do.

So when you think about the broader scope of the American electorate, we'll see that there are pretty serious challenges to the idea that citizens form meaningful preferences about public policy and that they choose candidates to vote for based on those preferences. And there are problems with the third thing, too, but that's for another day.

So the first challenge I'd like to talk about to this notion that the American public is broadly engaging in this kind of democracy is the fact that many people care very, very little about politics. About 30% of Americans did not vote in any of the 2018, 2020, or 2022 elections.

About a third of Americans and about half of young people think it does not really matter which party controls congress. Only about half of people talk about politics more than once or twice a week, and only about half of Americans can name all three branches of government. So there are really huge groups of the American public that are just not engaging with the political process at all.

And it should be clear, I do not say this to denigrate anybody. People have a lot of important things happening in their day to day lives that don't involve refreshing twitter 20 times a day like I do. But they really are just not engaging in this kind of preference formation and voting pattern that would be necessary for the folk theory of democracy to hold.

This is, in part, really problematic, because the people who do not engage in much politics look very different from people who do. The people who are the most active and engaged in politics tend to be more extreme than the people who are not. Here, for example, is some data from Pew.

Pew has a method of sorting Americans into a number of buckets, ideologically, ranging from the most extremely liberal to the most extremely conservative. And what we're seeing here is their rates of engaging in political activities, following what's going on in politics and government, and discussing politics and government almost every day.

And for both of these, you see kind of a u shaped pattern. The people with the most extreme political beliefs are the ones that are following and talking about politics the most. This is probably apparent to you if you spend any time on Twitter. This is some data on the attitudes of people who tweet a lot about politics and then everyone else who's on the platform.

So thinking about these people who are tweeting a lot about politics, they follow the news. They tend to put themselves in echo chambers, listening mostly to people who have similar beliefs to them. They're more ideological, they have more negative feelings about the other party. And for people who do not spend a lot of time tweeting about politics, these things just are not nearly as true.

So to the extent that this is the political conversation, the people that are part of it are very, very different from the people who are not, even though they all in some sense have equal standing in the electorate. So setting aside the people who are not engaging in the political process at all, there are still challenges to the notion that the people who are engaging are forming meaningful policy preferences.

And in fact, in general, stable and influential issue positions on matters of public policy are fairly rare. So if you think about the notion of someone forming an opinion about something, casting a vote based on it, and then approving or disapproving of public opinion of public officials actions based on those opinions, you may want their attitudes to look something like this.

So I mentioned the issue of housing. You might think I should take in information. I should form a single, coherent attitude. I am pro housing, I want more housing. And then when I'm casting a vote or evaluating public officials actions, I'll take my one coherent attitude into account.

But we have a lot of evidence that people's attitudes on matters of public policy look a lot more like this. It's not one single coherent attitude, but a bunch of different considerations, different notions about this issue, some of which might make me feel more pro housing and some of which might make me feel more anti.

And then when I go to cast a ballot, when I go to answer a question on a poll, when I have a conversation about this, I'm probably not going to remember all 20 or 50 or 100 of the things that I think about housing at once. I'll just pick one, whichever one comes to my mind, or maybe a couple.

And if that one that I pick happens to be something that's anti housing, maybe I drove by a new apartment building in my neighborhood. I don't like how it looks, I'm worried about my neighborhood changing. Then even if I have 35 other pro housing considerations in my mind, the one that happens to be at the top of my thoughts that day is gonna be influential over my voting and my opinions that I express.

So this kind of model helps explain some really common patterns that we see in public opinion. One of these is that if you ask someone how they feel about an issue today, and then you ask them again a year from now, they'll often give you a completely unrelated response both times.

People's attitudes change very readily, and that's consistent with the model in which people are kind of taking whatever ideas are top of their mind that day, rather than having a single coherent thought. Another thing that's probably related to this is the idea that question wording matters a lot in polling.

So there's been a lot of talk in the news about abortion in the last year or two. You probably noticed that polls on this can have really, really different results. And part of the reason we think for this is that different question wordings can bring different ideas to the top of people's minds.

Even if no one's getting any new information or changing their mind in a meaningful way, they can bring different bubbles to the top of what people are thinking about when they're giving you an answer. So both of these patterns, along with other evidence, suggest that people's attitudes look a lot more like this, and therefore, it's kind of hard to think about them having a single attitude shaping their vote choice and their approval of elected officials.

So you might think maybe people don't have strong attitudes on any individual issue, but maybe they have broader ideas, broader ideologies or principles that are shaping how they engage with politics. And that also turns out generally not to be the case. If you ask people to talk to you about what being liberal or being conservative means in politics, they'll often be able to tell you what it means to dress conservatively or something like that, but not whether estate tax reform is a conservative or a liberal policy.

And I think this makes sense even if you know some broader conservative or liberal principles. For example, if I think conservative means limited government, it takes a lot of steps and a lot of information to get from that general notion to a particular assessment of whether this estate tax reform proposal is conservative or is liberal.

And as we've seen, a lot of people are just not engaging enough in politics to go through that kind of arduous process of reasoning and gathering information. So most people have a pretty wide mix of liberal and conservative ideas and liberal and conservative policy positions. This was first pointed out in a really famous 1964 paper by a political scientist from Michigan named Philip Converse.

And if you kind of read the responses that he got when he asked people what liberal and conservative mean, you would hear very similar things today, even though the political system has changed a lot. People don't often have that great of an idea of what these things mean.

So people don't have strong issue positions. People don't have coherent ideologies. How do they make sense of politics? How are people going about their days casting their votes and talking about political issues? There are a lot of great answers to this question, and I'm happy to talk about other thoughts about this in the Q and A, but I'll focus here on one alternative model that is really, I think, motivating a lot of political behavior research right now, and that is groups.

The idea of politics through the lens of groups. Human beings are social animals. We are very, very good at dividing people into categories and assigning stereotypes to those categories, ranking them against each other. If you've ever taught a gym class of elementary schoolers and made there be a red group and a blue group, it's shocking how quickly they will decide, like, this is the good group or the bad group.

This is the smart group or the dumb group. People are really great at assigning importance to group labels, and this is no less true of politics than it is of life in general. So what do I mean by groups here? I mean people who share some sort of common identity marker, either because they have adopted it, chosen to identify with it, or because it is ascribed to them by people in their life.

And these groups have both definitional attributes that define their membership and also a set of norms and expectations, values, ideas about how members of the group should behave, even if those ideas are contested. So, for example, I am a Californian, I've lived here for a long time. And so you could call me a Californian, and I couldn't tell you that's not true, even if I don't feel particularly connected to that identity.

But I also might adopt that identity, I might think of myself as a Californian, and that might shape how I think about things. So in that sense, it could be both an ascribed and an adopted identity. And if you think about the question what are the stereotypes of Californians, what are Californians like?

What would it mean to vote like a Californian? Ideas probably come to your mind. So you might think, stereotypically, Californians cannot tolerate the kind of extremely humid weather for the Bay area that we're having right now, it's intolerable. You might think that Californians tend to care a lot about the environment or that they tend to vote liberal in elections.

These are obviously contested. Obviously, not all Californians are ascribed to these things. But if I were to ask you what the identity means, these are some of the things that might come to mind. There are a lot of different groups that shape people's politics in a lot of different ways.

Of course, partisan groups are the most important of these, but they're also ideological groups, racial groups, religious groups, groups based on geography, on occupation, on ruralness and urbanicity. And all of these things are important for how people think about politics and political issues. So what can groups do?

What do I mean by shaped political behavior? The set of things that groups can do for people's politics kind of fall into two buckets, which I like to think of as the heart and the brain. So thinking about the heart, groups can give people emotional and affective stakes in political contests.

So if you think about a person and their elected official, like a member of Congress, for example, they probably have never talked to one another. They probably don't share a lot of the same friends or a lot of the same hobbies. But maybe they share a group identity.

Maybe they're both Democrats or they're both Republicans, they're both union members or they're both Californians. That kind of sense of shared identity can help people feel more common ground with political representatives or members of their constituency that are not necessarily very socially proximate to one another. Groups can also give emotional stakes to politics.

For example, if I think of myself as a Democrat in the sense that I check the d box in most of the ballots that I vote in, but I don't really think of myself as identifying as a Democrat. Maybe the results of the election don't matter so much to me, other than the particular races that govern my life.

But if I identify as a democrat, if I really see the party's wins as my wins and their failures as my failures, then emotional stakes all of a sudden become involved. I might be proud or excited when the party wins, or sad or angry when the party loses.

And these group ties are really important in creating these emotional stakes. The other really important thing here is that when placed in conflict with one another, especially zero sum conflict, like an election. Group ties can make people more willing to expend resources, their time and their energy to make their group do better and the other group do worse.

And absent these kind of group attachments, people are less willing to engage in these costly behaviors on behalf of their team. So the other set of things groups can do have to do with the brain. And to my view, these are even more important. Groups can really, really help people understand political conflict and make educated voting decisions.

One of the most important things groups can do is they provide cues that help people infer what politicians believe and care about. So if I have, you know, 20 candidates on my ballot, I might go to the website of each of those candidates and look at all of their different policy positions.

And that would take a very long time, and it's not really necessarily a good use of my resources. Because if I have in front of me a race for one candidate is a Democrat and the other is a Republican, I can infer with 99% accuracy their opinions on almost all the issues they're gonna legislate on in a national political contest.

So it's a huge time saving heuristic to be able to use party identification or party membership to infer a whole host of political beliefs. And of course, this is true of party for partly institutional reasons. But you might also think if a candidate says they're a feminist or a candidate says they're a union member or they're in the NRA or something like that, you can infer, without doing any more research, what that candidate thinks about a number of issues that might be important to you.

So these group labels can help you make judgments about what candidates will do and care about and how they'll vote at a really minimal informational cost. On the other side of things, groups can help people make sense of issues. So I referenced the idea of estate tax reform earlier.

If there were an estate tax reform proposal in Congress, and someone wanted to learn whether to support or oppose that policy, they could find the bill text, or they could find a summary that someone wrote of the build text. And they could figure out exactly what it was gonna mean for people in their level of wealth, by that, again, is a lot of effort that a lot of people are not willing to exert.

But if they read a news article that says the Democrats are on this side and the Republicans are on this side, wealthy people are on this side, and poor people are on this side, most people are not going to need more information than that to understand how that proposal is going to affect their life.

So again, this is a real time saving, informationally useful piece of strategy that people can use to make sense of politics. Just to underscore this, I wanna give what I think is a really useful example. Please don't try to read all of this, I'm gonna tell you what it says.

So I mentioned before that when people answer the same survey question a year or two apart, they often give really different answers. And part of the reason is that when you ask people about some general policy like housing or the estate tax, they don't generally have clear ideas about what that means for their life.

For the groups that they care about, like we saw before, it's generally this mix of different considerations. But once people know where the parties stand on these issues, they have a much clearer sense of people like me support this and people not like me support this. And so if you look at the relationship between people's attitudes in one year and a year or two later, and you divide that up based on whether the people know where the parties stand or whether they don't.

The people who know where the parties stand on the issue have a much stronger relationship between their attitudes at different times than the people who don't. And in fact, in some cases, the people who don't know where the parties stand, it's just as good as a coin flip whether they give the same answer in two years to the same question.

So parties are really helping people understand what's going on in a way that helps them form these more stable, more consistent attitudes that are gonna be useful in making vote choices and holding elected officials accountable. So, as you might be able to tell, I'm a little bit of a pro groups person.

I think groups are really useful in helping people understand politics. And I think part of the reason it's important to me to say that is that groups can kind of get a bad rap. If you think about teamsmanship or tribalism, these are not generally good things. People do not think of group based reasoning as normatively good, and I think that's a reasonable conclusion.

If people are blindly following a party that they've chosen for no particular reason, that was just assigned to them randomly, and therefore has no relationship with the things that they believe or care about. And sometimes it can feel like that's the way it works. People often inherit the partisanship of their parents.

Here's a plot from Hugh showing that among Republican parents, 81% of their teenage children are Republicans. So maybe if you just got your partisanship straight from your parents and are following all of its policy prescriptions, it's a little bit silly to just follow your party's beliefs. But what research shows is that as people enter their 20s and 30s, they tend to break out of these kind of inherited partisanship and inherited religious affiliations.

And they tend to choose groups to align with that more closely match their independent values and how they see the world. So a great example of this is partisanship. Partisanship does not change very often for people in adulthood, but for people in their twenties, it changes actually fairly frequently.

And even in adulthood, important enough things can make people change their partisanship. For example, in the 1970s, abortion was not yet really an issue that was on the national partisan stage. There were a lot of pro choice Republicans and pro life Democrats. And then in the late 70s and early 80s, the parties chose sides.

They made it very clear, Republicans in the pro life party, Democrats of the pro choice party. And based on that, a substantial number of people, especially women, changed their partisanship. They switched to a party that shared their abortion opinion. And so, in doing so, maybe they found themselves in a party that more aligned with their values on this issue and maybe other issues.

And so when they went to make a vote choice in the next election or decide how to feel about the next issue, they were following a new party that better aligned with their values. Some things are not quite as easy to leave as a party you might think of a religion or a racial group.

But even if people can't just completely switch their affiliations, they can down weight identities that they don't feel like match the way they see themselves. For example, as I said, I'm kind of a californian, whether I like it or not. But if I felt like I didn't fit in, if it wasn't aligning with how I thought of myself, I might just not let that affect my life very much.

I might not let it give me emotional stakes or voting information if I didn't feel like it was a good way of thinking about people who share my ideas about the world. So in the end, people often end up with a constellation of group memberships and group ties that more or less do reflect something like their values.

And voting based on those is going to generally give them pretty good outcomes. But of course, thinking about politics through groups has huge downsides. And one of the maybe most salient these days is polarization. I'm sure you've all seen a plot like this before, we have a little pink curve of Republicans and a little blue curve of Democrats.

They're moving further apart from each other over time as Republicans and Democrats become more likely to take the opposite side from one another in various policy issues. And this is not just a policy divide, people often these days feel much more negatively towards people in the other party than they did before.

This is a plot of people who are saying they would be upset if their child married someone from the opposite party. These days, maybe a third or a half of people say they would be upset if this happened. And in the 1960s, these numbers were close to zero.

So there's been a really dramatic change. And a lot of literature has been spent trying to figure out exactly why. What's going on here. There are a lot of explanations, especially great ones at the elite level, that are not really under my purview today. But one that is, as more groups have come to sort based on party, as more of your identities are aligned with either the democratic or the republican party evenly, people's animosities towards one another have more and more reasons, more and more justifications.

So just to explain a little bit, in the 1970s, your ideology was not particularly predictive of your party, especially at the mass level. There were a lot of conservative Democrats and a lot of liberal Republicans. These days, that's just not the case. People who live in urban and rural areas are increasingly likely to belong to different political parties from one another.

And people with different beliefs about race and racial progress are increasingly likely to belong to different political parties from one another. So if you imagine someone from 1960 answering this question, would you be mad if your child married someone on the opposite party? Maybe not, that person might still share an ideology or share a geographic membership with you, still feel the same way as you do about race.

These days that's much less likely today. If you're imagining your child marrying someone of the opposite party, they're probably also on the opposite side of you on all of these other things. And that's just a much more kind of emotionally resonant proposition. And we think one of the main reasons why people find the divide these days so much harder to cross, at least as an interpersonal matter.