Partisanship Over Policy?

Shortly after joining the Dartmouth faculty in 2018, professor Mia Costa began to study how voters react to negative partisan rhetoric among their political representatives.

Two years later, the overwhelming response to her published research gave Costa a sense that she was onto something—and spurred her toward a deeper examination of the larger questions about polarizing rhetoric that culminated in March with the release of her first book, How Politicians Polarize: Political Representation in an Age of Negative Partisanship.

The assistant professor in the Department of Government analyzed more than 130,000 newsletters and 1.3 million tweets from members of Congress to understand the prevalence of negative partisan language, and directly surveyed hundreds of political elites to get a sense of why they say what they do. 

In a Q&A, Costa shares insights into what she learned about politicians' behavior, how voters feel about it, and what the growth of affective polarization and negative representation means for democracy in America.

What inspired you to pursue this particular book project?

It really started with an article I wrote in the American Journal of Political Science called Ideology, Not Affect. It's about affective polarization, which is this popular idea in American politics that Democrats and Republicans not only disagree with one another, but they actually really dislike one another. It's a form of polarization based on emotional affect rather than ideology.

The motivation for the article was learning what people think about representatives.  There was a lot of research on the mass-level extent of affective polarization among Democrats and Republicans. Since Democrats and Republicans dislike each other in the public, the logical conclusion that some people were making is that it must mean people respond well when their representatives then talk about how much they dislike the other side. But I had a suspicion that people actually don't want their representatives to make partisan attacks; they want their representatives to be good representatives. And that's exactly what I found.

You mention in the introduction that your book fills gaps in the study of political representation and affective polarization. Was that intentional from the onset or did it evolve more organically?

It's probably both. The way I started this was from the side of voters: How do voters react to this specific type of polarizing rhetoric by politicians? That was an obvious gap to me in both the representation literature and the affective polarization literature. We know that affective polarization is such a powerful force in the mass public, but what does that really mean for how people perceive their representatives' rhetoric and behavior? Once I started working on it more seriously, collecting way more data on what the rhetoric really looks like, and why political elites—candidates or elected officials—use this rhetoric, then it became even more obvious that we don't have a good understanding of it. 

My starting motivation was that there were already a lot of assumptions made in the affective polarization literature that because Democrats and Republicans dislike each other, that must drive representatives' rhetoric and behavior. But it wasn't really empirically demonstrated and that's what started me down that path. And as I started collecting more data on this more general topic, the other ways in which it filled out the story became more obvious.

That focus on "negative representation" in modern politics is at the heart of the book. Why do U.S. politicians rely so often on these negative partisan attacks? And how are they rewarded for this antagonistic behavior?

I outline three possible reasons in the book.

One is that they do it because it's what voters want. I don't find evidence for this, and I think it's the one that people think is the most obvious. If you were to ask people, "Why do politicians use all this negative language?" I think most people think it's because we're so polarized, and that's what their voters want. I argue that supply and demand doesn't line up, because I don't find evidence that voters like this or, when I survey the elite side, that they think voters want it. Politicians even report that voters don't like this, so then why do they do it?

That leaves two other reasons. One is that politicians are people, too. This is a little bit of a counterintuitive thing to say as a scholar of politician behavior, because if you're running for office or you've been in office for a while, you behave differently and have different incentives than a regular voter. But on the other hand, affective polarization runs deep, and if you're running for office or holding office, you're also a strong idealogue who is embedded in a really polarized environment. In my surveys I find that, at least at the local and state level, people who run for office or hold political office are likely to report very high levels of partisan animosity themselves. The fact that they're willing to report that in an anonymous survey suggests that they do genuinely hold these negative feelings and there might not be a rhyme or reason behind it besides that.

The other reason is that there are other external incentives and rewards that primarily come in the form of attention and power. Negativity specifically surrounding partisanship gets a lot more attention online. When members of Congress post on social media about the other side instead of their own side, and they use more negative language, they get more shares and likes. Rhetoric is likely to go farther when it's focused around attacking the other side. And I also show that it is more likely to get picked up on partisan news media and get amplified. That translates to more national prominence and attention, which I also measure through campaign donations. A lot of out-of-state donors happen to give more money to politicians who use more negative partisan language. The chicken or the egg, what comes first: Are donors giving to these politicians because they like that behavior, or are the politicians driven by trying to get those donations? That's not something I can empirically demonstrate, but I think it's part of the story.

Everyday voters who go to the polls don't really like hearing this language because it's not what they want out of politics. People are really turned off by partisan fighting, so it's not coming from voters. I think it's coming from the external reward system of a polarized environment where that behavior just gets more attention and it builds leverage for elites in power because they get more attention and they're more likely to be in leadership. It's part of the cycle.

The points you just made about the attention and notoriety at stake, and that a lot of the most negative representatives are the ones in highly uncompetitive seats, are really interesting. Do you see one or the other having a greater effect on this phenomenon of negative representation? Do the safe seats or the attention incentive provide more fuel for this behavior?

I don't know what is driving what—it's all related. It's hard to know what comes first, too. I don't think using negative partisan language is going to put you in a safe seat, lead to a more uncompetitive election, or cause you to raise more money, necessarily. I think they reinforce each other: The people who are in more safe seats have that potential leeway. I find that voters don't respond well to this, butt if you're an incumbent who's safe in your district, you're not going to turn off voters if you use this language and you also get more attention. 

It's reinforcing that the people who've been in office the longest and who are in safer districts and who happen to hold leadership positions are those who are more likely to get more money from out of state because they have more national notoriety. But then these are also the people who are the biggest partisan warriors. I think more freshman Congress members have to focus on the issues and talk to a broader set of voters, and have a greater  incentive to be more productive in other ways. They're not going to waste their time making partisan attacks; they're going to be a bit more measured and talk about the issues. 

What findings surprised you the most during your research?

It's not surprising to me now, because it totally makes sense with everything else we know about representation, but I think one thing that I was expecting was that I would find a higher frequency and prevalence of negative partisan rhetoric compared to policy language. But politicians, even on social media, do talk about substantive issues more than they use negative language. That part is surprising. 

Given what we hear about through the media and people who follow or study American politics, I was expecting to see higher rates of negative representation. But that negative partisan language is actually not as prevalent as we all think it is—and that's because it's what is covered, it's what we talk about, it's what we hear about. Negative partisan language goes farther, which is why we think it's all politicians are talking about.

For example, one critique Hillary Clinton got when she ran against Trump was that she wasn't really focused on the issues; a lot of Democratic strategists thought she was too focused on Trump when people wanted to hear about the issues. Part of me thought that's what I was going to find, that Democratic and Republican lawmakers are just focused on the other side and they're not actually talking about the issues—but that's not what I find. The good news is that there's actually a ton of policy language.

How have you incorporated this research into your teaching?

I teach a class called Political Representation that's an upper-level seminar and next year I'm going to teach a new class called Does Congress Represent America? In those classes some sections focus on how the language that Congress members or politicians at any level use matters for representation—and part of what I argue in the book is that it does. 

To be clear, it matters what voters think about representation because it's about the relationship we have with our governments, and that matters for democracy, but it also matters insofar as it predicts or relates to other forms of representational behavior. For example, in the book I show that there is a negative relationship between how productive and effective a Congress member is at actual lawmaking and the type of language they use. The more they focus on the other side, the less productive they are in the chamber; the more they talk about policy, the more productive they are. 

My classes are focused on representation and the quality of representation, and how to evaluate our governments in terms of how well they represent Americans. Studying questions in this way, where I focus on rhetoric and language, helps connect the symbolic behaviors to the substantive forms of representation that are based in lawmaking and power.

Is focusing on people rather than systems, which feel so big and hard to change, where we can find some hope for a future with less polarization? Is there hope?

There always has to be room for hope. I think part of the reason this happens is because of the larger incentives that a polarized environment creates. And so that's a harder nut to crack, of course. We can't change cognitive biases toward negativity. That's just part of human cognition. But we can probably change the way in which the media information environment rewards that negativity. I'm not the type of scholar that knows how to change what types of rules and regulations and incentives exist on the media side that could alter that behavior, but that's probably more changeable. There are also institutional procedures and rules that could potentially lessen polarized lawmaking. 

I do think there's hope in the finding that policy language and substantive issues are really still what comprises a lot of politicians' communication—it's really just about what we focus on and what is rewarded. The good news is that people generally don't like partisan attacks. The most partisan among us still don't like a focus on partisanship over policy, but we tolerate it because we think that's what the other side is doing. That, to me, is a silver lining: Negative language is not the most prominent. It's not what people want, it's just the most sticky—and that's an easier problem.