Polarization, journalism and the ‘pictures in our heads’: A Q&A with Yanna Krupnikov

Polarization challenges journalists’ ability to do their jobs. With divergent narratives on the political left and right, it can feel almost immobilizing to figure out ways to convey facts to people who seem to live in entirely different realities. Navigating how to build trust with those communities may feel demoralizing, and especially so if prominent figures eschew facts, manipulate people or unfairly challenge journalists or the motives of good journalism. In an environment like this, it’s easy to imagine why journalists might portray the divide as simply left versus right. 

Yanna Krupnikov, Stony Brook University

However, thinking of polarization too simply can obscure a different problem. So thinks Yanna Krupnikov. In a new book, “The Other Divide: Polarization and Disengagement in American Politics,” Krupnikov and her co-author John Barry Ryan, both scholars of political science at Stony Brook University, describe a too-often ignored part of the picture of U.S. polarization: the fissure between “people who make politics a central part of their lives and those who do not.” The thesis of the book, as they put it: “The growing partisan divide in America can only be understood in the context of the growing gulf between people who spend their day following politics and those who do not (i.e. the other divide).”

The pictures people have of division in American politics — which journalists help paint — are often oversimplified, Krupnikov argues. To her, partisanship tells you only part of the story about polarization. But what adds depth is something she and Ryan term “involvement.” Some people orient themselves toward politics — they think about it a lot, they follow every detail and they read, talk and post about it constantly. These people are “deeply involved.” But most people are not this way — while some are totally inattentive, most Americans pay attention in some way, just not constantly. 

Journalists who do “person on the street” interviews may intuit some of this other divide as they seek out people engaged enough to have quotable views. But through surveys, experiments and knowledge of academic literature, Krupnikov in the book and this video shows just how significantly those who are highly engaged differ from most people — and also how they may have an outsized influence because of the attention they get in the media. For example, most people dislike political elites of both parties, but don’t direct anger at ordinary voters. The “deeply involved,” however, do genuinely dislike rank-and-file members of the other party. And it manifests in other differences as well — in policy preferences, topics they prioritize, even views toward family life. 

This nuance matters for correctly capturing the picture of polarization, and devising interventions to contend with it. And because Krupnikov directly addresses journalism’s entanglement with this issue in her book, I wanted to talk to her about how it’s important for journalists to understand and reflect this other divide in order to do more representative reporting — and also reach more audiences.

I came to Krupnikov’s work through a colleague’s recommendation of a fall 2020 New York Times essay: “The Real Divide in America Is Between Political Junkies and Everyone Else.” It was right before the national election, and stories about polarization and divides were abundant. The implications for news coverage were clear. Now, the midterms are roughly seven months away, and newsrooms, if they prioritize it, still have time to think about these issues as they plan their coverage. But Krupnikov’s expertise in political psychology and expression — and her thinking about journalists’ decisions on framing stories and who gets quoted — is worth following year-round.

What is the “other divide”? How should it affect journalists’ coverage of politics, and polarization — and should they even use those words? What does it mean for who should get quoted? Should journalists more consciously take into account how their work shapes what journalist Walter Lippmann called the “pictures in our heads?” And how might it affect how journalists build trust — and grow audiences — beyond those political junkies? 

Our conversation is presented below, edited for clarity and length. 

Loker: Journalists are acutely aware that there is a political divide in this country. But let’s start with the basics. Can you describe the other divide you have been focusing on?

Krupnikov: So obviously, the divide everyone talks about is the partisan divide, which is the divide between Democrats and Republicans. And this divide is kind of all-consuming in American politics. And it’s been all-consuming, I think, in news coverage of politics as well. 

This other divide that we’re talking about should actually not be surprising. If we look at research, this is something that people who study political communication and political science have generally been worried about since the 1940s — which is the idea that some people are paying a lot of attention to politics, and some people are paying no attention to politics. Now, most of the time, if we go back to the earliest work, the worry is really about the people who are paying absolutely no attention. It’s people who can’t name who their congressperson is, people who can’t even name who’s running for office. 

What we suggest in the book is sort of a reformulation of that divide. There are certainly people who aren’t paying any attention who couldn’t tell you who’s running. But we suggest that a less explored divide — a divide that has now emerged even more clearly because of the media environment that we’re in — is this divide between people who are paying just a super-tremendous amount of attention, and basically everyone else. 

“The Other Divide” illustrates that “Person 3” is most dissimilar from “Person 1” and “Person 2.” Screenshot from video.

The people paying a tremendous amount of attention to politics are who we call the “deeply involved.” They are people who are checking all sorts of news throughout the day. They’re probably on Twitter looking at news, they’re constantly talking about politics. And they are unlike basically everyone else. This “everyone else” certainly involves people who aren’t following politics at all, but “everyone else” also involves people who are following politics. Maybe, you know, once a day, they’re following their local news and they actually know big events that are happening. They’re just not engaged with it on an hourly basis. They’re not constantly worried they’re missing something. And so this “other divide” is the difference between those deeply involved people — who make up about 20% in the various surveys we have in the book and who are just deeply, deeply engaged — and essentially everyone else.

Loker: The implications of that divide for journalism jumped out at me. For instance, you show in the book that journalists overestimate the level of what’s called affective polarization of the American public, or how much Americans of one political stripe dislike the other simply because of their affiliation. You asked a sample of journalists to guess the levels, and their guesses are actually closer to the polarization among the deeply involved — not the American public at large. Can you describe that more?

Krupnikov: We’re thinking here about this idea that has actually gotten a ton of news attention — that people dislike each other simply for identifying as a Democrat or as a Republican. And this, I think, has been taken as a really serious sign of a problem in American politics, because it’s one thing if I don’t like somebody from the opposing party because they propose a totally different set of policy issues than I would, but it’s a whole other thing where my dislike of somebody of the other party has nothing to do with political issues. “I just don’t like them because they are of this opposing party.” 

One of the questions that’s often used to measure this type of polarization asks people whether they would be happy or upset if their child married somebody of the opposing party and happy or upset if their child married somebody of their own party. Being unhappy with an out-party in-law and happy with an in-party in-law signals polarization. A particular measure that I’ve used with John Barry Ryan, who’s a co-author in the book, and my other co-author Samara Klar, changes the question slightly to clarify that this hypothetical in-law will never discuss politics. We find that once you assure people that no one’s ever going to talk to them about politics, we see much less polarization. We see this pattern across a number of different survey samples. 

Because we now have these measures from a number of national surveys, we asked a sample of journalists, “can you guess what proportion of people would be upset if their child married somebody of the opposing party who never talked about politics and happy if their child married someone of their own party who never talked about politics?” What we find is that journalists overestimate the size of this group – they believe that more than 50% of people are polarized. The actual number is closer to 20%.

But while the journalists are overestimating polarization in the general public, they are coming pretty close to correctly estimating the level of polarization among one particular group of people, and that is the deeply involved. 

Data Source: Figure 8.2. in Krupnikov and Ryan's "The Other Divide." The first bar reflects the average level of polarization as guessed by a sample of working journalists. The measure used to obtain the level of polarization is "How would you feel if you had a child who married someone who votes for the [Republican/Democratic Party], but who rarely talks about politics?" The level of polarization is difference in happiness level between having a child who married someone of your own party versus someone of the other party. In making their guess, the journalists were given the wording of the question and what it would mean by this question to be polarized. The next bars reflect the average level of polarization within the following samples: Sample recruited by the survey company GfK in 2016; survey company SSI (now Dynata) in 2018; sample from the Cooperative Congressional Election Study in 2018 and sample from Dynata in 2019.

American Press Institute

Since the book, we ran a similar study. This time, we asked a different sample of journalists to guess what proportion of the public is frequently posting to social media about politics and frequently reading political news. And we match that again to a national 2020 survey (this time, the Cooperative Congressional Elections Study). We see that the overestimation for reading about politics is not as large as it is for polarization, though the journalists still overestimate it. But journalists are more likely to overestimate the percentage of people posting about politics. So again, putting it together with the polarization result, journalists seem to be ascribing the behaviors of people who are deeply involved — polarized, posting about politics and following political news — to a larger swath of the American public.

Loker: If journalists often miss this mark when it comes to the people they’re imagining, what could they be mindful of in reporting stories?

Krupnikov: I think any time a particular perspective or a particular kind of idea becomes really dominant, there is a tendency to see and fit everything into that lens. There is this tendency now to channel everything through partisanship. And there are good reasons for this. There are many partisan divisions. We see these divisions even with things that aren’t necessarily strictly political. We saw this when partisanship affected people’s responses to the pandemic, for example. And so I totally appreciate this focus on partisans. I study political communications, so I’m all in on partisanship affecting how we communicate and how we respond to messages. 

On the other hand, I think once we have this perspective, there’s a tendency to try to look for the partisan difference in a lot of situations and a lot of contexts. Another tendency is to focus on partisan gaps rather than overall rates. So what is more important, the gap between Democrats and Republicans on a certain behavior, or the fact that majorities of both parties do or believe something? And I think that’s an open question, but there’s often a tendency to focus on the gap rather than the overall behaviors.

There are certainly a lot of examples where partisanship is really, really important. But on the other hand, I think the focus on partisan divides is going to come at the expense of considering other differences within the public that in some cases may explain patterns to a greater extent.

Loker: In the sense that the pattern could look like partisanship, but be something else?

Krupnikov: Yeah, partisanship is correlated with a lot of stuff. And so the question then becomes, is it partisanship that’s driving these ideas or is it something else that’s driving them? Another issue with this focus on partisanship is that social science is inherently really messy. Asking people questions is inherently really, really messy. Even if we get the measurement right, and we get all the analyses right, there may be a disagreement about the implications, or a disagreement about the significance of the pattern for what’s going to happen in the future. But once there’s this lens of seeing something as a partisan difference as being important, I think there can also be a tendency to perceive findings that speak to that difference as being more important. 

Loker: So things that are often talked about in terms of, like, left versus right — that may not always be the best way to talk about it?

Krupnikov: Obviously, there are going to be a lot of things where there’s a very clear partisan divide, which I think should be covered. But I do think that there are certain times where you get into these types of behaviors that actually very few people are doing, but the people who are doing them are very partisan and very polarized — and so then the question becomes, is this an important thing to cover as if it is yet another example of these partisan divides? 

I can make it more specific. One story that I’ve often seen is this idea that people are ending relationships because of politics. The statistics on how many people are actually doing this are really murky. Some surveys put it at 15%, others get it higher as we expand it to different types of behaviors. So is this something that should be represented as a broader pattern? I don’t know. But I think it is a type of story that needs to be contextualized with survey data. 

Loker: And so you brought up a particular story. I did want to ask about what your work might mean for the individual stories themselves. I know you talk a lot about who gets quoted. Can you expand on that?

Krupnikov: Journalists are under a lot of constraints in terms of who and what they cover and what stories are going to get readership. So I think that’s something to keep in mind. On the other hand, when the focus is on political divides and polarization and political conflict, there is a tendency to engage with people as “exemplars,” a person who is going to fit a mold. What this means is that when people see examples of politics in the news, they see the people who are at the upper end of polarization. It’s possible that these people make for more interesting stories. It’s probably more boring to read a bunch of person-on-the-street interviews where people are like, “Nah, I don’t really care. I just want the election to be over.” Or “I don’t really talk about politics.” So it makes sense that these voices aren’t necessarily in the news as often. On the other hand, if we think of the news as giving us stories, or, to quote Walter Lippmann, giving us the “pictures in our heads” of what other people look like, it means that the images we see in our heads are often of the most partisan, the most vocal, the most kind of intense people in politics. 

It means that even if somebody is covering a rally, you’re not necessarily going to talk or hear the voice of somebody who’s kind of just hanging out, maybe playing on their phone, just sort of there. You’re going to hear from somebody who is super into it. And I think the choice of that exemplar, is the most interesting in some way. 

The difficult part, then, is that even if there are surveys to suggest that these people might be in the minority, or even if there’s other quantitative evidence to suggest that there’s something unique about this group of people, there could be something more powerful to the exemplar. 

Loker: I know you’re not a journalist, and you acknowledged the constraints they have, but do you have any kind of practical guidance for journalists to get away from that framing or that behavior? What do you think concretely they might do?

Krupnikov: So, I will say first, that I admire journalists to such a tremendous extent. Their jobs are so difficult. 

I think one thing that would be great would be to try to fit these people that you’re talking to — these people in the street interviews — into a greater quantitative perspective. Is there a way to suggest what proportion of the public this person’s actually representing? This is really, really difficult. Even within a political party, it is difficult to figure out, is this person speaking for the average Democrat or Republican? So is there some way to use quantitative data to figure out the size of the group this person is representing, which I think is really hard. It’s hard for me and I work with quantitative data. 

The other thing that I think is more practical is to think how expertise is used in stories. Some of my other research – with John Barry Ryan, Katie Searles and Hillary Style — is about expertise in the news. We find that journalists tend to return to very same experts over and over throughout stories. But let me go back to what I was saying about how social science is really messy. Analyses, measurement — these are all choices that scholars make. They’re often very justifiable choices, but they are still choices, and social scientists often disagree. And so the question is, is what’s being presented by journalists as the expert consensus actually the expert consensus? If you are covering politics, is there a benefit to broadening the number of people you’re interviewing as experts, to get a broader scope of interpretations of what something might mean for American politics? I agree that quoting people on the street who just don’t care is kind of boring. But I think the place where you can bring in more nuance is in broadening the voices of experts.

Loker: Could we discuss the social environment that many of today’s journalists operate in. You show for instance that the deeply involved spend a lot of time following and talking about politics, and posting about it. That sounds a lot like some journalists. Are there any tendencies of the deeply involved that journalists might be on the lookout for in their own behavior? 

Krupnikov: The first thing I want to say is that the difference between journalists and the deeply involved is that following politics constantly is the journalists’ job. They’re being paid for this behavior. Whereas the deeply involved are not. They’re choosing to spend their time in this way. 

I do think that one of the tendencies, though, if you are embedded in a certain context, is that things that happen in that context loom very large — in a way that might not be the case for people who are functioning in a totally different environment. And so, I think one of these contexts has been Twitter. If you’re spending a lot of time seeing what people are talking about on Twitter, what other journalists are talking about, there’s this tendency to see these issues as being really, really huge when most people aren’t on Twitter, they have no idea why this thing is trending. So I think this is something to remember: the things that are important to your networks might not necessarily be all that important to people outside of those networks. 

And I think intellectually, everyone knows this. I think 100% journalists realize and know this, but in the moment this may not necessarily be top of mind. I think that this can be good and bad, ultimately. On the one hand, if we think of Twitter as a place where anyone can potentially have a voice, this is going to mean that people who in the past could never reach journalists now have the capacity to do so. And I think that there’s potentially a tremendous benefit to that. On the other hand, the downside is that debates and discussions that are specific to a very particular network, now become all-encompassing and really important. So I think it’s a double-edged sword. 

Another thing to keep in mind is that oftentimes the people who are going to be most critical of journalists, the people who are going to communicate with journalists the most, are people who probably have a tremendous amount of involvement in politics. And so that means you’re hearing from people who care the most about politics. How you interpret these signals, I think, should be through the filter of “these are the people who care the most.” And I think again, intellectually, everyone knows this and everyone realizes this, but psychologically, in the moment, it becomes hard to process this. 

Loker: I’m interested in what you think this “other divide” may mean for building audiences beyond the deeply involved. For example, many journalists in the business today want to engage wide swaths of their communities in events or engagement projects related to the news, things like listening projects or efforts to inform what topics they focus on. There’s different kinds of “engaged journalism” initiatives and I’m just curious how, with this idea in mind, journalists might make sure their efforts aren’t only reaching or appealing to the deeply involved. 

Krupnikov: Yeah. So actually, since this book, we’ve been doing some work on basically what leads people into politics or what leads people to avoid politics. Our research suggests that what is pivotal here is, actually, quite literally the word “politics.” Describing something as political makes a lot of people just want to immediately remove themselves, on the idea that it’s going to be really conflictual: “This is not going to be something that I enjoy.” 

But what’s interesting is that we’ve found that describing things that are entirely political in a way that doesn’t use the word politics actually makes people more willing to discuss them. In some measures, when we talk about issues — so we literally will name an issue — people are more willing to say they will talk about this, even with strangers. But once you say the word “politics,” that’s when they take themselves out. And so I think engaging people who are not deeply involved means describing things in other ways besides politics. 

The other thing that we find in the book is that people who aren’t deeply involved have a tremendous amount of skepticism about people who tweet about politics. People who are deeply involved have the perspective “I tweet about politics or others tweet about politics because they want to do something good.” They want to inform the public. People who are not deeply involved say that the reason others tweet or talk about politics is to either share their own opinions, or to find others who agree with them, or show how many people agree with them, or some other kind of deeply skeptical purpose. So I think in order to create more engagement, you have to get people to actually believe that what you’re trying to do is inform and engage rather than for some more self-serving purpose. 

Loker: What would a news organization inspired by your findings look like?

Krupnikov: The closest answer I can come to is that it would possibly look like a local news organization, where the focus is on connecting people to what is happening in their particular area, and what various decisions and outcomes mean for this particular area. 

A question is whether it’s an organization that avoids the game frame. And that I’m a bit more ambivalent about. I think on the one hand, it is difficult, if not impossible, to cover campaigns without the game frame. I think the game frame helps people who don’t know how to follow politics to pay attention to the news, brings people into the campaign, and communicates a lot of things that are important. On the other hand, I think in this kind of idealized news organization, the game frame is not used constantly. Issues are covered without necessarily framing them as a win or loss for the president. Because I actually do think that the American public can totally handle and be interested in coverage of issues without being told that somebody’s doing better or somebody’s doing worse. I think people definitely can handle an explanation of what a particular issue means for their lives, without that additional element. People can handle learning about inflation without being told whether this is a win or loss for Biden. 

Loker: Would that speak across this other divide you’re talking about? If you removed the game frame, would you keep the deeply involved while also getting the other people interested? 

Krupnikov: The deeply involved are going to keep being interested! You know, I think it would be an interesting thought exercise about what it would take for the deeply involved to no longer be interested in politics. I think it is going to take an awful lot. So the real question is, what does it take for the remainder of people to follow politics? And again, I think a lot of them are following politics — just not hourly reading or posting — so the better question is how can journalists engage with that group in a more productive manner? Like you’re not going to lose the deeply involved.

The American Press Institute supports a number of initiatives on community listening in news, including highlighting examples in its newsletter Need to Know. 

Trusting News, a project of the Reynolds Journalism Institute and API that helps journalists earn trust and demonstrate credibility, recently began a new initiative: the “Road to Pluralism.” If you’re a journalist interested in the themes in the above interview, consider getting involved.