Testing the real-world impact of partisan news

If there is one thing we all can agree on, it’s that Americans with strong political opinions can’t seem to agree with anyone on the other side of the partisan divide. Polarization runs deep.

Republican or Democrat, right or left, or some variation thereof — strong partisans of various stripes hold a distressingly dark view of their political opponents, see them as idiotic, and would prefer not to interact them, thank you very much. While not everyone or every issue is so riven, there’s little question that relations between the two major political sides have grown more rancorous.

Who’s to blame for this state of affairs?

While research generally suggests that “elite behavior, rather than communication, is driving political polarization” — that is, it’s especially the fault of politicians and other high-profile figures who stoke division — many people say it’s a media problem: that partisan news outlets, in particular, so thoroughly favor one side and demonize the other that it naturally accelerates polarization among the populace. Seems logical enough, right? But most evidence about the impact of partisan news on the American electorate has been based on surveys (which are inherently limited snapshots of self-reporting) and experiments (which can measure short-term effects but are less useful in detecting long-term, more enduring impact).

To develop a clearer picture of whether (and how) partisan news contributes to American political polarization, a team of researchers — Magdalena Wojcieszak, Sjifra de Leeuw, Ericka Menchen-Trevino, Seungsu Lee, Ke M. Huang-Isherwood, and Brian Weeks — conducted a massive study of “real-world” data that was recently published in The International Journal of Press/Politics, a leading venue for research on journalism and political communication.

The scope of this study is impressive. The project combined two two-wave panel surveys, which capture attitudes at Time 1 and allow researchers to query the same people at Time 2, looking for differences based on media exposure in between. To get around the limitations of people’s self-reported recollections of media use, each set of survey-takers had agreed in advance to allow the researchers to get access to three months of their browsing history. In total, then, the researchers had a year’s worth of browsing history from a diverse sample of more than 1,200 Americans — or about 38 million site visits in all. The researchers then teased out how much of that online activity focused on news domains — whether partisan or centrist sites — and, using a machine learning classifier, assessed how much of that news exposure was about politics in particular.

The study’s title unmistakably spells out the headline finding: “No polarization from partisan news.”

That is, actual online exposure to partisan news, whether congenial (e.g., Fox News for conservatives) or dissimilar (e.g., Fox News for liberals), didn’t appear to make participants’ policy attitudes any more extreme, didn’t lead people to hold more contempt toward supporters of the other party, and seemed to have no polarizing impact on Democrats or Republicans or even those who are strongly partisan. Similar things were true for exposure to political stories within partisan or centrist news websites.

“Taken together, these null results run counter to the popular narrative that partisan news is to blame for the ills of contemporary U.S. politics,” the authors write. “Although not aligned with past evidence, we argue that our null findings portray the reality of (very limited) effects of partisan news in the real world more accurately.”

Speaking of that real world and why news (however partisan) may not make much of a discernible difference: In what may be a reality check for journalists and media scholars, this study found that news is a tiny fraction of what people engage with online. About 1.69% of the web tracking data involved visits to news domains; of those, more than half were visits to centrist news sites, with liberal and conservative sites accounting for about 26% and 18% of all news browsing, respectively. An even smaller share of the overall news visits was focused on political news articles. In fact, even on partisan news sites, more than half of what people consumed was nonpolitical content such as sports analysis or cooking recipes.

“Ultimately,” the authors concluded, “an average participant encountered only one partisan political news article for every 200 sites they visited!”

The “powerful partisan news” narrative, the authors argue, is inaccurate “primarily because politics is a small drop in the overall ocean of what citizens do online. Theoretically, people use media that satisfy their needs and desires. Because politics is perceived as complex, boring, or overly divisive, people may avoid it altogether, especially as they have nearly unlimited entertainment and nonpolitical content at their disposal.” (Read: a bottomless well of Netflix, YouTube, etc.)

Partisan news, it should be noted, is an even smaller drop in that ocean of online content — less than 1% of all the URLs accessed by participants in this study.

What can we take away from this research? For one thing, it could be slightly concerning that so little news is being consumed relative to everything else, leaving citizens less informed about politics and public affairs. On the other hand, though, this study indicates that partisan news, in a real-world test, may not be nearly as polarization-fueling as conventional wisdom might suggest — even if, say, Fox News and other partisan outlets, along with the political elites who make frequent appearances via such outlets, are no doubt consequential as agenda-setters for the larger conversation about contentious politics.

Research roundup

“A sadness bias in political news sharing? The role of discrete emotions in the engagement and dissemination of political news on Facebook.”. By Ernesto de León and Damian Trilling, in Social Media + Society.

For the better part of a decade, we’ve understood emotion as a key to news sharing. BuzzFeed’s Jonah Peretti long ago argued that understanding emotion was key to his company’s social media success, and scholars have measured that idea in numerous ways. Sometimes they’ve found that arousing positive emotions leads to more sharing, sometimes that negative emotions are the key, and sometimes that just arousing strong emotions is more important than what type.

De León and Trilling add some useful shading to this discussion in a study of news articles on Facebook during the 2018 Mexican elections. They looked at the Facebook reactions (i.e., Like, Love, Wow, Angry, etc.) and shares of more than 16,000 news articles posted on Facebook during the election. One of their main contributions was to examine the different roles of the Angry and Sad reactions, as well as the Wow reaction. Sadness and anger have long been distinguished in emotional and information processing research, but as signals of emotion on Facebook, the two reactions have often been grouped together.

De León and Trilling found, as they expected, that negative news articles were shared more often than positive ones. But they were surprised to find out that sadness, not anger, was associated with greater sharing in news articles. They proposed a few possible explanations for this finding, including a desire to avoid conflict among close ties on Facebook and a use of news sharing as a way to collectively express grief.

They also found that Facebook’s Wow reaction was more negatively than positively associated (suggesting disbelief rather than amazement) and that Wow was significantly more strongly related with sharing than the Love reaction. On the whole, negative emotions ruled when it came to political news sharing, but we may want to look more closely at sadness as a key emotion in news consumption beyond the far better publicized anger.

“Conservative news nonprofits: Claiming legitimacy without transparency.” By Michael Buozis and Magda Konieczna, in Journalism.

The past few years have brought a trove of excellent research on both nonprofit journalism and conservative news as important niches within the modern news ecosystem. But until now, little work had been done on the overlap between the two — the small but potentially influential world of conservative news nonprofits. Buozis and Konieczna analyzed the discourse from and about many of these organizations to produce an insightful study on their relationship to mainstream journalism’s structures and norms.

Using the sociological concept of boundary work, Buozis and Konieczna revealed a delicate dance in which conservative news nonprofits seek to expand journalism’s boundaries to include themselves while flouting some of its key norms. The most prominent of those is funding: Almost none of the organizations they examined were transparent about their funding, something that is considered an ethical standard throughout the rest of nonprofit journalism.

These organizations often rhetorically committed to nonpartisanship and political independence, aligning themselves with journalistic norms. But Buozis and Konieczna argued that by obscuring their funding sources, they quietly undermined those commitments by failing to acknowledge the threat that private funding might pose to their journalistic independence.

The organizations employed a similar ambiguity in characterizing their relationship to the mainstream news media. They touted their associations with professional news associations and mainstream news organizations while also presenting their own work as a critique of mainstream journalism, saying most news organizations had abandoned the ideal of objectivity. The norms and structures of mainstream journalism, the authors conclude, are both integral to their journalistic aims but also serve “an important symbolic role representing the failures of journalism that these organizations claim to correct.”

“You are fake news! Factors impacting journalists’ debunking behaviors on social media.” By Magdalena Saldaña and Hong Tien Vu, in Digital Journalism.

We’ve all been on high alert about online misinformation for quite a while now, but journalists may be in a position to see more of it than many others, simply because they spend so much of their team scanning, collecting, and organizing information online. But what should journalists do about that misinformation? Should they ignore it? Debunk it on social media? Write about it as part of their work? There’s no clear guide for journalists who are encountering it each day.

Saldaña and Vu set out to find out what journalists do when they encounter misinformation on social media, and what factors make them more likely to intervene. They surveyed a random sample of 405 U.S. journalists to examine their opinions about misinformation and its relationship with their practices when they encounter it on social media.

They found an interesting, though perhaps unsurprising, paradox. Journalists did say they deeply cared about misinformation and its impact on journalism, democracy, and audiences. But they rarely did anything to respond to it when they found it on social media, other than ignoring it. Saldaña and Vu proposed a few possible reasons: Increasing workloads mean they don’t have time to deal with it, they don’t want to antagonize polarized audiences, they don’t want to give the misinformation oxygen, or they see that as the job of fact-checking teams or social media platforms.

But they also found a few notable factors that led journalists to debunk misinformation. Those who felt a stronger sense of obligation toward their followers were more likely to publicly debunk, as were women. And those who felt social media companies should be held responsible for misinformation were more likely to intervene, but privately. The norm, however, was simply letting it go, for better or worse.

“Journalists’ misjudgment of audience opinion.” By David Nicolas Hopmann and Andreas R.T. Schuck, in The International Journal of Press/Politics.

As Hopmann and Schuck described it, their study was prompted by a puzzle: Journalists have been shown to portray conservative opinions more often than liberal or progressive ones. Why? Hopmann and Schuck hypothesized that it’s because journalists actually believe public opinion is further to the right than it is, as another recent study has shown.

But they also tested three hypotheses about why journalists might misjudge audience opinion: (1) Decades of being told they’re liberal has led journalists to believe audiences are to their right politically; (2) conservative local governments lead to more conservative estimates of audience opinion; and (3) journalists interact most often with the most conservative parts of their audience.

Hopmann and Schuck tested these ideas with historical data from two parallel sets of surveys of German journalists and the German public, from 1993/1994 and 2005. They found support for the first two but not the third. Over both time periods, journalists misjudged their audiences as being more conservative than the audiences saw themselves. That effect was present for journalists across the political spectrum but greater for more liberal journalists. It was also greater for journalists in more conservative local governments, but showed no differences based on amounts of audience interaction.

As the authors noted, the data came from before the social media era, but they offer an interesting glimpse into the possible drivers of journalists’ perception of their audiences’ political opinions, and especially their misperceptions of those audiences as more conservative than they are.

“Journalists on Instagram: Presenting professional identity and role on image-focused social media.” By Diana Bossio, in Journalism Practice.

Journalists’ presence on social media (often Twitter) has been found to be a somewhat odd amalgam of personal branding, promotion of their own or their organization’s work, and guarded interaction with often angry audiences. Bossio took some of the questions behind those findings — what do journalists do on social media, and how do their professional roles impact their use of it? — to Instagram, which has received less scholarly attention than other platforms.

Bossio analyzed the Instagram accounts of 50 Australian journalists and interviewed 20 of those journalists. She found Instagram use that tended run along two tracks simultaneously: The first was self-branding, built heavily on Instagram’s “culture of microcelebrity,” with large amounts of behind-the-scenes photos and many aspirational images that balanced a more personal tone with subject matter that still largely revolved around professional settings.

The second was perhaps more notable, as Bossio found a strong theme of emotional and relational labor running throughout journalists’ Instagram practices. Particularly in their interviews, journalists suggested that much of what they did on Instagram was intended to relate to audiences, make them feel included, and build community. Bossio concluded that while researchers have tended to emphasize tension and burnout from social media among journalists, they have “often ignored the reciprocity of care and belonging forms of online expression bring to journalists” and the pleasure those practices bring journalists.