Do people learn more about politics on shopping sites than news sites?

When people want to make the case that journalism matters, they often begin with perhaps the most foundational and least controversial idea: that people rely on professionally produced news about public affairs to make decisions that shape the social and democratic wellbeing of communities.

And it’s undeniably true. For decades, research has shown that news media play this vital role, for societies and for individuals — that when people follow the news, they are more likely to be informed, vote, engage in their communities, and so on.

What’s not to like?

Well, apparently a lot, it turns out. As we’ve discussed extensively in this newsletter, research has also shown — especially in recent years — how thoroughly dissatisfied many people are with journalism. Growing numbers of people avoid the news, actively or unintentionally. They find it exhausting and anxiety-inducing. They feel overloaded. And the messy partisan politics they associate with it can make news seem “toxic.”

All of this was true to one degree or another during the legacy media era, but this frustration with news — or just indifference to it — has gone into hyperdrive in a digital media environment awash in bottomless alternatives to news and politics (think: everything else more interesting, from Netflix and YouTube to games and social media). With more media choice, research shows, news consumption typically declines.

And yet, there’s also the possibility that people will stumble upon news and political information in the course of their online wanderings, or as a result of what algorithms serve up to them. Search engines and social media can be powerful engines for incidental exposure to news, and some people happen upon news and politics even when going online for very different things, such as romance or trivia, much as moviegoers once saw newsreels before films and television viewers learned about politics on The Oprah Winfrey Show.

So, some big questions emerge: How much news and politics are people actually exposed to online? And where is that coming from — is it from visits to news sites or from people’s individual browsing to non-news domains, and does that matter? Plus, it’s an open question how much political information people are really getting when they visit news sites, given that many consumers may be more interested in news about sports or entertainment rather than politics.

These questions were addressed by a team of researchers led by Magdalena Wojcieszak, in a paper published recently in the journal Political Communication. The team’s methods were impressive. The researchers studied 7,266 online users across three countries with different media and political systems (the U.S., the Netherlands, and Poland). They relied on a three-wave panel study: every three months, the same participants completed 20-minute surveys and, importantly, also shared their online browsing history (at least on their desktops/laptops, not on mobile).

This way, the researchers could test for effects over time, and they also could simultaneously assess attitudes and beliefs (via survey) as well as actual behaviors (via browsing data), a powerful combination that helps alleviate problems with many studies that depend on self-reports alone (i.e., what people can remember about what they saw online). “From over 106 million visits to over 65 million URLs in over 655,000 domains,” Wojcieszak and colleagues wrote, “we match the visited domains to identifiable news domains as determined by our open-source lists of news sites per country, for a total of 2,366 news domains visited across the countries.” But it was important to establish the existence of political information on non-news sites, so they created a “multi-lingual classifier” that could detect political content both in news and outside (e.g., in shopping or entertainment sites).

Their headline finding? Non-news websites expose people to more political content than news websites.

“Out of every 10 visits to political content, 3.4 come from news and 6.6 from non-news sites,” the researchers noted.

This is partly a function of infrequent news use by most people: of participants’ total online browsing, just 3.4% of it was to news sites, which tracks with other research suggesting that news constitutes about people’s 4% of total online consumption. And when participants in this study did visit news sites, they usually read about something other than politics: Only between 14% (Netherlands) and 36% (U.S.) of the visits to news sites involved political content.

So, where did they get their political information? From entertainment, shopping sites, celebrity gossip pages — all the things that people prefer in comparison to news and public affairs. These non-news sites emerged as the dominant source of political information simply by virtue of the volume of activity there compared with the little time spent on news sites.

“Even though politics in these kinds of [non-news] sites comprised only 1.6% of all visits,” the authors write, “the aggregate popularity of webmail, entertainment, shopping sites, or celebrity gossip means that an average citizen encounters most political content outside news [our emphasis added]. People, especially Americans and especially those with low political interest…encounter politics more frequently outside news outlets than within.”

Just as significantly, the researchers found that the effects of encountering political information — for example, the impact on people’s intentions to participate in politics or their propensity to believe misinformation — were “astonishingly similar” between news and non-news sites.

Wojcieszak’s team concluded: “exposure to political content outside news domains had the same — and in some cases stronger — associations with key democratic attitudes and behaviors as news exposure.”

There are a couple ways to think about these findings. On the one hand, they paint a somewhat dispiriting picture of the tiny role that news consumption plays in most people’s lives — a role that appears comparatively small even within the realm of learning about politics, of all things. On the other hand, perhaps it’s comforting to know that no matter where or how people gain information about politics, that political knowledge still may have relatively positive effects for democracy and society. In all, this research suggests we have more to learn about the place and influence of news and politics in people’s everyday lives.

Research roundup

“Attacks against journalists in Brazil: Catalyzing effects and resilience during Jair Bolsonaro’s government.” By Joao V.S. Ozawa, Josephine Lukito, Taeyoung Lee, Anita Varma, and Rosental Alves, in The International Journal of Press/Politics. Violence against journalists has long been a threat around the world, but it has become an increasing concern over the past several years, especially with the rise of populist anti-journalist rhetoric since the late 2010s. The abhorrence of acts and threats of violence against journalists is self-evident, but one of the primary long-term concerns with that violence is that it will create a chilling effect in which journalists are intimidated into self-censorship.

Ozawa and a team of researchers from the University of Texas set out to find an empirical answer to the question of whether that chilling effect has occurred in the case of Brazil. Based on a watchdog group’s reports (PDF), they built a database of the state’s physical attacks against journalists as well as social media criticisms of the press from former president Jair Bolsonaro and his fellow right-wing politicians. They then tracked critical coverage of the Bolsonaro administration in Brazil’s leading news organizations and compared the three datasets over time. They also interviewed 18 journalists who were the victims of attacks during the Bolsonaro administration.

It’s an impressive array of data, and the results are multi-layered and nuanced. But one of the main findings was that rather than a chilling effect, a catalyzing effect took place — manifesting not as more coverage (attacks had no effect on future critical coverage), “but as persistent coverage despite ongoing criticisms and threats (social and physical).”

Not all news was good: The researchers found that the social media criticisms of Bolsonaro and his fellow populist politicians did lead to a significant rise in attacks against journalists. And they found in their interviews that media ownership had a significant chilling effect and that journalists often felt little support from their editors. But journalists also said that network support from fellow journalists and professional organizations was crucial, and the authors called for those networks of solidarity to be strengthened, especially in Latin America.

“Speaking the language of market segmentation: How newsworkers describe their organization’s target audience.” By Stephanie Edgerly and Kjerstin Thorson, in Journalism. The audience has been more visibly at the center of journalism for a couple of decades now, but in some ways their influence may come through not in themselves, but in how journalists imagine them to be. This helps shape a massive range of news practices — how journalists pursue those audiences, what kinds of stories they produce, how they frame them, how they distribute them, and so on. Seen this way, journalists’ perception of the audience may be at least as powerful a force in news as the audience itself.

Edgerly and Thorson examined those perceptions of audiences with a view toward market segmentation, a staple of the news business since the 1980s. They surveyed 1,387 American journalists, asking them, “How would you describe the target audience of the organization where you currently work?” They then broke the answers down into demographics, psychographics, and some demographic sub-categories, like race and ethnicity or social class.

Not surprisingly, the language of market segmentation (that is, demographics and psychographics) was widespread, appearing in 87% of descriptions. Two out of five descriptions contained only demographic terms, with a quarter containing only a psychographic description and about a fifth containing both. Demographics occurred more commonly in local journalists’ and TV journalists’ descriptions, with psychographics less common among those journalists, as well as those in small newsrooms, and in newspapers and radio. (Newspaper journalists were more likely to include minimal descriptions with neither demographics nor psychographics.)

What was perhaps most interesting was the gap many journalists described between what their target audience was and the more inclusive audience they wanted to have. Some of those journalists articulated an ability to cover that gap, with responses like, “Like many local newspapers, our readership tends to skew older and whiter than the city the paper serves, but leaders at the paper seem like they are genuinely trying to broaden our audience to younger and nonwhite people.”

But for others, the more traditional, elite, and narrow audience was treated as a simple function of market-driven journalism. “A common response was to acknowledge a sense of inevitability,” the authors wrote. As many journalists perceived it, “if a news organization is to survive, it must rely on wealthy, white, educated, and/or older audiences.”

“How loud does the watchdog bark? A reconsideration of losing local journalism, news nonprofits, and political corruption.” By Nikki Usher and Sanghoon Kim-Leffingwell, in The International Journal of Press/Politics. One of the truisms regarding the decline of local news is that it opens the door to increased corruption, because the journalistic watchdog is no longer able to eye business and public officials as closely. Several studies have attempted to measure whether this has in fact occurred, but it’s a complex issue to look at empirically, especially since it’s so difficult to measure corruption that goes entirely unreported.

With that in mind, Usher and Kim-Leffingwell measure corruption effects in the U.S. by looking at public prosecutions of corruption, reasoning that they may be more likely if local news is strong, since corruption has to be exposed to be prosecuted. They measured local news decline by looking at local newspaper employment as well as local newspaper circulation, and they also examined the role of local news nonprofits through both their presence and their funding as a proxy for their strength.

They found no significant effects of newspaper employment numbers on corruption prosecutions, but higher newspaper circulation was correlated with more corruption prosecutions. “If news about a corrupt public official is circulating widely throughout a community,” the authors reasoned, “then it is likely that sanction is not only warranted [but also] expected.”

Their most encouraging results surrounded nonprofit news organizations. Both the presence of nonprofit news outlets and their funding were positively related to corruption prosecutions. The challenges with measuring these corruption effects remain — we still can’t measure corruption that didn’t occur because of the threat of exposure by local journalists — but the authors concluded that their study provides some of the first empirical evidence that “nonprofit interventions in failing local commercial news markets may be an important safeguard for keeping public officials accountable.”

“Blurring boundaries: A longitudinal analysis of skills required in journalism, PR, and marketing job ads.” By Jana Bernhard and Uta Russmann, in Journalism & Mass Communication Quarterly. Journalists have long sought to maintain a professional (and rhetorical) boundary between themselves and public relations, claiming exclusive expertise and moral high ground for their profession. But they’ve also bemoaned the increasingly porous nature of that boundary, especially with many journalists moving into PR as their profession has become more precarious and with “brand journalism” finding journalists and PR professionals producing editorial products under clients’ direction.

Bernhard and Russmann looked at the overlap between those two professions (as well as marketing) over time through the lens of the skills required to join them. They analyzed 336,000 job postings from Germany and Austria between 2015 and 2020, using both automated text analysis and some further manual categorization to come up with 25 “hard” skills (programming skills, managerial skills, audiovisual skills, etc.) and 10 “soft” skills (work ethic, creativity, teamwork, etc.) among them.

They found relatively small differences in the rates of both hard and soft skills demanded across the disciplines, with the biggest differences in some expected areas, like sales skills being heavier for marketing jobs and content generation heavier for journalism. The demand for digital skills increased significantly from 2015 to 2020, particularly digital marketing and digital media generally. The fastest-growing skill demanded was programming, though it was still much less frequently demanded than the others.

The authors concluded that this digitalization of journalism, PR, and marketing has accelerated the overlap of these three fields, making it increasingly difficult for journalism to distinguish itself — especially since its blurring with PR and marketing also influences its editorial independence, a primary professional characteristic of journalism. The three fields are still distinguishable from each other in terms of skills, they wrote, but the margin between them is growing quite thin.

“Journalism in flux: The changing news industry in Latin America, 2013–2021.” By Summer Harlow, Vanessa de Macedo Higgins Joyce, and Amy Schmitz Weiss, in Journalism Studies. There’s a basic narrative of the past 10 to 15 years of journalism that has been so deeply ingrained in the minds of journalists and scholars that it’s reached the point of rote history: shrinking newsrooms, a scramble for revenue, once-promising startups, erosion of trust, and the ascendance of a digitally centered era driven by social media, data, audience metrics, and now automation. As a broad narrative, that’s pretty accurate and widely encompassing. But it’s helpful to take a step back and look for the trends and shifts under the surface of that narrative that might be escaping our attention.

That’s what Harlow and her colleagues have done with this study. They surveyed Latin American journalists across 20 countries in 2013, 2017, and 2021, asking many of the same questions each time about their work practices and the most important changes at their news organizations. They found several of the troubling trends with which we’ve become familiar: smaller newsrooms, more concern about disinformation and governmental interference, and more worries about journalists’ physical safety and mental health.

But they also found some positive trends which have helped to combat those threats. Journalists reported significantly more collaboration in recent years, especially transnationally. They also reported several signs of increased professionalization: better training, more investigative units, strengthened resolve for editorial independence, and the deeper institutionalization of digital tools.

Social media use offered a useful representative case. Latin American journalists’ amount of use of social media has not significantly changed since 2013, even within platforms like Facebook, Twitter, YouTube, and Instagram. But the professionally purposeful nature of their use — for reporting, branding, monitoring news, and interacting with audiences — have all significantly increased over that time. They also were more likely to have a social media policy at their organization. The overall picture was a macro-level glimpse of not necessarily the increased use of a technology, but the institutionalization and professionalization of that technology over the past decade.