Your research finds that America has two very different media ecosystems — one brings together the political center and the left, and the other is largely confined to the right. What are the key differences between them?
On the right, audiences concentrate attention on purely right wing outlets. On the left and center audiences spread their attention broadly and focus on mainstream organizations. This asymmetric pattern holds for the linking practices of media producers. Both supply and demand on the right are insular and self-focused. On the left and center they are spread broadly and anchored by professional press.
These differences create a different dynamic for media, audiences, and politicians on the left and right.
We all like to hear news that confirms our beliefs and identity. On the left, outlets and politicians try to attract readers by telling such stories but are constrained because their readers are exposed to a range of outlets, many of which operate with strong fact-checking norms.
On the right, because audiences do not trust or pay attention to outlets outside their own ecosystem, there is no reality check to constrain competition. Outlets compete on political purity and stoking identity-confirming narratives. Outlets and politicians who resist the flow by focusing on facts are abandoned or vilified by audiences and competing outlets. This forces media and political elites to validate and legitimate the falsehoods, at least through silence, creating a propaganda feedback loop.
What tools do you use to map out these media ecosystems, and how do they work?
We use Media Cloud, a data collection and analysis platform that we at Harvard’s Berkman Klein Center for Internet and Society have been developing in collaboration with our colleagues at the MIT Center for Civic Media for a decade.
For this study, we collected 4 million stories from the U.S. presidential election and the first year of the Trump presidency. We analyzed what outlets these stories linked to, how often and how they were tweeted, and how often they were shared on Facebook. We use network analysis of the link network to analyze the authority different sites have among media producers, and of who tweets what and how often to measure audience attention. We use the tweeting of Trump and Clinton followers to identify the mix of political orientations of a site’s audience. And we use a range of text analysis techniques to understand what these outlets discussed and when.
You argue that the most serious problem is Fox News, not Facebook. Why so?
Because that’s where the eyeballs are.
The highly asymmetric pattern of media ecosystems we observe on the right and the left, despite the fact that Facebook and Twitter usage is roughly similar on both sides, requires that we look elsewhere for what is causing the difference.
Surveys make it clear that Fox News is by far the most influential outlet on the American right — more than five times as many Trump supporters reported using Fox News as their primary news outlet than those who named Facebook. And Trump support was highest among demographics whose social media use was lowest.
Our data repeatedly show Fox as the transmission vector of widespread conspiracy theories. The original Seth Rich conspiracy did not take off when initially propagated in July 2016 by fringe and pro-Russia sites, but only a year later, as Fox News revived it when James Comey was fired. The Clinton pedophilia libel that resulted in Pizzagate was started by a Fox online report, repeated across the Fox TV schedule, and provided the prime source of validation across the right-wing media ecosystem.
In 2017 Fox repeatedly attacked the national security establishment and law enforcement whenever the Trump-Russia investigation heated up. Each attack involved significant online activity, but the spikes in attention and transition moments are associated with Hannity, “Fox & Friends” and others like Tucker Carlson or Lou Dobbs.
Why does your research suggest that there isn’t any epistemic crisis on the left, even though some left-wing people have seriously mistaken beliefs?
On both sides of the political spectrum, manipulative clickbait sites lure people to read stories that feed their outrage. Russian propaganda is aimed at both sides on Facebook and Twitter. We all suffer confirmation bias. We all occasionally fall prey to this desire.
But “post-truth” or epistemic crisis are population-level effects — where an entire chunk of the population loses the ability to tell truth from identity-confirming fiction — not individual or small group effects. Because people on the left use a broad range of media mostly subject to mutual policing for facts and misinformation, they regularly encounter stories that give them a reality check. This limits the spread of disinformation in the population as a whole, although it cannot prevent discrete instances of false belief.
On the right, millions of people wake up with “Fox & Friends,” commute with talk radio, have lunch with Limbaugh, or settle in for an evening of Carlson, Hannity and Ingraham, as their president tweets against the fake news media. These are their most trusted news sources, and they have been telling their audiences for over 20 years that they can’t trust anything they hear outside their own propaganda bubble. The outrage industry amplifies, reinforces and monetizes the desire to satisfy confirmation bias, while insulating its audiences from sources that could give them a reality check (and reduce profits).
Does your research shed light on arguments about whether Trump and right-wing media have any responsibility for recent terrorist attacks and attempts?
We have a detailed chapter on how Trump and Breitbart interacted to force the Republican Party to make immigration the number one agenda item, and framed it in starkly Islamophobic terms. We show text-analysis-based evidence that the term “Globalist” originated in antisemitic white supremacist sites, was transmitted to the mainstream by Breitbart while its antisemitic evocation of a global Jewish conspiracy was still clear, and was generalized by Lou Dobbs and Fox more generally.
This article is one in a series supported by the MacArthur Foundation Research Network on Opening Governance that seeks to work collaboratively to increase our understanding of how to design more effective and legitimate democratic institutions using new technologies and new methods. Neither the MacArthur Foundation nor the network is responsible for the article’s specific content. Other posts can be found here.