A telling schism in our fractured, post-truth world is that between “fake” and “real” news. Those who insist on this division associate fake news with Donald Trump, Republicans and the unvaccinated—and place great store in the causal power of information. Only because of fake news, many believe to this day, could a monstrosity like Trump get elected president. Starved of real news, democracy dies in darkness.
I have always found the evidence in support of these propositions to be pretty threadbare. The initial report that pro-Trump “fake election news stories” had “generated more engagement on Facebook than top election stories from 19 major outlets combined” was put forward by Buzzfeed in November 2016. It was a methodological mess. Even if one accepted all the claims, the implications were left hanging. Suppose hundreds of thousands of Catholic voters read an article stating that the Pope had endorsed Trump: Was there any evidence that minds had been changed? The report provided none.
In fact, the subject was never broached. The magical power of the written word was simply assumed—a common failing among those of us who churn out words by the pound. That has been the typical maneuver of participants in the fake news panic. The proliferation of lies is shown to be a fact, while the “so what” is passed over in silence. And, of course, anyone wishing to quantify the spread falsehood will have a field day with the internet—where, as N.N. Taleb has observed, the digital information tsunami is composed mostly of noise.
The web is many things—including the mother of lies. But it’s a long leap from this description of the landscape to proving that “Fake news is bad news for democracy.”
Belief in the power of the word requires a parallel faith in the gullibility of the public. Fake news can only be dangerous if ordinary people can be made to believe in preposterous lies. That’s a risky posture for an analyst: It divides the world into the clever, which is mostly us, and the foolish, which is entirely them. More to the point, it’s almost certainly wrong.
The human mind is a hard object to budge. That is the conclusion Hugo Mercier arrives at in “Not Born Yesterday,” after reviewing the psychology of persuasion. If recipients of information “were excessively gullible,” Mercier writes, “they would be mercilessly abused … until they reached a point where they simply stopped paying any attention to what they were being told.” For good evolutionary reasons, the default isn’t blind faith but hunkering down. When deprived of the means to evaluate information, people “revert to a conservative core, rejecting anything they don’t already agree with, being much harder, not much easier, to influence.”
Fake news and impossible rumors are sometimes deployed after the fact to justify bad behavior, according to Mercier. Bigots who hate and envy Jews might loot their stores and abuse their persons because they wish to do so, then explain, “I heard they kill children in secret rituals.” Such people use rumors to “provide the ‘facts’ that sanction what they want to do anyway.” Cause and effect are inverted; crucially, no persuasion occurs—no changing of minds.
The same principle holds for the 2016 election. Consumers of fake news weren’t bamboozled into voting for Trump; they were feeding a pre-existing hunger.
In 2016, fake news was “a distraction” that “played a relatively small part in the overall scheme of things.” So says a study by Harvard’s Berkman Klein Center that included, among the authors, some of the brightest luminaries in internet research. Another large-scale study, fronted by Brendan Nyhan and conducted in 2019, came to an even more uncompromising verdict. Here is Nyhan on the findings:
[I]t turns out that many of the initial conclusions that observers reached about the scope of fake news consumption, and its effects on our politics, were exaggerated or incorrect. Relatively few people consumed this form of content directly during the 2016 campaign, and even fewer did so before the 2018 election … And, most notably, no credible evidence exists that exposure to fake news changed the outcome of the 2016 election.
Most of the factors that drain fake news of persuasive power apply to “real” or mainstream news as well. So we are led inevitably to the question: To what extent, if any, does influence depend on the institutional status of the news provider? Can The New York Times, say, change our minds in a way that Breibart News can’t?
As it happens, a massive recent study tackled this question directly. The study looked at a year’s online browsing history from a “diverse sample” of more than 1,200 Americans, analyzing around 38 million site visits in total. That included all institutional flavors of news: prestigious, “centrist” outlets, but also highly “partisan” sites like Fox News on the right or MSNBC on the left. Specifically, the authors of the study sought to tease out the degree to which the news media contributes to the political “polarization” of the public.
That the public is increasingly polarized and intolerant of opposing views has been confirmed by a ton of research. That the news media must be implicated in this trend is practically a truism. Yet, whatever part the news has played in polarization, the study learned that it did not involve persuasion—changing minds. Here is a summary of the findings:
[A]ctual 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 towards supporters of the other party, and seemed to have no polarizing impact on Democrats or Republicans or even those who are strongly partisan.
Post-journalism, or abandoning the pretense of objective reporting for openly partisan advocacy, is not responsible for our rabid political environment but merely seeks to profit from it. This business model is unlikely to succeed: The study discovered that only 1.69% of the public’s online activity entailed visits to news domains. Lack of influence thus largely coincides with lack of interest. Whether this should make us happy or sad is a question for another day—but there can be no doubt that our online time is spent on TikTok dances, Amazon purchases and pornography, rather than the news.
Together with a handful of scholars like Andrey Mir, I maintain that it is the structure of information, not the content, that can potentially influence opinion and behavior. That was true in the age of newspapers (see Walter Lippmann) and that of television (see Marshall McLuhan). It’s no less true of the digital age. The structure of the web produces a “Tower of Babel effect” that drives reasonable persons to shriek in the most violent and extreme fashion if they wish to be heard at all.
Yet even here, persuasion may not be part of the story. Allow me to cite just one more study, published in December 2021, which I find utterly fascinating. In an attempt to quantify “social organization and political polarization in online platforms,” Isaac Waller and Ashton Anderson looked at 5.1 billion comments made in 10,000 communities on Reddit over a span of 14 years. Parsing the data, the authors determined that a “significant polarization event” occurred after Trump’s election in 2016. That conforms with our intuitive sense of the history of the digital environment.
But what caused this “polarization event”? I would have thought the answer to be equally intuitive: A host of mild-mannered Reddit users were transformed into howling partisans. Yet that isn’t what the study found. “Contrary to conventional wisdom,” the authors write, “individual-level polarization is rare.” Instead, “the system-level shift in 2016 was disproportionately driven by the arrival of new users.”
While we shouldn’t make too much of a single study, this one presents a plausible alternative story to how the structure of digital information affects opinion and behavior. Those mild-mannered users didn’t suffer a sudden personality disorder after 2016. They simply found the post-Trump Tower of Babel to be an uncongenial habitat and withdrew—or, more accurately, they were swept out by partisan hordes.
In other words, it was a population exchange. If this story is correct, the most apt metaphor for the effects of the digital environment may not be “mass conversion” but rather “barbarian invasion.”