This Just In: Fake News Probably Has Less Impact Than You Think

Written

This Just In: Fake News Probably Has Less Impact Than You Think

By a large margin, more people saw actual election news, not the fabricated kind.
A woman looks at her cell phones
Only 14% of voters regarded social media as the “most important” source of election-related news. | Reuters/Mike Segar

The notion that “fake news” was a major factor in the 2016 Presidential election may be, well, fake.

Although many fabricated news items circulated online prior to the election, far fewer people saw those pieces than actual news stories, Stanford professor of economics Matthew Gentzkow says.

It’s unclear how many of those who did see fake news stories switched their votes as a result, but for fabricated news to have changed the election’s outcome, one fake news story would have had to be as persuasive effect as 36 television ads, Gentzkow said during a recent talk at the Stanford Graduate School of Business. Gentzkow appeared in the first in a series of seminars examining political divisions and conflicts in democracies. The series, organized by students and Stanford GSB’s Center for Social Innovation, has invited such speakers as futurist Gerd Leonhard, Stanford political economy professor Saumitra Jha, and University of Chicago global conflict studies professor Oeindrila Dube.

In a post-election survey of 1,200 people, Gentzkow and co-researcher Hunt Allcott of New York University showed respondents headlines from actual news stories, such as Hillary Clinton’s stumbling by her van at a 9/11 memorial event, and fake items, such as the Pope’s pre-election endorsement of Donald Trump. The survey showed that only about 15% recalled seeing fake stories and just 8% believed them.

To create a placebo, the researchers concocted other election-related headlines, or “fake fake news,” and included those in the survey; responses to those headlines were similar to responses to the actual fake items. By contrast, a majority of those surveyed remembered seeing major actual news items. The research is detailed in a new working paper, “Social Media and Fake News in the 2016 Election,” released in mid-January.

“Happily, more people recalled seeing the true stories than the fake stories,” Gentzkow says. “The extent to which these (fake) stories circulated and were seen has been quite substantially overstated.”

The extent to which fake stories circulated has been substantially overstated.
Matthew Gentzkow

Gentzkow notes that a majority of voters obtain some of their news from social media and that most fake items, which are almost always online, favored Trump. But fake news still had limited influence, because just 14% of voters regarded social media as the “most important” source of election-related news, the study indicated. Broadcast and cable television remains the most important source of news for voters, while 18% of people said they “often” obtain news from social media and 20% say they “sometimes” do so.

“The reality is that digital media and social media are an important but not a dominant force in news and information,” says Gentzkow.

At the same time, social media’s influence on news consumption is increasing, although it’s unclear by how much. Before the rise of social media, people obtained most of their online news from large, politically centrist sites. Those who did visit extremist sites tended to be heavy Internet users who also went to many other online sources as well, including mainstream ones, Gentzkow says. But as news is increasingly funneled through social networks, which themselves are segregated politically and otherwise, people are more likely to find themselves in “echo chambers” and encounter only ideas with which they already agree.

Political polarization, too, is on the rise. Interestingly, that phenomenon can’t be attributed solely to growing usage of the Internet and social media, because it’s greatest among people age 65 and older, a demographic that uses social media far less often than younger groups do, Gentzkow says.

The nation’s current strain of polarization goes deeper than political or policy differences, he adds. It’s rooted in people’s feelings and opinions about others’ personal traits, such as intelligence or selflessness. “It’s much more about what the literature calls ‘affective polarization,’ or how do we see — and how do we feel about — the other side.”

For media inquiries, visit the Newsroom.
Explore More

Insights

Gert McMullen. Credit: Elena Zhukova
August 3, 2020
Written

Maker: AIDS Memorial Quilt — Sewing as a Catalyst for Change

“My greater hope,” Mike Smith says, “is that people will see what we’re doing and realize that they, too, can do something worthwhile — that it’s easy to be helpful if you want to be.”

Insights

Susan Athey. Credit: Tricia Seibold
January 16, 2020
Video

Machine Learning and AI for Social Impact

Innovations developed at big tech firms could transform the nonprofit world, with a little help from academia.

Insights

Szu-chi Huang, associate professor of marketing at Stanford GSB, smiling outdoors in front of a green background. Credit: Nancy Rothstein.
November 27, 2019
Video

Water Instead of Soda: Marketing Good Nutrition to Kids

How marketers are using the power of persuasion to combat childhood obesity.