You'll Never Guess What Group of People Shared the Most Fake News Around the 2016 Election

Illustration for article titled You'll Never Guess What Group of People Shared the Most Fake News Around the 2016 Election
Photo: Getty

Old people may have the worst internet journalism literacy, according to common sense, and partially backed up by new research on the spread of so-called “fake news” immediately before and after 2016 presidential election.


A study published today in the Science Advances journal suggests that elderly Facebook users in the U.S. were far more likely to share false and misleading news stories than any other age group.

The study, conducted by researchers at New York University and Princeton University, focused on Facebook activity in the months surrounding the election. The team started by gathering a panel of 3,5000 Facebook users and non-users in spring of 2016. Right after the presidential election, researchers requested that the Facebook users in that group install an app that shared their Facebook data, including political and religious views, with researchers. Facebook approved the app to be used for research. Almost half of the participants who were Facebook users consented to sharing their profile data.

In the following months, researchers monitored the links participants posted on Facebook. They compared those links with a list of websites based on the fake news sites list that BuzzFeed assembled for its in-depth analysis on how “fake news” outperformed “real news” on Facebook prior to the election.

The study found that more than 90 percent of participants didn’t share any links from the sites included in the “fake news” list. But most people who did share articles on the “fake news” list were older than 65.

The results also show that people who identified as Republican and conservative were more likely to share false and misleading stories than Democrats and liberals (18 percent versus 3.5 percent). However—while researchers did try to remove partisan websites from their list (like Breitbart)—many of the websites included on the list were being used to spread pro-Trump and anti-Clinton propaganda.

Reached by Gizmodo, Facebook did not want to comment on the study itself. But a spokesperson shared information on the company’s efforts to provide context that’s intended to help users check the validity of stories in their feeds. They also shared information on research showing how Facebook’s efforts are apparently helping to address the false news problem.


Of course, a lot has happened in the online and media landscape since 2016. As the authors of the study wrote in a Washington Post op-ed about their research, “Our results focus on social media behavior from over two years ago. Online, that’s virtually an eternity.”

In the Washington Post article, they also acknowledge they “don’t know whether ‘being old’ is associated with sharing more fake news, or if it’s being in the generation born before 1950" and they can’t say “whether people raised in the modern digital environment—digital natives—will share more fake news as they age.”


Hopefully, research like this will help social networks figure out where to focus their efforts in the ongoing and potentially futile war on bullshit.

[Science Advances, Verge, Washington Post]


Former senior reporter at Gizmodo


Dense non aqueous phase liquid

I’m guessing most news discriminated throughout facebook is simply the product of confirmation bias. We tend to like what we like.

So this kind of says that news produced for conservative consumers tends to be more fake than news produced for liberals’ consumption.

Weighting in statistics is kinda important. In this case, facebook usership:

Weighted to total US population:

0-14 years: 18.73% (male 31,255,995/female 29,919,938)
15-24 years: 13.27% (male 22,213,952/female 21,137,826)
25-54 years: 39.45% (male 64,528,673/female 64,334,499)
55-64 years: 12.91% (male 20,357,880/female 21,821,976)
65 years and over: 15.63% (male 22,678,235/female 28,376,817) (2017 est.)

Facebook weighting w.r.t two groups:

15 or 18 to 24 y.o: 39.4/( 22.2 + 21.1) = say around 90% penetration of group

65+: 21.1/(22.6 + 28.3) = 61% penetration

Ergo: there might be very discerning and smart old folks who only watch PBS Newshour and read New York Times NOT on facebook.