Science reports:

Did you see the article claiming Kamala Harris joked about killing Mike Pence and Donald Trump? Or the one about large numbers of Trump votes being secretly switched to Joe Biden? If stories like this, run by fake news sites such as Infowars or Gatewaypundit, popped up in your social media feed about the 2020 U.S. presidential election, they probably came from a tiny group of people with a massive impact.

A mere 2000 or so “supersharers” spread 80% of content from fake news sites in a sample of more than 600,000 U.S. voters on X (formerly Twitter), according to an analysis published today in Science. The posters were more likely to be women and older—challenging the stereotype of social media manipulators as young, alt-right men—and they had a huge reach: More than one in 20 users in the data set followed at least one of these supersharers.

The research “is a valuable addition to our understanding of who shares unreliable news on social media,” says Brendan Nyhan, a political scientist at Dartmouth College who was not involved in the work. It also points to a possible solution, he says: “Simple limits on retweets would constrain the spread of this information while having little effect on the vast majority of users.”

The new findings back up previous studies. In 2019, for example, Nir Grinberg, a computational social scientist at Ben-Gurion University of the Negev, and colleagues showed that in a sample of more than 16,000 Twitter users taken around the 2016 U.S. presidential election, 80% of tweeted news from untrustworthy websites came from just 16 users. But who were these superspreaders?

To find out, Grinberg’s team dove into a far bigger data set comprising 660,000 U.S. X users who used their real name and location, allowing the researchers to match them with voter registration data. About 7% of all political news shared by these users on any given day came from untrustworthy websites such as Infowars and Gatewaypundit, the researchers found. And just 2107 users were spreading 80% of the fake news.

The average supersharer was 58 years old, 17 years older than the average user in the study, and almost 60% were women. They were also far more likely to be registered Republicans (64%) than Democrats (16%). Given their frenetic social media activity, the scientists assumed supersharers were automating their posts. But they found no patterns in the timing of the tweets or the intervals between them that would indicate this. “That was a big surprise,” says study co-author Briony Swire-Thompson, a psychologist at Northeastern University. “They are literally sitting at their computer pressing retweet.” [Continue reading…]