Twitter ‘supersharers’ of fake news tend to be older Republican women

About 80 percent of the fake news shared on Twitter during the 2020 US presidential election came from just 0.3 percent of users, according to researchers from Israel and the US.

These “supersharers” were disproportionately likely to be older, Republican women from Texas, Florida, and Arizona, according to a study published in the journal Science by Sahar Baribi-Bartov, Briony Swire-Thompson, and Nir Grinberg.

And their output was not the result of automation, it’s believed, but rather reflects “manual and persistent retweeting,” the academics say. “These findings highlight a vulnerability of social media for democracy, where a small group of people distort the political reality for many.”

… a vulnerability of social media for democracy, where a small group of people distort the political reality for many

From dataset of 664,391 registered US voters who were active on Twitter (now known as X) during the US presidential election between August and November 2020, just 2,107 supersharer accounts were identified.

On average during the study period, about 7 percent of all political stories shared by the survey participants linked to fake news sources. But 2,107 supersharers were responsible for 80 percent of the nonsense.

The study’s authors, affiliated with Ben-Gurion University in Israel and Northeastern University in the US, characterize these enthusiastic social media users as supersharers because their posts reached a broad audience.

“Despite being only 0.3 percent of the population, supersharers reached 5.2 percent of registered voters in our sample and are about a fifth of the heaviest consumers of fake news,” they explain.

These supersharers thus had more reach than Russia’s foreign influence campaign on Twitter in 2016, based on estimates that 3.4 percent of Americans on Twitter at that time followed a Russia-controlled account. And measured in ad spending, the researchers estimate that political candidates would have had to spend $20 million for equivalent message distribution.

The academics contend that these avid tweeters, by flooding the digital commons with their views, undermine the democratic conceit that people have an equal voice in public debate – a debatable proposition in the US given that media ownership has long conveyed disproportionate social influence.

They compare the phenomenon of supersharing to authoritarian governments using information to sway public opinion, but argue that influence campaigns by voters in a democracy haven’t been thoroughly studied. And they contend that misinformation research needs to consider not just coordinated campaigns but also political distortion carried out by the shoutiest on social media.

To address concerns about manipulation driven by persistent posting, the authors argue for turning down the volume on the social media megaphone by limiting the reach of online posts.

Platform interventions that target supersharers or impose retweet limits could be highly effective

“Our research shows that platform interventions that target supersharers or impose retweet limits could be highly effective at reducing a large portion of exposure to fake news on social media,” the researchers argue.

“Interventions are always tricky and somewhat complicated,” Nir Grinberg, an assistant professor at Ben-Gurion University, told The Register. “On the one hand, our work shows that platforms could very effectively cut down the mass volume of fake news spreading on their platform by suspending or slowing down supersharers.

“On the other hand, we also found that such intervention is still very far from eliminating the fake news that’s available to supersharer followers, suggesting that they are embedded in an ecosystem where misinformation is prevalent.

“I personally think that what we’re seeing right now is platforms with very loose limits on speech, and my personal opinion is that we are seeing mostly potential harms and abuse of this unlimited quantity of speech, while the benefits of letting it run without limitation are not that clear.” ®