Supersharers of fake news on Twitter during the 2020 US presidential election

Social media have made it possible to rapidly expose large segments of society to misinformation. Most studies to date have focussed on the role of foreign influence campaigns and bots, but relatively few have examined the role of ordinary citizens in spreading misinformation online.

Recent studies have found that a small fraction of people, “supersharers”, are responsible for the majority of fake news shared on social media. Who are these people – if they are people at all – and how much of an impact do they have?

The study involves a large panel of more than 600,000 registered US voters who were active on (the social media platform formerly known as) Twitter during the 2020 US presidential election.

Results show that on an average day, 7% of all news shared by the panel linked to fake news sources. An extremely small fraction (0.3%, or 2,107 people) accounted for 80% of the tweets linking to fake news sources (side note: i.e. one of the fake news websites that were found in this study by Grinberg et al.). This happened throughout the entire election period. On a daily basis, the average supersharer posted considerably more links to political news (16) than the heaviest sharers of non-fake political news (5) and the average panelist (0.3).

These supersharers are not shouting into the void: 5.2% of registered voters on Twitter directly follow a supersharer. Supersharers had significantly higher network influence (i.e. they are very well-connected) and had more people interacting with their content than other Twitter users.

A fifth of the heaviest consumers of fake news in the panel follows a supersharer. Moreover, the average follower of a supersharer was 2.5 times as likely to see fake news compared to the average panelist. Finally, supersharers were responsible for nearly a quarter of the fake news available to their average follower, and were the only source of fake news for about 11% of their followers.

So who are these supersharers?

The researchers found that supersharers have a significantly higher proportion of women, older adults (on average about 58 years old) and Republicans, and were significantly more likely to be Caucasian compared to the panel and heavy sharers of non-fake political news. Supersharers were also found to be overrepresented in three US states (Florida, Arizona, and Texas), have an average of 0.3 fewer education years, and a $2500 higher annual income compared to those in the panel.

Based on results of the bot detection tool Botometer, only a small percentage (about 7%) of supersharer accounts can be considered bots. Posting times of supersharers are not significantly different from those in the panel, nor does tweet metadata suggest that supersharers use apps that support automation more. The only real difference is that supersharers retweet more often (75%) compared to the panel (33%). This may explain why many supersharers managed to evade Twitter’s attempts to purge fake news from its platform.