How fake news spreads online

The spread of misinformation on social media is an alarming phenomenon that scientists still do not fully understand. Although data shows that false claims are increasing in the online world, most studies looked at small samples or just the spread of individualized false news.

Professor Sinan Aral of the Massachusetts Institute of Technology in the United States, along with his colleagues Soroush Vosuughi and Deb Roy, have decided to analyze the potential for the spread of both true and false information on Twitter since the creation of the microblogging in 2006 until the year 2017.

Information Chains

To begin with, scientists assembled approximately 126, 000 “tweet cascades” (uninterrupted strings of common and unique retweets) involving information spread across 3 million people, more than 4.5 million times.

The result was considered disturbing by the researchers: According to them, false news spread faster than true, and significantly. The work prepared by the team was published in the journal Science.

How the work was done

The researchers began their work by identifying hundreds of fake and true news, using information from six independent fact-checking sites. These organizations showed considerable agreement (between 95 and 98%) about the truth or falsity of this information.

Then they searched Twitter for mentions of these news, following the sharing activity for the “source” tweets (the first mention of an information on Twitter), tracking all the “cascades” of retweets from each of the source tweets. . Then the scientists analyzed how they spread online.

Amazing results

For all categories of information - politics, entertainment, business, and so on - researchers have found that fake news has a much wider reach and spreads faster and more widely than the real thing.

According to Aral and his team, fake news is 70% more likely to be retweeted, even when the source account’s lifetime, activity level, number of followers, and followings are verified, and if Twitter confirmed the account as genuine. In addition, false political information spreads faster than any other kind of news.

Robots? What robots?

To the scientists' surprise, users who spread fake news averaged fewer followers and followed, were less active, spent less time microblogging, and had fewer verified accounts against those who spread true news.

And despite concerns about the role of automated robots in spreading fake news, they found that human behavior contributed more to the spread of information than bots did. Using existing bot detection algorithms, the researchers found that bots accelerated the spread of true news at about the same rate as they accelerated the spread of false news, suggesting that false information spreads more than true information as a result of human activity.

What is the explanation?

What motivates people to spread false news with greater intensity than the real one? One explanation is the "novelty factor." For researchers, the unprecedentedness of fake news may attract more human attention, encouraging sharing, and giving participants a more "informed" status.

Scientists' analysis seems to confirm this hypothesis. Using computerized methods to infer emotional content from word usage, they found that fake news inspired more replies on Twitter expressing surprise than true news. The truth, on the other hand, inspired more joy and confidence. Such emotions can shed light on what inspires people to share fake news.