It’s not bots, it’s us: False news is shared much faster than real stories on Twitter
A lie can be halfway around the world before the truth has got its shoes on, as the old expression goes. A variant of that phrase was first uttered in the 1850s, and if a lie could get that far in those days, just imagine how it fares now, aided with superfast broadband and an online culture that promotes writing first and thinking later.
You don’t need to imagine. A new study from the Massachusetts Institute of Technology (MIT) published in Science reveals exactly how much more viral false news is when compared to verifiable, newsworthy accounts. False news stories are, the researchers discovered, 70% more likely to be retweeted than true stories. Not only that, but it takes true stories about six times as long to reach 1,500 people on Twitter than it takes a false story, on average.
The phrase “false news” may stand out in an era where “fake news” is the term of choice for every media outlet including, usually, this one. That’s an important distinction, according to the researchers. “We avoid the term ‘fake news’ because A) it has been used recently as a political strategy to label stories that politicians don’t like as ‘fake news’ whether is is true or false and B) because it implies the news was faked, meaning someone intentionally created it to deceive,” Sinan Aral, a co-author on the paper tells me via email. “We do not [have] data on what the intentions of the creators was.”
Whatever the intent, the outcome was the same. The researchers examined around 126,000 stories tweeted over 4.5 million times by around three million accounts between 2006 and 2017. Using six independent fact-checking organisations, they labeled each story “true” or “false” if it achieved 95-98% agreement between the organisations. Unbroken retweet chains – or cascades – reached a depth of ten around 20 times faster than facts.
Not all false news is created equal. “Analysis of all news categories showed that news about politics, urban legends, and science spread to the most people, whereas news about politics and urban legends spread the fastest and were the most viral in terms of their structural virality,” the paper reads.
But what it does have in common is its reach when compared to the truth. “Whereas the truth rarely diffused to more than 1,000 people, the top 1% of false-news cascades routinely diffused to between 1,000 and 100,000 people,” the paper explains. You might think that this is due to a few high-profile troublemakers who routinely distribute unverified accounts, either for fun, profit or political reasons, but that doesn’t seem to be the case, according to Aral. “False news spreaders tend to have fewer followers, follow fewer people, are less often verified, tweet less often and have been on Twitter for a shorter period of time,” he explained. “False news spreads despite these characteristics not because of them.
“But, I am sure that when a person with lots of followers tweets a false tweet it spreads the information more than when a person with fewer followers tweets false news,” he added – though pointed out that the researchers did not examine this directly.
Another piece of conventional wisdom brought into question by the research is the role of bots in the dissemination of false news. “Although the inclusion of bots, as measured by the two state-of-the-art bot-detection algorithms we used in our analysis, accelerated the spread of both true and false news, it affected their spread roughly equally,” the paper reads. “This suggests that false news spreads farther, faster, deeper, and more broadly than the truth because humans, not robots, are more likely to spread it.”
While the researchers make no judgement over whether this false news is believed by those humans who spread it or those who it reaches (“we did not get into the heads of those consuming or sharing false news,” Aral says,) the researchers speculate that the novelty factor of false news stories may encourage its spread, when compared to mundane truths. They took a sample of 5,000 users known to share both true and false news stories and analysed their reaction to a random sample of 25,000 tweets. “We saw a different emotional profile for false news and true news,” says Soroush Vosoughi, a co-author of the paper. While users responded to false news more with surprise and disgust, he explained, true stories generated replies that scored highly for sadness, anticipation, and trust.
“Although we cannot claim that novelty causes retweets or that novelty is the only reason why false news is retweeted more often, we do find that false news is more novel and that novel information is more likely to be retweeted,” the paper reads.
With that in mind, what can be done to stop the flow of false news? Aral has a number of suggestions on this point ranging from the flagging of false news (as Facebook tested back in 2016) to reducing the economic incentive to spread it by reducing its reach. While the report doesn’t dismiss the role of bots and state level meddling, it seems that human nature makes that disruption all the more potent and the solutions could be just as sociological as they are technical.
Photo: Melanie Gonick, MIT