The truth about online fake news is becoming clearer. A new study shows that on Twitter, phony stories reach more people than truthful ones do. Fake stories also spread far faster.
Fake news refers to stories based on false or misinterpreted information. These stories try to dupe readers into believing something that isn’t true. Some might try to make public figures look bad or claim people did something they didn’t. Others might try to discredit scientific findings. Such stories are often shared on social media platforms such as Twitter and Facebook. But scientists have lacked data on how widely they were shared, or by whom. So a team of researchers decided to investigate.
They recently analyzed more than 4.5 million tweets and retweets. All had been posted between 2006 and 2017. And their disturbing finding: Fake news spreads faster and further on Twitter than true stories do.
Filippo Menczer studies informatics and computer science at Indiana University in Bloomington. He was not part of the new study but says its findings are important for understanding the spread of fake news. Before this, he notes, most investigations used a few people’s observations rather than a mountain of scientific data. Until now, he says, “We didn’t have a really large-scale, systematic study evaluating the spread of misinformation.”
Deb Roy, who did work on the new analysis, studies media and social networks at the Massachusetts Institute of Technology in Cambridge. In the past, he also has worked as a media scientist for Twitter. To study how news spreads on Twitter, Roy and his colleagues collected tweet cascades. These are groups of messages composed of one original tweet and all retweets of that initial post. They examined about 126,000 cascades centered on any of about 2,400 news stories. Each of those original news stories had been independently confirmed as true or false.
The researchers then collected data on how far and fast each cascade spread. Discussions of bogus stories tended to start from fewer original tweets. But they tended to soon spread extensively. Some chains reached tens of thousands of users! True news stories, in contrast, never spread to more than about 1,600 people. And true news stories took about six times as long as false ones to reach 1,500 people.
Overall, these data show, fake news was about 70 percent more likely to be retweeted than was real news. The team reported its results in the March 9 Science.
Not just bots
Roy’s team also wanted to know who was responsible for spreading false news. So they looked at Twitter accounts that were involved in sharing fake stories. Some had been run by computers, not people. These so-called web robots, or bots, are computer programs that pretend to be human. They have been designed to find and spread certain types of stories.
Some people had assumed that bots drive most fake news moving across the internet. To test that, Roy and his colleagues looked at data both with and without bot activity.
Bots spread false and true news about equally, the data showed. So fake news could not be blamed just on bots, Roy's group concluded. Instead, people are the main culprits in retweeting fake news.
Why might people be more likely to spread tall tales? These stories may seem more exciting, says data scientist Soroush Vosoughi. He works with Roy at MIT and is a coauthor of the new study. Compared to the topics of true-news stories, fake-news topics were more different from other tweets that users had viewed in the two months before they retweeted a story. Tweet replies to the false news stories also used more words indicating surprise.
The researchers didn’t inspect the full content of every tweet. So they don’t know exactly what users said about these stories. Some people who retweeted fake-news posts may have added comments to debunk them. But Menczer calls the new analysis a “very good first step” in understanding what types of posts grab the most attention.
The study also could guide strategies for fighting the spread of fake news, says Paul Resnick. He works at the University of Michigan in Ann Arbor. Though he was not part of the new study, he uses computer science to study how people behave online. One approach might be for social media platforms to discourage people from spreading rumors, he says. That approach might have more impact than simply booting off bots that behave badly.
Sinan Aral at MIT has some other ideas. He is another coauthor of the new study and an expert on how information spreads through social networks. One way to fight fake news might be to help users identify true stories online, he suggests. Social media sites could label news pieces or media outlets with truthfulness scores, Aral suggests. In fact, at least one September 2017 study has already looked into that. The bad news: Flagging potentially false headlines or news sites only works a little, it found. Sometimes the tactic could even backfire.
Platforms also might try to restrict accounts reputed to spread lies, Aral says. But it’s still unclear how successful such actions might be, he adds. Indeed, he notes, “We’re barely starting to scratch the surface on the scientific evidence about false news, its consequences and its potential solutions.”
bot (short for web robot) A computer program designed to appear that its actions come from some human. The goal is to have it interact with people or perform automated tasks such as finding and sharing online information through social-media accounts.
coauthor One of a group (two or more people) who together had prepared a written work, such as a book, report or research paper. Not all coauthors may have contributed equally.
colleague Someone who works with another; a co-worker or team member.
computer program A set of instructions that a computer uses to perform some analysis or computation. The writing of these instructions is known as computer programming.
computer science The scientific study of the principles and use of computers. Scientists who work in this field are known as computer scientists.
data Facts and/or statistics collected together for analysis but not necessarily organized in a way that gives them meaning. For digital information (the type stored by computers), those data typically are numbers stored in a binary code, portrayed as strings of zeros and ones.
dupe To fool.
informatics The study of how humans create, process and understand information. Informatics is useful in many areas such as healthcare, ecology and studies of human behavior.
information (as opposed to data) Facts provided or trends learned about something or someone, often as a result of studying data.
internet An electronic communications network. It allows computers anywhere in the world to link into other networks to find information, download files and share data (including pictures).
media (in the social sciences) A term for the ways information is delivered and shared within a society. It encompasses not only the traditional media — newspapers, magazines, radio and television — but also Internet- and smartphone-based outlets, such as blogs, Twitter, Facebook and more. The newer, digital media are sometimes referred to as social media. The singular form of this term is medium.
network A group of interconnected people or things. (v.) The act of connecting with other people who work in a given area or do similar thing (such as artists, business leaders or medical-support groups), often by going to gatherings where such people would be expected, and then chatting them up. (n. networking)
online (n.) On the internet. (adj.) A term for what can be found or accessed on the internet.
retweet To share a tweeted post with someone else on Twitter.
robot A machine that can sense its environment, process information and respond with specific actions. Some robots can act without any human input, while others are guided by a human.
social media Internet-based media, such as Facebook, Twitter and Tumblr, that allow people to connect with each other (often anonymously) and to share information.
social network Communities of people (or animals) that are interrelated owing to the way they relate to each other. In humans, this can involve sharing details of their life and interests on Twitter or Facebook, or perhaps belonging to the same sports team, religious group or school.
technology The application of scientific knowledge for practical purposes, especially in industry — or the devices, processes and systems that result from those efforts.
tweet (in social media) Message consisting of 140 or fewer characters that is available to people with an online Twitter account.
tweet cascade An initial posting (tweet) on the social-media platform Twitter together with all the retweets of that original post.
Twitter An online social network that allows users to post messages containing no more than 280 characters (until November 2017, the limit had been just 140 characters).
Web (in computing) An abbreviation for World Wide Web, it is a slang term for the internet.
Journal: S. Vosoughi et al. The spread of true and false news online. Science. Vol. 359, March 9, 2018, p. 1146. doi:10.1126/science.aap9559.
Journal: G. Pennycook and D. Rand. Assessing the effect of ‘disputed’ warnings and source salience on perceptions of fake news accuracy. SSRN. September 14, 2017.