Among more than 16,000 American registered voters who interacted with fake news sources on Twitter during the 2016 U.S. presidential election, individual consumption of such content was extremely concentrated. Only a small fraction — many older, conservative and politically engaged — accounted for most fake news exposures and shares on the social media platform, a study has found.
Fake news — defined by the study authors as news-like content by publishers who lack traditional editorial standards — has become a topic of public concern, particularly with regard to its role in the most recent presidential election.
The findings, published in the January 25 issue of Science, revealed that only about 1% of the Twitter users studied accounted for 80% of exposures to fake news content. Furthermore, 0.1% were responsible for 81% of the fake news shared. For most voters in the study, representing both sides of the political spectrum, political news exposure came from factual media outlets.
According to lead author Nir Grinberg of Harvard University and his team of researchers, the results of the study are "reassuring," in contrast to claims of political echo chambers and fake news garnering more attention than real news during this period.
The study comes on the heels of a recently published Science Advances study by Princeton University's Andrew Guess and colleagues that evaluated the spread of fake news on the social media platform Facebook during the last presidential election. Like the work in Science, the study by Guess and his team reports that many social media users did not share fake news articles during the 2016 presidential campaign, and the small number who did were older and more politically conservative. Such individuals shared nearly seven times as many articles from fake news sources compared to those in the study's youngest age group.
The two studies in Science and Science Advances add to the small but growing body of research that attempts to understand the fake news phenomenon in the modern age, from disingenuous news sources and the pernicious misinformation they pedal to its broader impact on society and factually informed citizens.
Worries about fake news are nothing new, according to Grinberg and the authors. In 1925, Harper's Magazine published an article warning about the ways in which emerging technologies were making it difficult for people to disentangle rumor from truth.
"The motivation of the [current] research was to better understand the news ecosystem, which has been undergoing radical change over the last generation," said Northeastern University researcher and co-author David Lazer.
Unlike other earlier studies, the authors focused on the sources of fake news, rather than individual false or questionable stories, by defining fake news as content coming from sources that do not adhere to established journalistic practices or have the "organizational intent or processes to produce accurate information," said Lazer.
In the relatively short history of fake news studies, researchers have noted the difficulties of quantifying the scope and scale of the phenomenon — a task complicated by the challenge of measuring human behavior through immense amounts of social media data, which can be further obfuscated by the generally unknown role of automated "bot" accounts.
To overcome these difficulties and better understand how U.S. citizens experience misinformation and interact with fake news sources, the authors leveraged Twitter data linked to public voter registration records, studying the tweets sent by more than 16,000 accounts from August to December 2016. By identifying the various celebrity and high-profile people followed by each account, the researchers inferred the political leanings of each Twitter user.
According to Lazer, fake news was modestly prevalent, accounting for about 5% of the total election-related content in the study, and heavily concentrated in a small group of individuals.
"I think the most surprising thing was the level of concentration on sharing, that almost all the fake news in our sample of 16,000 came from just 16 people," said Lazer, who admitted that he did not believe the striking results at first.
The researchers' conclusions — that fake news was largely read and shared by a tiny fraction of Twitter users — conflict with the results of other similar studies, including a 2018 Science study by Soroush Vosoughi and colleagues from the Massachusetts Institute of Technology which reported on the pervasive and rapid diffusion of fake news.
"We cannot ignore the irony that a field of research on misinformation has come to resemble the very thing it studies," writes Derek Ruths, a McGill University researcher and author of a related Science Perspective in the January 25 issue.
According to Ruths, there is little consensus concerning the definition of fake news, which is likely to blame for the conflicting conclusions of various studies. For example, Vosoughi and team focused on fake news at the level of the individual story and concluded that fake news spreads more rapidly and widespread on social media than real news.
Ruths emphasized that when underlying approaches are more carefully considered, the two studies together tell a more detailed and cohesive story — one of small and specific communities that engage with fake news sources, who are far more likely to produce misinformation.
One of Grinberg and team's most curious findings, also mirrored by Guess and his team, suggests that political conservatives are more disposed to consume and share fake news. While remaining largely unexplained, Ruths suggests that these conclusions may not be as cut and dry as they initially seem. In a more likely scenario, political liberals also likely embed and spread misinformation in ways that have not yet been identified, he writes.
Lazer hopes that continued research will provide the public with an accurate understanding of the scale of fake news and inspire social media platforms and publishers to work with academics to understand this type of misinformation.
"I think generally we need to reflect on the role institutions play in getting us to more accurate understandings of the world. We can directly witness and evaluate very little about our understandings in the world, we need to rely on mapmakers, newspapers, science," said Lazer. "We need to be able to trust those institutions - and those institutions in turn need our trust."