Why do people spread fakenews? Can’t you see which messages are incorrect? Or do you care about the truth of the matter as long as the message fits your own attitude? A study now provides a third explanation: According to this, most people are basically motivated not to share false content, but think too little about whether a headline is really right. Conversely, if they are reminded to keep the information accurate, they will share more, higher quality messages.
Fake news regularly reaches thousands of users on social media and helps shape political opinion. Social media platforms have already tried various strategies to prevent the dissemination of untrue information: For example, Facebook displays warnings when a message has been identified as untrustworthy, and Twitter experimented with the uncommented sharing of links during the US election campaign to restrict. How successful such strategies are, however, is still unclear.
True or Politically Appropriate?
A team led by Gordon Pennycook from the University of Regina in Canada has now investigated in several consecutive experiments why people share fakenews and what can be done about it. In a first experiment with 1,002 Americans, the researchers asked half of the subjects to rate the accuracy of 36 headlines. The headlines partly agreed with the political convictions of the test subjects, and partly they contradicted it. The other half of the participants should instead indicate whether they would share the respective article on social media.
“If the participants were explicitly asked about the correctness, they rated true headlines significantly more often than false headlines,” report the researchers. Politically consistent headlines were judged to be correct somewhat more often than those that contradicted one’s own attitude. However, when it came to sharing the content, a different distribution emerged: “Whether a heading was politically consistent or not had a much greater effect on the intention to share it than the truthfulness,” according to the researchers. Overall, 37.4 percent of the time, respondents considered sharing news that was wrong but agreed with their views, even though they only believed such headlines to be right 18.2 percent of the time.
People want to share real content
When asked directly, most of the study participants said it was very important to them that the messages they share were correct. “Our results suggest that the vast majority of people across the ideological spectrum only want to share correct content,” says co-author David Rand of the Massachusetts Institute of Technology (MIT) in Cambridge. “It’s not like most people just say, ‘I know this is wrong and I don’t care.'”
But why are so many people still sharing fake news? The researchers found an explanation in another experiment. 1,507 subjects were asked to decide which political headlines they would share on social media. Before this decision-making task, however, the researchers asked some of the test subjects to rate the correctness of a random, non-political headline – thus drawing their attention to the concept of correctness. Indeed, participants who had previously thought about the credibility of news were less likely to share false statements when it came to the political headlines. In two further experiments with representative samples of the US population, the researchers replicated this result.
Field experiment on Twitter
To test whether these results could be applied in practical terms on social media, the researchers carried out a field experiment on Twitter. “We created a number of bot accounts and sent messages to 5,379 Twitter users who regularly shared links to websites with misinformation,” explains Rand’s colleague Mohsen Mosleh. “Just like in the survey experiments, the message asked if a random apolitical headline was correct in order to get users to ponder the concept of accuracy.” In fact, users who received this message were more likely to share news in the following 24 hours from websites that professional factchecker rated as higher quality.
From the researchers’ point of view, these findings suggest that one of the top reasons for fake news sharing is inattentiveness. With a final experiment with 710 participants, they substantiated this hypothesis: If the subjects were first asked to evaluate the correctness of a headline and only then to decide whether they would share it, the percentage of false news that the subjects would share halved from about 30 to 15 percent. From this, the researchers conclude that 50 percent of the previously shared false headlines were shared due to carelessness about accuracy. About a third of the false headlines shared were believed to be true by attendees – meaning about 33 percent of the misinformation was spread due to confusion about its accuracy. Only 16 percent of the wrong news was shared despite respondents realizing it was wrong.
Possible interventions in practice
“Taken together, these studies suggest that when deciding what to share on social media, people are often distracted from considering the accuracy of the content,” the researchers write. One reason for this is that social media present a lot of content in a short time and users get quick reactions to shared content. The concept of the platforms is designed more for user retention than accuracy. “But it doesn’t have to be like that,” say Pennycook and colleagues. “Our findings can easily be translated into interventions that social media platforms could use to increase user focus on accuracy.” For example, they could occasionally encourage users to judge the correctness of headings in their feed.
“The extent to which such standardized requests from platforms have the same effect as a private message is not tested by the study and would be an open question for future research,” says media professor Lena Frischlich from the Ludwig Maximilians University in Munich, who is not at the Study was involved. In principle, however, the study could certainly help to better understand and combat the spread of fakenews.
Source: Gordon Pennycook (University of Regina, Canada) et al., Nature, doi: 10.1038 / s41586-021-03344-2