The spread of fake news on social media has been a problem for years. But there may be a simple solution to stop people sharing inaccurate information.
Gordon Pennycook at the University of Regina in Canada and colleagues have found that people can identify fake news easily, but may unwittingly share misinformation on social media because they aren鈥檛 thinking analytically.
The researchers presented more than 2500 people from the US with real headlines and images taken either from mainstream news stories or from a cache of stories that had been debunked by independent factcheckers as entirely false.
Advertisement
In the first part of the study, some participants were asked to indicate if they would consider sharing the headlines on social media. Many said they would, whether or not the headline was actually true.
In detail, it turned out that people鈥檚 likelihood of sharing a headline was more contingent on whether or not it supported their political beliefs. For instance, if a headline was false but concordant with someone鈥檚 political leanings 鈥� whether towards the political left or right 鈥� the chances of the person sharing it were 37.4 per cent. But the chances of sharing dropped to 24.0 per cent for headlines that were discordant with someone鈥檚 beliefs but true.
As part of a follow-up study, the team asked another group of people to judge the accuracy of headlines before asking them whether or not they would consider sharing them on social media. They found that people who were given this accuracy prompt were significantly less likely to consider sharing false headlines.
Building on this second finding, the researchers explored whether people using Twitter in their daily lives could be encouraged not to share stories that they suspect to be fake. The team identified thousands of users who had previously shared news from potentially misleading websites, and used Twitter bots to message almost 5500 of them.
The researchers asked these users to rate the accuracy of a single non-political headline. Afterwards, they measured the trustworthiness of the news shared by these users, using a trustworthiness scale developed by independent factcheckers.
More trustworthy
The researchers found that their prompt asking users to rate news accuracy seemed to have an effect. The news that the Twitter users shared in the 24 hours following a prompt was rated as 3.5 per cent more trustworthy than the news shared before they had been contacted.
While some people may deliberately share false content, the improvement suggests that people are more likely to spread misinformation because they aren鈥檛 thinking about truth and accuracy while on social media, says Pennycook.
鈥淚t鈥檚 an entertainment medium that people go on often when they are, for example, taking a break from thinking,鈥� he says. 鈥淚t probably isn鈥檛 very conducive to people stopping and checking about whether something they鈥檙e reading is true or not.鈥�
鈥淲e need to do things to nudge them towards being a little bit more reflective,鈥� says Pennycock.
The team would like to partner with social media platforms to scale up the experiment to see what effect an accuracy prompt has on users.
Reference:
