Researchers: Flagging fake news on Facebook is ineffective

Yale University researchers have concluded that tackling fake news on social media through fact-checkers and flagging questionable posts has had little effect.

According to the psychologists, flagging possible fake news stories has only a very limited impact on whether readers perceive the content as ‘true’. Adding comments such as ‘this message is being questioned by fact-checkers’ results in study participants being only slightly more likely to correctly label messages as false. Compared to the ‘doubt flags’ situation, there was a drop of only 3.7 percentage points in the number of flagged messages deemed accurate.

At the same time, the researchers found that marking can also backfire. Particularly among members of the Trump supporters group and the group of adults under the age of 26, it was found that flagging fake news can ultimately increase the likelihood that these people believe fake news. This has to do with the multitude of fake stories on social media. Due to this sheer number, it is impossible for fact checkers to check every story. The presence of markers on some fake stories turned out to lead to the situation that these two groups were more likely to believe other fake stories without markers. The researchers had not specifically designed the study for age groups; it was only after the analysis of the data, based on an algorithm, that a pattern was seen in adults between the ages of 18 and 25.

7,500 people were interviewed for the study. The researchers initially used a control group and presented them with 24 randomly selected headlines; 12 of those were real, while the other 12 stories were fake. The group was then asked how they rated the accuracy of the news stories. The headlines of the stories were all from posts that appeared on Facebook in 2016 and 2017. The members of the control group were able to recognize real news as such in 59 percent of the cases; fake news was believed in 18.5 percent of the cases. This experiment was repeated several times with several other groups, only 6 of the 12 fake stories were marked as ‘disputed’.

The researchers also state that no firm conclusions can be drawn from this study; one of the researchers told Politico that the effects are only minimal and that more research is needed. In addition, they hope that Facebook will make more data available in the future. A Facebook spokesperson has responded to the investigation, questioning the method. He pointed out that the research was conducted via an internet survey and not via Facebook itself. And according to the spokesperson, fact-checking is just one of the methods Facebook uses to combat fake news.

The research has been published as a paper entitled Assessing the Effect of “Disputed” Warnings and Source Salience on Perceptions of Fake News Accuracy.