• Featured Chart
  • November 3, 2021

Sharing fake news

Source: dikushin

Nearly everyone would like to believe the news that proves them right and ignore the reports that prove them wrong. Recent research suggests that this confirmation bias prevents people from being skeptical and makes them potentially more likely to spread lies on social media.

However, in a paper in the American Economic Review, authors Marta Serra-Garcia and Uri Gneezy take a step back and show that individuals are overconfident in their ability to detect lies even without the motivation to believe them.

In an experiment, the authors first incentivized a group of participants (“senders”) to record short videos to convince others that they were seeing and describing a true news event. In one case, the senders were looking at a real headline and picture from The New York Times. But in another case, they were viewing a blank screen. The researchers then asked a different group of participants (“receivers”) to guess whether the sender was describing a real news story or was lying. 

The receivers were hardly better than a coin flip at detecting lies, but believed they were right about two-thirds of the time.

These receivers were then either incentivized to share true videos or to share videos that they thought would be believed by a second group of receivers, regardless of whether the video was true or not.

Figure 4 from the authors’ paper shows how often the first group of receivers shared lies under the two scenarios.



Figure 4 from Serra-Garcia and Gneezy (2021)


The chart shows that when the first group of receivers were incentivized to share a truthful video, 62 percent of them shared a lie—a rate that was significantly higher than the actual fraction of lies from senders (red horizontal line). When receivers were incentivized to share a video that would be believed, they shared lies 58.2 percent of the time.

The results indicate that receivers were significantly more likely to share lies than truthful videos, even when they were incentivized to share truthful videos. The researchers also found that the second group of receivers were significantly more likely to watch and believe a video if the first group of receivers believed it was true.

The authors argue that this combination of overconfidence in detecting lies and reliance on shared content being truthful may help explain why fake news is so prominent and influential.

Mistakes, Overconfidence, and the Effect of Sharing on Detecting Lies appears in the October 2021 issue of the American Economic Review.