Media
Why People Share Misinformation
How to pay more attention to the truthfulness of what you like.
Posted March 28, 2021 Reviewed by Gary Drevitch
Key points
- New U.S.-based research finds that we're good at spotting inaccurate social media headlines, but if they are in line with our politics, we often share them anyway.
- When deciding what to share, we value getting likes and demonstrating our political allegiances more than we value accuracy.
- The pattern of sharing inaccurate posts occurs among both Republicans and Democrats.
You’re scrolling through social media and notice that one of your really smart friends has once again shared a somewhat tantalizing story that you know is wrong. The story has a good hook in its title, but it’s ultimately just misinformation. Why does your smart friend do that? A new article in the journal Nature explains why so many of us unthinkingly forward misinformation and it’s a little disappointing, if understandable (Pennycook et al., 2021). This international team of researchers found that most people do actually value accuracy, and want to share truthful articles, but in the heat of the social-media moment, they consider factors other than accuracy—like the number of likes it might receive.
A Nonpartisan Pattern of Sharing Inaccurate Posts
In a series of experiments, Gordon Pennycook and his colleagues (2021) set out to explore the contrast between what we believe and what we share. First, they reported that participants in their U.S.-based studies were quite good at identifying which headlines were accurate and which were not, regardless of whether the headlines matched their political leaning. Despite this proficiency, participants were about twice as likely to share a false article that represented their political viewpoint as they were to rate it as accurate. Specifically, close to 20% of people rated false headlines that matched their politics as accurate; yet close to 40% of people said they would share these same news stories.
In one particularly telling example, the researchers explained that only about 16% of Republicans judged the following headline as accurate: “Over 500 ‘Migrant Caravaners’ Arrested with Suicide Vests.” But more than half of Republicans said they might share this article. (This pattern repeats on both sides of the partisan divide.) The authors reported that people are more attuned to getting “likes” from their followers and making their allegiances clear than they are to sharing accurate information.
What Can Social Media Platforms Do?
It may seem hopeless, but these researchers identified some tactics that social media platforms might use so that people prioritize accuracy in their shares. In one study, the researchers recruited more than 5,000 Twitter users who had shared articles from sites that regularly traffic in misinformation. The researchers sent private messages to these Twitter users and asked them to rate the accuracy of an apolitical headline. After the intervention, the accuracy of these users’ posts increased by 5 to 10% — they had been primed to think more about accuracy. An increase of 5 to 10% may seem small, but when you consider the sheer number of Twitter posts, such a decrease in the amount of shared misinformation could be quite substantial.
What Can We Do?
There are two main takeaways from this series of studies. First, it’s important to realize that the people sharing news stories may not even believe them—and that they might not even share them if they took a moment to think about them. So maybe we should judge a bit less and politely push back when a story seems false. Second, make it a habit to ask yourself how accurate each headline seems before you share. If you don’t trust yourself to slow down and do this, maybe put a post-it note that reads “Is this really true?” on the back of your laptop or smartphone as a regular reminder. And if in doubt, scroll on.
References
Pennycook, G., Epstein, Z., Mosleh, M., Arechar, A. A., Eckles, D., & Rand, D. G. Shifting attention to accuracy can reduce misinformation online. Nature (2021). https://doi.org/10.1038/s41586-021-03344-2