Social Media’s Reward Structure Fuels Misinformation Spread, USC Study Finds

A groundbreaking study from University of Southern California researchers has identified what may be the primary driver behind the proliferation of fake news online: social media platforms’ fundamental design that rewards users for habitually sharing attention-grabbing content.

Published Monday in the Proceedings of the National Academy of Sciences, the findings challenge long-held assumptions that misinformation spreads primarily due to users’ lack of critical thinking skills or because of political bias clouding judgment.

The research team discovered that just 15% of the most habitual news sharers were responsible for disseminating between 30% and 40% of all fake news. This concentrated effect prompted researchers to examine the underlying motivations driving these users’ behavior.

“Our findings show that misinformation isn’t spread through a deficit of users. It’s really a function of the structure of the social media sites themselves,” explained Wendy Wood, USC emerita Provost Professor of psychology and business and an expert on habit formation.

The study, which involved 2,476 active Facebook users ranging from 18 to 89 years old, found that social media platforms function similarly to video games, with built-in reward systems encouraging continued engagement. Users who frequently post and share eye-catching information receive validation through likes, shares and comments – reinforcing the behavior regardless of content accuracy.

“Due to the reward-based learning systems on social media, users form habits of sharing information that gets recognition from others,” the researchers noted. “Once habits form, information sharing is automatically activated by cues on the platform without users considering critical response outcomes, such as spreading misinformation.”

Gizem Ceylan, who led the study during her doctorate at USC Marshall and now works as a postdoctoral researcher at Yale School of Management, emphasized that while individual attributes contribute to misinformation spread, platform design plays a more significant role.

“We know from prior research that some people don’t process information critically, and others form opinions based on political biases, which also affects their ability to recognize false stories online,” said Ceylan. “However, we show that the reward structure of social media platforms plays a bigger role when it comes to misinformation spread.”

The impact of these sharing habits proved substantial. The researchers found that users’ social media habits doubled or even tripled the amount of fake news they shared, outweighing other factors including political beliefs and critical reasoning abilities. Most strikingly, frequent, habitual users forwarded six times more fake news than occasional or new users.

“This type of behavior has been rewarded in the past by algorithms that prioritize engagement when selecting which posts users see in their news feed, and by the structure and design of the sites themselves,” explained Ian A. Anderson, a behavioral scientist and doctoral candidate at USC Dornsife who co-authored the study.

Through several experiments, the team uncovered additional insights. They found that habitual users share both true and false news indiscriminately, demonstrating a general insensitivity to content accuracy. Interestingly, these users shared politically discordant news—stories challenging their own political beliefs—as frequently as content aligning with their viewpoints.

The researchers also tested alternative reward structures, discovering that incentivizing accuracy rather than popularity doubled the amount of truthful content users shared. This finding suggests misinformation spread isn’t inevitable but rather a product of current platform design choices.

The study offers three significant conclusions: habitual sharing of misinformation can be changed; users could be incentivized to develop sharing habits favoring truthful content; and effectively reducing misinformation would require restructuring online environments that currently facilitate its spread.

These findings point to a potential solution beyond content moderation. Rather than focusing exclusively on which information is posted, social media platforms could implement structural changes to their reward mechanisms to limit misinformation proliferation.

The research was supported by the USC Dornsife College of Letters, Arts and Sciences Department of Psychology, the USC Marshall School of Business, and the Yale University School of Management.