Popular Social Media Users More Likely to Share Misinformation, Study Finds

A comprehensive analysis of X (formerly Twitter) data has revealed that users with more followers, higher tweeting frequency, and newer accounts are significantly more likely to share low-factuality content online.

Researchers analyzed over 1.67 million users and nearly 14.7 million tweets from 2022, assigning each user a factuality score based on the credibility of news sources they shared in their most recent 500 tweets containing URLs. Using these scores, users were categorized into high, middle, and low factuality groups.

The study specifically focused on regular users rather than automated accounts or public figures by applying filtering criteria such as capping follower counts at 10,000 and limiting average daily tweet activity to 32 posts.

“Our findings highlight clear patterns in social media behavior that correlate with the quality of information users typically share,” said a researcher familiar with the study. “The social network metrics we examined can serve as initial indicators for identifying users who might be more prone to sharing misinformation.”

Among the most striking findings was the relationship between popularity and factuality. Users with higher follower counts were significantly more likely to share low-factuality content compared to those with fewer followers. Statistical validation using Mann-Whitney U tests confirmed the significance of this pattern.

The research also found that users with higher daily tweet frequencies were more likely to fall into the low factuality category. This aligns with existing literature on information overload, suggesting that higher activity rates may correspond with less careful evaluation of shared content.

Similarly, users who followed more accounts were generally more likely to share low-factuality content. However, regression analysis revealed a more nuanced relationship, showing a reverse U-shaped pattern where both low and high factuality users typically follow fewer accounts than middle factuality users, with the effect being stronger for high factuality users.

Account age emerged as another significant factor. The longer a user had been registered on the platform, the more likely they were to share high-factuality content. This finding supports theories connecting digital literacy with information sharing behaviors, suggesting that experience on social media platforms may correspond with better content discernment.

“What’s particularly interesting is how these factors interact with each other,” noted a social media researcher not involved in the study. “For instance, the effect of tweet frequency on factuality differs depending on how many accounts a user follows. Among users following many accounts, higher tweet activity actually associates with higher factuality – which wasn’t what we expected.”

The researchers employed multinomial regression analysis to examine these factors both independently and in combination, finding significant relationships between all four social network metrics and factuality levels. The robustness of these findings was verified through eight alternative dataset constructions, consistently showing that tweet frequency and account age were the most reliable indicators across different testing conditions.

These findings have important implications for platform policies, fact-checking organizations, and media literacy initiatives. By identifying easily accessible social network metrics that correlate with misinformation sharing, platforms could potentially develop more targeted approaches to combating the spread of false information.

The research adds valuable insight to ongoing discussions about social motivations, attention economics, and digital literacy in the age of social media, suggesting that basic user metrics may provide meaningful signals about content reliability.