Peer Corrections, Not Algorithms, Prove Most Effective Against Misinformation

When X (formerly Twitter) launched its user-driven fact-checking system, skeptics doubted that social media users could effectively police the very misinformation they often help spread. A new study, however, reveals this “crowdchecking” approach has produced remarkable results.

Research published in Information Systems Research shows that when community-generated correction notes appear beneath posts containing questionable information, authors are 32 percent more likely to delete those posts compared to when such notes remain private.

“Trying to define objectively what is misinformation and then removing that content is controversial and may even backfire,” explains co-author Huaxia Rui, the Xerox Professor of Information Systems and Technology at the University of Rochester’s Simon Business School. “In the long run, I think a better way for misleading posts to disappear is for the authors themselves to remove those posts.”

The comprehensive study—conducted by researchers from the University of Rochester, the University of Illinois Urbana–Champaign, and the University of Virginia—analyzed 264,600 posts on X that received community notes during two distinct periods: one before a US presidential election (June–August 2024) and another two months post-election (January–February 2025).

X’s Community Notes feature operates on a threshold mechanism. For a corrective note to become publicly visible, it must first achieve a “helpfulness” score of at least 0.4 after evaluation by other contributors. The platform employs a bridging algorithm that prioritizes ratings from users who have disagreed in past evaluations, preventing partisan voting blocks from manipulating a note’s visibility.

This design created a natural experiment. By comparing posts with notes just above the visibility threshold to those just below it, researchers could isolate the causal effect of public exposure on author behavior.

The findings were consistent across both study periods, pointing to a universal pattern rather than a temporary phenomenon tied to election cycles when misinformation typically surges.

What drives authors to delete posts after receiving public corrections? According to the researchers, it’s primarily reputational concern.

“You worry that it’s going to hurt your online reputation if others find your information misleading,” Rui notes. Public correction notes signal to the wider audience that “the content—and, by extension, its author—is untrustworthy.”

The study found that users with verified accounts (indicated by blue checkmarks) were especially quick to delete posts flagged with Community Notes, suggesting those with established online presences have more substantial reputational concerns. Additionally, posts with higher engagement were more likely to be deleted when publicly corrected, as the potential reputational damage increases with visibility.

Speed also plays a crucial role in this ecosystem. The researchers discovered that faster public display of corrective notes led to quicker retractions by post authors. This timing factor is particularly important considering that false information typically spreads more rapidly than corrections on social media platforms.

The study’s findings highlight how social media’s inherent dynamics—status, visibility, and peer feedback—can be leveraged to improve information quality without resorting to heavy-handed moderation or algorithmic filtering.

“Crowdchecking strikes a balance between protecting First Amendment rights and the urgent need to curb misinformation,” the researchers conclude. Rather than relying on censorship or expert determination of truth, it harnesses collective judgment and public correction in a way that respects user autonomy.

Rui admits he was initially surprised by the strength of the findings, particularly given today’s polarized information environment.

“For people to be willing to retract, it’s like admitting their mistakes or wrongdoing, which is difficult for anyone, especially in today’s super polarized environment with all its echo chambers,” he says.

The research team had initially questioned whether public correction notes might backfire, causing users to double down on misleading claims. Instead, they found the opposite—a mechanism that effectively encourages voluntary removal of problematic content.

“Ultimately,” Rui concludes, “the voluntary removal of misleading or false information is a more civic and possibly more sustainable way to resolve problems” in our increasingly complex information landscape.