Human bias, not algorithms, fuels distorted social-media reality

Researchers at Virginia Tech have identified a cognitive pattern they label “rareness bias diffusion,” whereby users preferentially engage with and repost uncommon or sensational content, causing a cumulative distortion of perceived reality on social media. In a peer-reviewed article in MIS Quarterly, Alice Jang and Viswanath Venkatesh detail six experiments, including scenarios of monster attacks, that show participants overwhelmingly share the rarest versions of events.

The team argues that platform algorithms cannot fully mitigate this bias because it originates from human decision-making. They cite examples such as exaggerated fears over a cyclospora outbreak, the over-hyped Tide Pod Challenge, and disproportionate coverage of wars, all of which shape public sentiment and even policy responses. The authors suggest that journalists and users adopt more contextual reporting and selective following habits to counteract the skewed narrative. Their findings highlight the broader impact of personal sharing choices on collective understanding of health, safety, and global events.