Peter Salovey President | Yale University
Peter Salovey President | Yale University
Conservative groups often claim that social media platforms exhibit bias against their viewpoints. Research conducted into the 2020 election sheds light on this issue, revealing some significant findings.
The study found that accounts sharing pro-Trump or conservative hashtags were suspended at a rate 4.4 times higher than those sharing pro-Biden or liberal hashtags. This discrepancy might suggest a bias against conservatives. However, the research also indicated that users associated with conservative content were more likely to share links from sources considered low-quality or misinformation-heavy, which social media platforms aim to restrict according to their policies.
These findings imply that the increased suspension rates are linked to Twitter's neutral efforts to limit misinformation spread rather than intentional ideological bias by the platform. The differences in suspension rates arise from varying media consumption patterns between political groups.
To determine what constitutes misinformation in a polarized environment, researchers used a media quality score based on crowd-sourced ratings from politically balanced groups. These raters represented both sides of the political spectrum equally, reducing partisan bias risks in evaluations. The score reflects consensus judgments about news domains' trustworthiness without skewing toward any single political perspective.
The research offers two key lessons for social media platforms today:
1. Polarization and separate media universes: The study underscores severe polarization in the U.S., where individuals consume entirely different types of media depending on their political background—mainstream sources for some and fringe or hyper-partisan ones for others. This fragmentation mirrors broader issues within political discourse as people operate within isolated information bubbles.
2. Careful enforcement during elections: As suspensions predominantly affected one side during an election year, it is crucial for platforms to implement policies cautiously. Even neutral intent enforcement can appear biased and potentially erode public trust or affect election outcomes. Transparency and consistency are vital to ensure anti-misinformation measures do not unintentionally amplify perceptions of political bias during sensitive periods.
A related study titled “Differences in misinformation sharing can lead to politically asymmetric sanctions” explores these dynamics further.
Alerts Sign-up