Republican and Democrat Users Both Affected by Algorithm Manipulation

Partisan identity didn’t protect users from algorithmic influence, according to research that examined both Democrats and Republicans. The experiment found that subtle feed manipulations increased polarization among members of both parties, suggesting that algorithmic effects transcend ideological differences and affect the full political spectrum.
Over 1,000 participants included roughly equal numbers of Democrats and Republicans, allowing researchers to test whether partisan identity moderated algorithmic influence. Results showed similar effect sizes for both groups—when exposed to more divisive content, members of both parties felt more negatively toward political opponents, and when divisive content was reduced, animosity decreased for both groups.
This bipartisan impact matters for several reasons. First, it suggests that algorithmic polarization isn’t merely reinforcing existing divisions but actively creating new hostility. If only one party were affected, polarization might be asymmetric, but when both sides experience similar increases in animosity, the overall political climate deteriorates for everyone.
Second, the finding complicates narratives about which party benefits from or suffers more from social media polarization. Rather than being a problem primarily affecting one side of the political spectrum, algorithmic amplification of divisive content appears to systematically worsen relations between citizens across ideological divides.
Third, bipartisan effects suggest potential for bipartisan solutions. If members of both parties are harmed by current algorithmic approaches, there may be political opportunities for reforms that reduce polarization. Platform changes that down-rank divisive content could benefit citizens across the political spectrum, potentially creating coalitions for healthier digital public squares.