Engagement, User Satisfaction, and the Amplification of Divisive Content on Social Media

In a pre-registered randomized experiment, we found that, relative to a reverse-chronological baseline, Twitter's engagement-based ranking algorithm may amplify emotionally charged, out-group hostile content and contribute to affective polarization. Furthermore, we critically examine the claim that the algorithm shows users what they want to see, discovering that users do *not* prefer the political tweets selected by the algorithm. Finally, we explore the implications of an alternative approach to ranking content based on users' stated preferences and find a reduction in angry, partisan, and out-group hostile content but also a potential reinforcement of echo chambers. The evidence underscores the necessity for a more nuanced approach to content ranking that balances engagement, users' stated preferences, and sociopolitical outcomes.