Research

Working Paper

Algorithmic Personalization and Digital Addiction: A Field Experiment on Douyin (TikTok)

(with Qianyun (Carole) Gao, Yu Gao, and Juanjuan Meng)

Abstract: We conducted a field experiment to investigate the impact of deactivating personalized recommendation algorithms or abstaining from Douyin (the Chinese version of TikTok) on users’ screen time and well-being. We found that while both treatments reduced Douyin use during the intervention, only the deactivation of personalized recommendations induced a lasting effect post-intervention. Further analysis indicates that deactivating personalized content bolstered cognitive competence, while abstaining from Douyin led to a significant decline in subjective well-being. In alignment with an adjusted habit formation model, these results highlight the effectiveness of disabling personalized algorithms as a strategy to combat digital addiction.