Social Media Algorithm Recommendations and Social Connection Among Digital Natives: The Moderating Role of Information Cocoon

Abstract

This study explores the relationship between social media algorithm recommendation types (personalized interest vs. diverse social recommendations) and digital natives’ (18–25 years old) social connections (emotional vs. behavioral), as well as the moderating role of information cocoons. Adopting a mixed-methods design, it conducted a cross-sectional survey (N=2,156) and semi-structured interviews (N=45) with participants from five countries (China, Germany, India, Spain, Ghana). Survey results showed diverse social recommendations positively predicted both emotional and behavioral connections, while personalized interest recommendations had a non-significant positive effect on emotional connection and a weak positive effect on behavioral connection. Information cocoons significantly moderated the link between diverse social recommendations and social connections (weakening the positive association), but their negative moderating effect on personalized interest recommendations was insignificant. Interview findings further indicated diverse recommendations boosted cross-group interactions, whereas personalized ones narrowed information horizons; active information-seeking could mitigate information cocoons’ negative impacts. These findings deepen theoretical insights into algorithmic media’s effects on social development and guide rational algorithm use.

Previous article in issue Next article in issue