Algorithmic personalization on social media carries a dual psychological impact: it may alleviate cognitive burden while simultaneously provoking autonomy-related concerns. This study develops and empirically tests a dual-pathway reactance framework explaining how algorithm-induced psychological strain contributes to social media discontinuance. Integrating the Stressor–Strain–Outcome perspective with Psychological Reactance Theory, we conceptualize information overload and uniqueness neglect as parallel psychological experiences that increase psychological reactance and, in turn, discontinuance intention. Using survey data from X users and structural equation modeling, we find that personalization significantly reduces uniqueness neglect but does not significantly affect information overload. Both uniqueness neglect and information overload increase psychological reactance, with identity-related perceptions exhibiting a comparatively stronger association. Reactance strongly predicts discontinuance intention, confirming its central motivational role. Interaction analysis reveals that algorithm awareness moderates the cognitive pathway: personalization reduces perceived overload among users with higher levels of algorithm awareness, while reductions in uniqueness neglect remain stable across awareness levels. These findings advance theory by clarifying the asymmetric psychological consequences of personalization and positioning psychological reactance as a key mechanism underlying social media discontinuance.
抄録全体を表示