Abstract
INTRODUCTION: The present study provides novel insights into the psychological mechanisms linking social emotions and eating behaviors by integrating large-scale social media analysis with individual-level assessments. METHODS: Two complementary approaches were employed: Study 1 analyzed 1,902 Weibo posts containing "diet" and "social" keywords through latent Dirichlet allocation topic modeling and semantic network analysis to identify thematic structures and interactional patterns; Study 2 surveyed 1,199 participants (aged 18-33) using the Intuitive Eating Scale and self-reported social situation texts, applying Linguistic Inquiry and Word Count (LIWC) and LASSO regression to identify language features of intuitive eating. RESULTS: Study 1 revealed six psychological themes and a semantic framework connecting social-dietary interactions, health discourse, emotional states, and body image concerns, while Study 2 demonstrated that negatively valenced words (e.g., sensitive, tiring) were associated with lower intuitive eating, whereas positively valenced words (e.g., relaxed, positive) were associated with healthier eating patterns; moreover, negative emotion scores in social texts showed significant correlations with poorer intuitive eating (p < 0.05). DISCUSSION/CONCLUSION: These findings illustrate associations between social emotional expression and eating behaviors, highlighting implications for emotion-sensitive interventions and the design of healthier online social environments.