“Envy in the Feed”: Pattern-Specific Social-Media Predictors of Depression and Anxiety among Chinese College Students
X. Fan, X. Jin, T. Lyu, X. Hou, W. Guo, S. WangIntroduction
Depression and anxiety are prevalent mental health concerns among Chinese college students. Algorithm-curated “highlight reels” on WeChat, Weibo and TikTok amplify social comparison and fear of missing out (FOMO), both associated with heightened risk of depression and anxiety. But previous work rarely tested whether specific use patterns predict mood symptoms while adjusting for education level or social support.
Objectives
To examine the association between social media use and symptoms of depression and anxiety, and to derive parsimonious, pattern-specific algorithms that quantify the risk of depression and anxiety from social-media behaviors.
Methods
Nationwide cross-sectional online survey (March 2024–May 2025) of 1,030 students (16–30 y) recruited through student-union mailing lists in China; 93.6 % response rate; 67.9 % women; mean age 20.8 ± 2.3 y. Measures: Depression & Anxiety: Zung Self-rating Depression Scale (SDS, Chinese cut-off: ≥53) and Self-rating Anxiety Scale (SAS, ≥50). Social-media inventory: Social Network Site Intensity Scale (SNSIS, 6–30), Active/Passive Use Sub-scales (5-15), Social Media Envy Scale (2–10), Bergen Social-Media Addiction Scale (6-30), and self-reported daily duration (h). Covariates: Age, gender (0=male, 1= female), educational level (1=college, 2=bachelor, 3=master, 4=doctoral), and social support rating scale (SSRS).
Results
Overall, 51.4% of participants screened positive for depression (11.3% moderate, 5.6% severe) and 33.1% for anxiety (8.3% moderate, 7.7% severe). Median daily social-media exposure was 6 h (IQR 5–10). Backward-LR logistic models identified a 4-item composite (SNS intensity, passive use, education, gender) that predicted depression with AUC = 0.67, and a 3-item composite (envy tendency, addiction, social support) that predicted anxiety with AUC = 0.78; both AUCs were superior to any single social-media metric (DeLong’s test, all p < 0.05,
Figure 1-2
). Final prediction equations (Nagelkerke R²): Depression: logit(p) = −3.41 + 0.051 SNS-intensity + 0.131 passive-use + 0.71 education + 0.25 gender; R² = 0.121. Anxiety: logit(p) = −2.81 + 0.116 envy/FOMO + 0.164 addiction − 0.036 social-support; R² = 0.285. Structural equation modeling further showed that FOMO mediated 41 % of the effect of SNS intensity on depression and 52 % on anxiety (standardized paths: a = 0.50, b = 0.16 and 0.31, respectively), indicating that envy in the feed is a key explanatory pathway (
Figure 3
).
Image 1: Long description.