A multi-site study integrating direct exposure logs (idealized content density, influencer ratio, beauty filter prevalence) with surveys in India, China, and one Western context. Use piecewise or spline models and moderated mediation analyses to assess whether social comparison kicks in only past certain exposure thresholds and among high social comparison orientation teens. This approach tests thresholded and context-dependent mediation processes rather than assuming linear, always-on mediation. It incorporates cultural and gendered norms as contextual moderators. The study explains why some prior research finds direct effects without mediation and others find comparison-driven pathways. It aligns with algorithm design concepts (exploration/exploitation), suggesting feeds may keep teens below comparison thresholds unless prior behavior pushes them into an “escalation zone.” The impact is a policy-relevant “safe exposure” framework specifying thresholds for idealized content density or influencer ratio beyond which comparison harm accelerates, guiding age-aware ranker guardrails.
References:
If you are inspired by this idea, you can reach out to the authors for collaboration or cite it:
@misc{gpt-5-when-does-social-2025,
author = {GPT-5},
title = {When Does Social Comparison Mediate? A Thresholded, Contextual Mediation Model Reconciling Conflicting Findings},
year = {2025},
url = {https://hypogenic.ai/ideahub/idea/3ABiqdA9rLzagZWDWKvt}
}Please sign in to comment on this idea.
No comments yet. Be the first to share your thoughts!