Short-Video Platforms’ Impact on Group Psychology: A Computational Social Science Analysis

Abstract

This study employs causal inference methods to analyze user behavior patterns on short-video platforms, investigating how content characteristics, algorithmic recommendations, and social network structures influence user engagement and group psychology. Using techniques such as Propensity Score Matching (PSM), Regression Discontinuity Design (RDD), and Instrumental Variables (IV), this study establishes robust causal relationships. The findings indicate that algorithmic promotion plays a crucial role in content diffusion, significantly increasing user engagement. Emotionally charged content is more likely to be shared, with highly positive or negative videos being 2.5 times more likely to go viral than neutral content. Key Opinion Leaders (KOLs) have a significant impact on information diffusion, with influencer interactions increasing content visibility by 80%. Furthermore, distinct behavioral patterns among different user groups suggest that platform algorithms not only shape information flow but also influence emotional resonance and opinion convergence within digital communities. The key innovation of this study lies in the application of causal inference techniques to uncover real behavioral mechanisms on short-video platforms, avoiding the limitations of traditional correlation-based analysis. The findings provide valuable practical implications for social media marketing, public opinion management, and digital content governance. Policy recommendations based on this study include improving platform transparency, optimizing content recommendation diversity, and assessing algorithmic fairness. Future research directions include cross-platform data integration, dynamic causal tracking, and cross-cultural comparative studies to further enhance the understanding of digital media ecosystems.