算法推荐如何塑造网络情绪扩散：机制、异质性与因果识别——以微博平台为例

How Algorithmic Recommendation Shapes Online Emotional Diffusion: Mechanisms, Heterogeneity, and Causal Identification—Evidence from Weibo

摘要: 在数字化时代，社交媒体平台的算法推荐系统已成为信息传播的核心驱动力。本文以微博平台为例，探讨算法推荐如何塑造网络情绪扩散的机制、异质性及因果识别问题。文章综述了情绪扩散的理论基础和社交媒体算法推荐的传播特点，分析了算法通过个性化过滤、参与度驱动和社群回声等机制影响情绪传播的路径与强度。我们关注不同情绪类型和用户群体在情绪扩散中的差异，并讨论在算法环境下辨析情绪传播因果效应的方法和挑战。通过微博平台案例，揭示算法推荐更易放大愤怒等高唤醒情绪、抑制悲伤等低唤醒情绪扩散的趋势，以及算法与用户互动在引发情绪共鸣、极化和传播中的作用。最后，文章讨论了研究发现对维护健康网络心理生态的启示，呼吁平台优化算法策略、用户提高媒介素养，并加强学界对算法影响网络情绪扩散的因果识别研究。

Abstract: In the digital age, algorithm-driven recommendation systems on social media platforms have become the principal engines of information dissemination. Using Weibo as a case study, this article examines how recommendation algorithms shape the diffusion of online emotions, with a focus on the underlying mechanisms, heterogeneity, and issues of causal identification. We first review theoretical foundations of emotional contagion and the distinctive features of algorithmic curation on social media. We then analyze how personalization filters, engagement-oriented ranking, and echo-chamber formation alter the pathways and intensity of emotional spread. Particular attention is paid to differences across emotion types and user groups. Methodological challenges and strategies for disentangling causal effects of algorithms on emotional diffusion are also discussed. Empirical evidence from Weibo reveals that algorithms tend to amplify high-arousal emotions such as anger while suppressing low arousal emotions such as sadness, and that algorithm–user interactions foster emotional resonance, polarization, and rapid propagation. Finally, we consider the implications for nurturing a healthy online affective ecology, calling for platform-level optimization of recommendation logic, enhanced media literacy among users, and deeper scholarly engagement with causal identification in studies of algorithmic influence on emotional diffusion.