群体情绪感知的偏差机制：基于注意与线索整合的综述

Mechanisms of Bias in Group Emotion Perception: A Review Focusing on Attention and Cue Integration

摘要: 快速感知群体情绪对适应复杂社会互动至关重要。研究发现，个体通过“集群编码”机制高效提取群体情绪的统计特征，但该过程存在“情绪放大效应”的系统性偏差。本文系统梳理了该领域研究进展。首先，集群编码的高效性与注意资源的有限性构成了理解群体情绪感知的基础框架。其次，放大效应作为核心偏差现象，其机制存在一般性注意偏向与特异性威胁驱动的理论争议，并受到情绪类别、暴露时间等因素调节。进一步，研究不仅在加工机制层面揭示了注意在信息采样与决策整合阶段可能存在的分离，同时也在研究视角上，从单一的面孔信息拓展至对身体姿势、语音等多线索整合机制的探索。最后，本文总结了现有研究在生态效度、理论纵深及方法整合方面的主要局限，并展望了结合计算建模、神经科学与跨文化比较等多元路径的未来方向。

Abstract: Rapid perception of group emotions is crucial for adapting to complex social interactions. Research indicates that individuals efficiently extract statistical features of group emotions through an “ensemble coding” mechanism, yet this process exhibits a systematic bias known as the “emotional amplification effect.” This article systematically reviews recent advances in this field. First, the efficiency of ensemble coding and the limited nature of attentional resources constitute a foundational framework for understanding group emotion perception. Second, the amplification effect, as a core perceptual bias, is subject to theoretical debate between general attentional bias and threat-specific drivers, and is modulated by factors such as emotion category and exposure duration. Furthermore, studies have not only revealed a potential dissociation between attentional sampling and decision integration in the processing mechanism, but have also expanded the research perspective from facial information alone to the integration of multiple cues such as body posture and vocal tone. Finally, this article summarizes the main limitations of existing research in terms of ecological validity, theoretical depth, and methodological integration, and suggests future directions incorporating computational modeling, neuroscience, and cross-cultural comparisons.