Contextualizing Public Sentiment Analysis of Sexual Violence on Chinese Social Media: A Mixed-Method Study

Sexual violence remains a severe global public health burden, yet disclosure rates are particularly low in Chinese contexts due to cultural norms of family privacy and stigma. Social media provides a semi-anonymous channel through which survivors and the public express emotions that are rarely captured in clinical or survey research. This study aimed to characterize public emotional responses to sexual violence in China by analyzing large-scale social media discourse and to interpret the psychosocial mechanisms underlying these emotional patterns. A Mixed-Method Design was utilized to analyze N = 58,214 original user-generated posts collected from Weibo, Zhihu, and Xiaohongshu. The quantitative phase employed the DLUT Chinese Sentiment Lexicon to compute Emotion Intensity Scores, calculate polarity, and generate emotion co-occurrence matrices. Two independent coders achieved high inter-coder agreement (kappa = 0.87) to identify core themes explaining the statistical patterns . Negative emotions dominated the digital discourse (82.5%), with an overall polarity of -0.68. Disgust was the most intense (M = 6.8) and frequent (42.5%) emotion, empirically linked to “Bodily Contamination” and physiological rejection of the act. Anger (22.5%) manifested as “Moral Outrage,” primarily targeting perpetrators and systemic injustice. While Fear and Sadness were moderate individually, they co-occurred in 62% and 45% of survivor narratives, respectively, reflecting trauma-related helplessness . Public discourse on sexual violence is dominated by high-intensity Disgust and Anger, signaling profound physiological rejection and moral outrage. Critically, the prevalence of Fear in medical discussions identifies “Medical Shame” as a major barrier to care. These findings underscore the urgent need for mandatory Trauma-Informed Care protocols and comprehensive consent education to reduce retraumatization and clarify social boundaries.