Exploring public discourse and sentiment toward aging on Chinese social media: Machine learning analysis of Weibo and Douyin

The rapid evolution of information and communication technologies has transformed online environments, offering new platforms for interaction and shaping social identities. This study investigates public perceptions of older adults on two major Chinese social media platforms, Weibo and Douyin, within the context of Social Identity Theory, which posits that identities are dynamically formed through social interactions. Despite older adults' challenges in navigating digital spaces, which often exacerbate negative stereotypes, emerging platforms also present opportunities to highlight their diverse and positive attributes. The study addresses a research gap by examining both pandemic and post-pandemic periods, a time marked by increased internet use among older adults and evolving public attitudes. Prior research has focused on single platforms or short time frames, leaving a need for a comprehensive analysis across multiple platforms and time periods. This study collects a substantial dataset—844,966 posts from Weibo and 53,996 from Douyin—using key terms related to aging. Methodologies include web scraping and trending video analysis to capture a broad spectrum of content and user engagement. Employing machine learning techniques such as Latent Dirichlet Allocation (LDA) for thematic and sentiment analysis, this study explores how discourse and sentiment towards older adults have shifted over time. The findings reveal a complex portrayal of older adults, with both reinforcing stereotypes and positive representations evident. This longitudinal approach provides insights into whether the social identity of older adults has improved, worsened, or remained static, highlighting emerging positive perceptions within public discourse. The study’s results underscore the importance of ongoing, multidimensional analyses to fully understand the impact of social media on the representation of older adults.