Internet altruistic behaviors and affective well-being among family caregivers: a natural language processing approach

OBJECTIVES: This study investigates the typology of Internet al.ruistic behaviors (IABs) and their associations with affective well-being at the provider, recipient, and group levels within a large online community of family caregivers of people living with dementia in China. METHODS: We employed natural language processing (NLP) techniques to analyze 25,024 chat messages from the largest WeChat-based support group for caregivers of people living with dementia in Shanghai, collected between June 1, 2022, and June 1, 2024. The RoBERTa model was used to identify the typology of IABs of every text, and sentiment analysis was applied to quantify the affective well-being. Finally, linear mixed-effects models were used to test the associations between IABs and affective well-being. RESULTS: We identified 3 categories of IABs: Internet support, Internet guidance & reminder, and Internet sharing. IABs were found to enhance affective well-being at all 3 levels. Specifically, Internet support and Internet guidance & reminder were significantly associated with improved recipients' affective well-being, whereas all 3 IAB types were positively linked to the provider's affective well-being. At the group level, higher intensity of IABs predicted increased group affective well-being. DISCUSSION: By extending the NLP approach to analyzing interactions among family caregivers of people living with dementia, this study contributes to a deeper understanding of how digital altruism is associated with affective well-being. Findings support the design of IAB-enhancing mechanisms that facilitate emotional support, guidance, and knowledge-sharing for family caregivers.