Online health community for change: Analysis of self-disclosure and social networks of users with depression.

Abstract

A key research question with theoretical and practical implications is to investigate the various conditions by which social network sites (SNS) may either enhance or interfere with mental well-being, given the omnipresence of SNS and their dual effects on well-being. We study SNS' effects on well-being by accounting for users' personal (i.e., self-disclosure) and situational (i.e., social networks) attributes, using a mixed design of content analysis and social network analysis. We compare users' within-person changes in self-disclosure and social networks in two phases (over half a year), drawing on Weibo Depression SuperTalk, an online community for depression, and find: ① Several network attributes strengthen social support, including network connectivity, global efficiency, degree centralization, hubs of communities, and reciprocal interactions. ② Users' self-disclosure attributes reflect positive changes in mental well-being and increased attachment to the community. ③ Correlations exist between users' topological and self-disclosure attributes. ④ A Poisson regression model extracts self-disclosure attributes that may affect users' received social support, including the writing length, number of active days, informal words, adverbs, negative emotion words, biological process words, and first-person singular forms. We combine social network analysis with content analysis, highlighting the need to understand SNS' effects on well-being by accounting for users' self-disclosure (content) and communication partners (social networks). Authentic user data helps to avoid recall bias commonly found in self-reported data. A longitudinal within-person analysis of SNS' effects on well-being is helpful for policymakers in public health intervention, community managers for group organizations, and users in online community engagement.