Sickness and Health: Homophily in Online Health Forums

This work explores the link between health and social relations by creating an automated metric of similarity of positive or negative affect (sentiment) between peers in online health forums. We analyze textual communication between peers and demonstrate that those who communicate often have similar average sentiment scores. Sentiment is the author’s immediate affective state, their positive or negative orientation. We hypothesize that average sentiment over time indicates overall happiness or sadness as similar analysis has been utilized to identify depression and depression at risk college students [19]. These results follow the analysis of Framingham study data demonstrating that happy people tend to associate with one another and that happiness spreads within social networks [4].