TY - JOUR
T1 - Write for life
T2 - Persisting in online health communities with expressive writing and social support
AU - Ma, Haiwei
AU - Smith, C. Estelle
AU - He, Lu
AU - Narayanan, Saumik
AU - Giaquinto, Robert A.
AU - Evans, Roni
AU - Hanson, Linda
AU - Yarosh, Svetlana
PY - 2017/11
Y1 - 2017/11
N2 - Expressive writing has been shown to improve physical, mental, and social health outcomes for patients struggling with difficult diagnoses. In many online health communities, writing comprises a substantial portion of the user experience, yet little work has explored how writing itself affects user engagement. This paper explores user engagement on CaringBridge, a prominent online community for writing about personal health journeys. We build a survival analysis model, defining a new set of variables that operationalize expressive writing, and comparing these effects to those of social support, which are well-known to benefit user engagement. Furthermore, we use machine learning methods to estimate that approximately one third of community members who self-identify with a cancer condition cease engagement due to literal death. Finally, we provide quantitative evidence that: (1) receiving support, expressive writing, and giving support, in decreasing magnitude of relative impact, are associated with user engagement on CaringBridge, and (2) that considering deceased sites separately in our analysis significantly shifts our interpretations of user behavior.
AB - Expressive writing has been shown to improve physical, mental, and social health outcomes for patients struggling with difficult diagnoses. In many online health communities, writing comprises a substantial portion of the user experience, yet little work has explored how writing itself affects user engagement. This paper explores user engagement on CaringBridge, a prominent online community for writing about personal health journeys. We build a survival analysis model, defining a new set of variables that operationalize expressive writing, and comparing these effects to those of social support, which are well-known to benefit user engagement. Furthermore, we use machine learning methods to estimate that approximately one third of community members who self-identify with a cancer condition cease engagement due to literal death. Finally, we provide quantitative evidence that: (1) receiving support, expressive writing, and giving support, in decreasing magnitude of relative impact, are associated with user engagement on CaringBridge, and (2) that considering deceased sites separately in our analysis significantly shifts our interpretations of user behavior.
KW - CaringBridge
KW - Death
KW - Expressive writing
KW - Online health community
KW - Social support
KW - User engagement
UR - http://www.scopus.com/inward/record.url?scp=85050615020&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85050615020&partnerID=8YFLogxK
U2 - 10.1145/3134708
DO - 10.1145/3134708
M3 - Article
AN - SCOPUS:85050615020
SN - 2573-0142
VL - 1
JO - Proceedings of the ACM on Human-Computer Interaction
JF - Proceedings of the ACM on Human-Computer Interaction
IS - CSCW
M1 - 73
ER -