Levonian, Zachary2023-01-042023-01-042022-07https://hdl.handle.net/11299/250412University of Minnesota Ph.D. dissertation. 2022. Major: Computer Science. Advisors: Loren Terveen, Svetlana Yarosh. 1 computer file (PDF); 242 pages.When a person has a health crisis, the availability of social support affects both their physical and mental health. Online communities can make support available by providing a place to connect with peers who have had similar experiences. However, finding relevant peers to talk to and learn from is challenging. Algorithmic systems for peer matching could help people find relevant peers, but designing such systems requires an understanding of how people use online communities for support—when, how, and to whom they connect. I collaborated with a large existing online community—CaringBridge.org—to understand how patients experiencing a health crisis and their non-professional caregivers use CaringBridge to seek and receive support. Based on this understanding, I created a recommendation system to facilitate peer connections on CaringBridge. CaringBridge users of my system received email recommendations for peer users they may wish to connect with. By collecting survey and usage feedback, I advance an understanding of when support seekers and providers connect with potentially-supportive peers. Taken together, my work describes quantitatively and qualitatively the use of health-related online communities for receiving and providing social support. My work has implications for the deployment of peer-matching systems that facilitate supportive communication.enhuman-computer interactiononline communitiesrecommendationsocial computingsocial supportThesis or Dissertation