This dissertation examines the networks of supportive relationships for individuals suffering from alcoholism, and the how these networks may have changed in response to the availability of online venues for social support exchange. To do this, the research combined two theoretical perspectives from social support and social network theory, namely the Convoy Model of Social Relations and the Strength of Weak Ties. Together, these two theories provide a multilevel framework for systematizing research into complex social phenomena. The benefit of these two perspectives, when combined, is an ability to evaluate both structural and psychosocial aspects of the topic under study. Support groups are theorized to exhibit properties that would not be predicted by either theory alone, creating a special theoretical case. Differences are also theorized to exist between participants of online and face-to-face support groups on the basis of both psychological and health-related indicators.

This framework was utilized to conduct two studies. First a pilot study was conducted to validate a new research instrument utilizing undergraduate students (N = 50) at UC Santa Barbara. The main study utilized a mixed method approach to examine both traditional face-to-face and computer-mediated support group participants (N = 413). Results from a cross-sectional survey validate many aspects of the combined convoy-SWT framework, and show significant limitations of online support group participation. Further, 20 follow-up semi-structured interviews were conducted to gain additional contextual insight. These indicated that participants encountered both helpful and detrimental aspects of familial support, longitudinal effects in support network structure, stigmatization from AA participation outside group, and perception of different benefits and uses of online versus face-to-face support groups. The dissertation concludes with a discussion of how the convoy-SWT framework could be utilized to investigate additional complex social phenomena using longitudinal and sociometric approaches.