Analytic Methods for Network Data Ott, Miles Q creator 2014 16, 100 p. born digital Thesis (Ph.D. -- Brown University (2014) Hogan, Joseph Director Harrison, Matthew Reader Barnett, Nancy Reader Gile, Krista Reader Brown University. BIOMED: Biostatistics sponsor theses This dissertation presents methods for statistical analysis of social network data. First, we develop a Bayesian hierarchical model that calibrates self and peer-reports in order to synthesize information collected from multiple sources in the social network context. Secondly we demonstrate how current methods for analyzing fixed choice design social network data can induce biases. We propose a new survey design which collects information on the total number of relationships. Lastly, we investigate how assumptions in prevalence estimation in network sampling designs can produce biased estimates and introduce a new prevalence estimator for network sampling studies that offers considerable improvement over existing estimators. respondent-driven sampling Bayesian calibration augmented fixed choice design Social networks RPB 20141006 engEnglish10.7301/Z0NZ8604In CopyrightCollection is open for research.dissertations