Date(s) - Oct 25, 2010
2:00 PM - 3:00 PM
Center for Community Health
Assistant Professor, UCLA Department of Medicine Statistics Core; Senior Statistician, UCLA Semel Institute Center for Community Health
The concept of multi-level intervention is based on the assumption that individual behavior is interwoven with multiple layers; one person’s predictor may influence not only that person’s outcome measure, but also that person’s partners’ outcome measures. In this talk, I will briefly review the actor-partner interdependence models (APIMs) used in the study of family system when data from multiple family members are gathered. Next, we propose a Bayesian longitudinal APIM, which is a hierarchical, bivariate longitudinal model with actor’s and partner’s predictors, to estimate effects of interest as well as various levels of correlations. We illustrate our approach using data from HIV-affected families in China. *Joint work with Li Li.
Li-Jung Liang, Ph.D. (UCLA Biostatistics, 2005), is Assistant Professor at the Medicine Statistics Core of the UCLA School of Medicine and Senior Statistician at the UCLA Semel Institute Center for Community Health. Prior to her doctoral studies, she worked in the pharmaceutical and biotechnology industries for over 10 years. Her statistical research interests are hierarchical (multi-level) modeling, longitudinal data analysis, Bayesian methods, Markov model, and Markov chain Monte Carlo computation. She works with collaborators in several application areas, HIV/AIDS, bioinformatics, clinical trials, behavioral and social sciences, and nutrition.