I was wondering if anyone could provide feedback on an analytical question. I have longitudinal sociocentric network data of a college class. At each timepoint (I have network data at 2 timepoints), participants nominate up to 10 of their friends. I am considering using latent profile analysis to identify latent groups of people based on their changes in alcohol use behavior using this data. Specifically, I plan to use average number of drinks, maximum number of drinks, and number of heavy drinking episodes at both timepoints to identify latent groups. Is it possible to conduct a latent profile analysis with social network data? Specifically, is it possible to account for the interdependence in the data?
Hannah Doucette, PhD
T32 NIAAA Postdoctoral Fellow
Center for Alcohol and Addiction Studies
Department of Behavioral and Social Sciences
Brown University School of Public Health