Title: Bayesian two-level model for partially ordered repeated responses
Authors: Xiaoqing Wang - The Chinese University of Hong Kong (Hong Kong) [presenting]
Xiangnan Feng - The Chinese University of Hong Kong (Hong Kong)
Xinyuan Song - Chinese University of Hong Kong (Hong Kong)
Abstract: Owing to the questionnaire design and problem nature, partially ordered data that are neither completely ordered nor completely unordered are frequently encountered in behavioral and medical researches. However, among literature, little attention has focused on longitudinal observations with partially ordered data structure. We propose a Bayesian two-level regression model for analyzing longitudinal data with partially ordered responses. The first-level model is defined for partially ordered observations taken at each time point nested within individual and the second-level model is defined for individuals to assess the effect of their characteristics on longitudinal responses. A full Bayesian approach with Markov Chain Monte Carlo (MCMC) algorithm is developed for statistical inference. A simulation study demonstrates that the developed methodology performs satisfactorily. An application to a longitudinal study concerning adolescent substance use is presented.