Title: Robust semiparametric Bayesian methods in growth curve modeling with nonnormal missing data
Authors: Xin Tong - University of Virginia (United States) [presenting]
Abstract: Despite the wide application of growth curve models in health research, few studies have dealt with two practical issues of longitudinal data analysis -- nonnormality of data and missing data. A semiparametric Bayesian approach is proposed for growth curve modeling, in which intraindividual measurement errors follow unknown distributions with Dirichlet process priors. To deal with missing data, a multiple imputation technique is applied for missing completely at random and missing at random data. A Monte Carlo simulation study is conducted to evaluate the proposed method and compare it to the traditional growth curve modeling. A real data analysis in health research is also provided to illustrate the application of the semiparametric growth curve modeling.