Modeling High-Dimensional Longitudinal Data with Structural Equation Modeling

• Activity Number: 308
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Xinming An and Yiu-Fai Yung*+ and Qing Yang
• Companies: SAS Institute and SAS Institute and UCLA
• Keywords: high dimensional longitudinal data ; structural equation modeling

Abstract

Longitudinal studies involving latent variables such as depression that cannot be observed directly are often encountered in biomedical and social science research. Multiple responses are used to characterize these latent quantities, and repeated measures are collect to capture their trends over time. As a result, high dimensional longitudinal data are collected. Substantive research questions may concern issues as interrelated trends among latent variables that can only be addressed by modeling the high dimensional longitudinal data jointly. While statistical analysis of univariate longitudinal data has been well developed, methods for multivariate high dimensional longitudinal data are still under development. Structural equation modeling (SEM) has been widely used in social science to model longitudinal data. It can be easily extended to handle high dimensional longitudinal data and enjoy several appealing modeling abilities. However, its usage has been very limited in other fields. In this paper, we will illustrate how to model high dimensional longitudinal data using SEM with different features, for example unbalanced data, time-varying covariate and latent covariate.