Latent Class-Profile Analysis for Modeling Stage-Sequential Processes of Drug-Taking Behaviors

• Activity Number: 380
• Type: Contributed
• Date/Time: Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Hwan Chung*+
• Companies: Michigan State University
• Keywords: Latent class analysis ; Latent stage-sequential analysis ; Longitudinal data ; Mixture model ; Multivariate categorical data

Abstract

This study proposes a mixture modeling approach for longitudinal data, which will be referred as Latent Class-Profile Analysis (LCPA). LCPA provides a set of principles for systematic identification of two-level homogeneous subgroups of individuals: at the first level, LCPA identifies discrete classes of individuals who have similar responses to the items at each occasion (latent class); at the second level, LCPA examines individuals' class membership over the entire set of time points so as to classify the population into two or more subgroups based on class sequencing (latent class-profile). Applied to the longitudinal study of drug use, all individuals in a second-level subgroup are specified to have a similar sequential pattern of drug-taking behavior over time. The use of the proposed model is illustrated with observations from the National Longitudinal Survey of Youth 1997.