Extension of the cross-classified multiple membership growth curve model for longitudinal data

Abstract

Student mobility is a common phenomenon in longitudinal data in educational research. The characteristics of education longitudinal data create a problem for the conventional multilevel model. Grady and Beretvas (2010) introduced a cross-classified multiple membership growth curve (CCMM-GCM) model to handle Student mobility over time by capturing complex higher level clustering structure in the data. There are some limitations in the CCMM-GCM model. By creating dummy coded indicators for each measurement occasion, the new model can improve the accuracy and provides an easier and more flexible structure at the higher level. This study provides some support that the new model better fits a dataset than the CCMM-GCM model