Examining the Dynamic of Clustering Effects in Multilevel Designs: A Latent Variable Method Application

Abstract

This note is concerned with the study of temporal development in several indices reflecting clustering effects in multilevel designs that are frequently utilized in educational and behavioral research. A latent variable method-based approach is outlined, which can be used to point and interval estimate the growth or decline in important functions of level-specific variances in two-level and three-level settings. The procedure may also be employed for the purpose of examining stability over time in clustering effects. The method can be utilized with widely circulated latent variable modeling software, and is illustrated using empirical examples.