Advances in Multi-level Psychometric Models: Latent Variable Modeling of Growth With Missing Data and Multilevel Data

This paper describes three important methods areas of multivariate analysis which are not always thought of in terms of latent variable constructs, but for which latent variable modeling can be used to great advantage. These methods are: random coefficients describing individual differences in growth; unobserved variables corresponding to missing data; and variance components describing data from cluster sampling. The methods are illustrated using mathematics achievement data from the National Longitudinal Study of America Youth.