ABSTRACT
The purpose of this chapter is to introduce some recent trends in Latent Growth Curve Models (LGM). Basic features of Structural Equation Modeling (SEM) are used to describe these models. To illustrate this type of longitudinal data analysis, a small sample of data from the National Institutes of Health (NIH) are used. These data include the budgets of 20 NIH Institutes and up to 6 occasions of measurement (1992 to 1998). These data are analyzed using maximum-likelihood estimation (MLE) available with the LISREL-8 (Jöreskog & Sörbom, 1993) and Mx (Neale, 1993) computer programs. In a first analysis, standard autoregressive changes models are described and fitted. In a second analysis, standard LGMs are fitted to the same data using average cross-products (i.e., moments). In a third analysis, these same data are then explored using raw data and fitted with unbalanced pedigree techniques (Neale, 1993). More complex dynamic model extensions are also described.
