ABSTRACT
In the first part of the chapter, we extend our basic latent two-level latent variable model developed in the last chapter to examine relationships between covariates with random slopes. The second part of the chapter extends the use of latent constructs at two levels in the investigation of structural parameters between latent variables. We extend the examination of “error free” structural parameters, particularly between groups, which is a type of model that is well suited to SEM. Our goal is to illustrate how the SEM approach is ideal for modeling multivariate outcomes and for examining separate sets of structural relationships within and between groups involving direct effects, indirect effects, and reciprocal effects between variables.
