SEMs with latent variables embody simultaneous equations with multiple exogenous and endogenous variables (path analysis), along with measurement error models (confirmatory factor analysis). Thus SEMs are the synthesis of methods developed in econometrics and psychometrics. This course provides an introduction to the theory, and empirical applications of SEMs within the "LISREL" framework.
The course will examine the five steps that characterize most applications of SEMs:
The course will also address estimation problems (improper solutions), and the problems involved in the analysis of non-normal. The concept of equivalent models will also be discussed.
The course will be of interest to those who plan to utilize the general structural equation model with latent variables or its specializations.
Familiarity with elementary matrix algebra will be useful, though not essential, for understanding LISREL syntax.