PSGA 7830:
Structural Equation Modeling with Lab (3 credits) Syllabus via Blackboard, eRes, or email.
The course introduces structural equation modeling (SEM) methodology with emphasis on its applications in behavioral and social sciences (psychology and education in particular). It provides a historical background for the development of SEM and gives an overview of model components (structural regressions and measurement models). It also addresses important caveats in using SEM and making causal interpretations. Specific topics include model representations and path diagrams, model identifications and model fitting (parameter estimation), model evaluations (significance of parameters and assessment of fit), and model refinement (modification indices and comparison of model fits). Path analysis, confirmatory factor analysis, and structural regressions with latent variables are discussed in details. LISREL will be used to practice applications of various models to analyzing real data. Lab work will provide practical experiences in using SEM software and focus on developing and fitting models as well as interpretations of results. Other topics including multiple-group modeling (model invariance), mean structure models, and latent change models (simple growth curves) will be introduced with empirical examples. Applications of measurement models in the context of classical test theory are also covered. These include concepts of measurement errors, reliability, and multitrait-multimethod (MTMM) validations.
