Course Description

Structural equation modeling (SEM) is a very general framework for specifying and evaluating parametric directional and non-directional relationships among variables. There can be any number of independent and dependent variables as well as hypothetical latent variables. Using latent variables allows estimation of relationships that are not affected by measurement error. Specific types of SEM include multiple regression, path analysis, confirmatory factor analysis, and growth curve models, among others.

This Short Course provides an overview of the basic concepts of SEM, with a particular focus on confirmatory factor analysis, and then moving to the general model for structural relations among latent variables. Each session will incorporate work in the computer lab, when participants have an opportunity to apply the material covered in the lecture. These lab exercises will be presented for SAS software (i.e., proc calis) and R (sem, lavaan), but with the main emphasis on R.

Because SEM is essentially a framework for specifying and estimating regression models, it will be expected that course participants have a strong background in multiple linear regression analysis. Specific topics are as follows:

  1. From EFA to CFA: basic ideas
  2. Specification and identification of confirmatory factor analysis (CFA) models;
  3. Special CFA models: Higher order models, hierarchical models, and multi-trait multi-method models; and
  4. The full SEM: Models with structural equations among latent variables.

Copyright © 2019 Michael Friendly. All rights reserved.

friendly AT yorku DOT ca

                  ORCID iD