Introduction to Structural Equation Modeling

A 3-Day Remote Seminar Taught by
Paul D. Allison, Ph.D.

Read reviews of this seminar

To see a sample of the course materials, click here.

Structural Equation Modeling (SEM) is a statistical methodology that is widely used by researchers in the social, behavioral and educational sciences. First introduced in the 1970s, SEM is a marriage of psychometrics and econometrics. On the psychometric side, SEM allows for latent variables with multiple indicators. On the econometric side, SEM allows for multiple equations, possibly with feedback loops. In today’s SEM software, the models are so general that they encompass most of the statistical methods that are currently used in the social and behavioral sciences.

Here Are a Few Things You Can Do With Structural Equation Modeling

  • Test the implications of causal theories.
  • Estimate simultaneous equations with reciprocal effects.
  • Incorporate latent variables with multiple indicators.
  • Investigate mediation and moderation in a systematic way.
  • Handle missing data by maximum likelihood (better than
    multiple imputation).
  • Adjust for measurement error in predictor variables.
  • Estimate and compare models across multiple groups of individuals.
  • Represent causal theories with rigorous diagrams.
  • Investigate the properties of multiple-item scales.

Because SEM is such a complex and wide-ranging methodology, learning how to use it can take a substantial investment of time and effort. Now, you have the opportunity to learn the basics of SEM from a master teacher, Professor Paul D. Allison, in just three days.

Starting November 4, we are offering this seminar as a 3-day synchronous*, remote workshop. Each day will consist of a 4-hour live lecture held via the free video-conferencing software Zoom. You are encouraged to join the lecture live, but will have the opportunity to view the recorded session later in the day if you are unable to attend at the scheduled time.

Each lecture session will conclude with a hands-on exercise reviewing the content covered, to be completed on your own. An additional lab session will be held Thursday and Friday afternoons, where you can review the exercise results with the instructor and ask any questions.

*We understand that scheduling is difficult during this unpredictable time. If you prefer, you may take all or part of the course asynchronously. The video recordings will be made available within 24 hours of each session and will be accessible for four weeks after the seminar, meaning that you will get all of the class content and discussions even if you cannot participate synchronously.

Closed captioning is available for all live and recorded sessions.


The empirical examples and exercises in this course will emphasize Mplus, but equivalent code will be presented for SAS, Stata and lavaan (a package for R). Mplus is one of the best SEM packages because of its superior capabilities for missing data, multi-level modeling, and ordinal and categorical data.

To fully benefit from the course, you should have your own computer loaded with a recent version of SAS, Stata, Mplus or R (with the lavaan package installed). Whichever package you choose, you should already be familiar with basic data management operations and the commands/procedures for doing linear regression, logistic regression, etc.


This course is designed for researchers with a moderate statistical background who want to apply SEM methods in their own research projects. No previous background in SEM is necessary. But participants should have a good working knowledge of basic principles of statistical inference (e.g., standard errors, hypothesis tests, confidence intervals), and should also have a good understanding of the basic theory and practice of linear regression. 

Course Outline

1. Introduction to SEM
2. Path analysis
3. Direct and indirect effects
4. Identification problem in nonrecursive models
5. Reliability and validity
6. Multiple indicators of latent variables
7. Exploratory factor analysis
8. Confirmatory factor analysis
9. Goodness of fit measures
10. Structural relations among latent variables
11. Alternative estimation methods.
12. Multiple group analysis
13. Models for ordinal and nominal data

reviews of introduction to structural equation modeling

“I left this course feeling like I had a good grasp on these methods for the first time, and feeling empowered to apply them in a practical setting with real-world data.”
  Lauren Wilson, Duke University

“SEM has come up a lot in my current line of work at University of Michigan. Since I had never taken a formal training on this, I was very unclear on the concepts. This course emphasizes simple examples to explain new concepts. It develops concepts very systematically. At the end of this course, I have a very good idea about path analysis, latent factor models, and confirmatory factor models. I feel confident that I can draw a conceptual diagram of a theoretical model and implement it in a statistical program. Thank you so much!”
  Bidisha Ghosh, University of Michigan

“This course provided a great conceptual and technical introduction to SEM. The instructor is an expert in SEM and it showed! Plus he provided clear explanations and showed great patience in answering all our questions. I feel I can now go back to my job and apply what I’ve learned and extend my SEM skills, now that I have a solid foundation.”
  Elizabeth Tarlov, Edward Hines, Jr. VA Hospital

“This course is immediately applicable to my own work and work with students. Dr. Allison is available for detailed explanations. There was a clear integration of theoretical and statistical concepts.”
  Paul Harrell, Eastern Virginia Medical School

“I learned a lot about SEM on principle and examples which will guide me to solve real problems in my work.”
  You Li, University of Massachusetts

“This was a great course and covered a number of topics in depth. I signed up to learn more about how SEMs could be used to test mediation models but the section on FIML was interesting and I will try it out on my next missing data problem.”
  Laura Pyle, University of Colorado