Structural Equation Modelling using Mplus

Course Content:
This course introduces Structural Equation Modelling (SEM) using lectures, examples and lab sessions.  The software for examples is Mplus which is widely regarded as one of the most advanced and accessible packages for estimating SEM.  Topics covered include getting to know Mplus by running standard regression models, extending regression into path analysis by including mediating variables, and adding latent variables to produce full structural equation models.  Models for both continuous and categorical data will be covered.

Course objectives

  • To introduce participants to the Mplus interface , and teach them how to transfer data from other software packages into Mplus
  • To learn how to run a simple regression model in Mplus and how to interpret the output and understand issues of heteroscedasticity and centring
  • To understand how to run models with more than one dependent variable and use the model to tese hypotheses 
  • To introduce participants to logit and probit models for categorical data, interpreting log odds ratios and latent responses and thresholds.
  • To introduce path analysis and the fitting of mediation models for continuous and categorical variables
  • To understand the rationale of latent variable models and to fit Confirmatory Factor Analysis for continuous data
  • To introduce Rasch modelling and Item response Theory, modelling approaches analogous to Confirmatory Factor Analysis but for categorical data
  • To cover basic model building issues in SEM, and to understand important related concepts such as measurement vs. structural models, identification issues, and diagnostic testing

Topics covered in this course include:

  • What is Structural Equation Modelling?
  • Mplus
    • What is Mplus?
    • Demo version
    • Getting to know Mplus
  • Regression Models
    • Regression Assumptions
    • Standardised regression results in Mplus
    • Heteroscedasticity
  • Models for Binary Categorical Data
    • Binary outcomes
    • Logit regression
    • Probit regression
  • Estimators
    • Maximum likelihood estimator (ML)
    • Weighted least squares estimator (WLS)
  • Path Analysis
    • Causality
    • Multiple equations
    • Direct and indirect effects
    • Competing mediators
    • Mediation, moderation and confounding
    • Path model assumptions
  • Confirmatory Factor Analysis
    • Unmeasured causes
    • Factor models
    • Identification
    • Missing data
    • Confirmatory analysis model input file
    • Model fit
    • Approximate fit
    • Comparative fit
    • Summary of model fit indices
    • Multi-factor solutions
    • Likelihood ratio test
    • Crossloadings
  • Item Response Theory (IRT) Models
    • Item characteristic curves (ICCs)
  • Structural Equation Models
    • Examples of structural equation models
    • Recursivity 
    • Modification indices
    • Model building strategy 
    • Equivalent models 
    • Evaluating models
    • Sample size
  • Assessing Measurement Invariance
    • Multi-group models
    • Scale factors
    • Difftest

Request the training materials

View all courses available