Using Structural Equation Models to Uncover Relationships between Observed and Latent Variables

See how to:

• Understand the basics of SEM
  • Confirmatory Factor Analysis (CFA) – a multivariate statistical procedure to test how well measured variables represent the number of constructs and confirm or reject a theory
  • Path Analysis - an extension of the regression model built from the correlation matrix that compares two or more casual models and uses a diagram flow chart to model path and causation
  • Structural Equation Modeling (SEM) – Multivariate statistical analysis technique that combines factor analysis with multiple regression analysis to analyze structural relationship between measured variables and latent constructs, estimating multiple and interrelated dependence in a single analysis
• Select models and handle measurement areas using an industrial case study
  • Navigate the 3-panel SEM window (From List, To List, Diagram)
  • Use SEM to see if there is any built-in error in your measurements
  • Understand how to examine and compare restricted and independent model results
  • Build multiple models iteratively and interpret requirement Rule Status before running models
  • Locate and interpret model statistics and Report Warnings, and then modify or rerun models if necessary
  • Use Local Data filter to filter by variable
  • Save factor scores and apply latent variables to observed variables
• Use SEM to test a hypothesis using a sports case study

Note: Q&A included at time 36:43.

Resources

• SEM JMP Discovery presentation and blog