Developer Tutorial: Building Structural Equation Models in JMP Pro
Statistics, Predictive Modeling and Data Mining
This session is for JMP Pro users who understand multiple regression analysis and have missing data, are analyzing situations that include latent variables, and/or want to test differences across subpopulations.
Structural equation modeling (SEM) is a framework that enables researchers to model relationships among both observed and latent variables. Structural equation models can range from simple (e.g., linear regression, confirmatory factor analysis) to complex (e.g., path analysis with latent variables, conditional latent growth curves), and can be fit to data from multiple population segments or groups.
Analysts and scientists use SEM for developing surveys, measuring latent variables and understanding their associations with other variables, fitting linear models with cutting-edge algorithms in the presence of missing data, accounting for measurement error in their models, and testing competing theories about the structure of relationships among variables.
One of the developers of SEM in JMP Pro introduces SEM and demonstrates how to use the techniques. She explains how to use SEM to leverage observed variables to measure error-free latent variables, explore associations across latent variables, and test differences across population. She explains why and how the development team implemented, in JMP Pro 17, systematic model comparisons of statistical effects across groups, linking exploratory factor analysis with SEM to improve survey development, visualizing model predictions, and accessing multiple types of covariance and correlation matrices and heat maps. The session includes time for Q&A.
This JMP Developer Tutorial covers: constructing, interpreting and using Structural Equation Models