Reference manual for Stata's gsem command, which fits generalized structural equation models, covering syntax, options, and stored results.
This document is the reference manual entry for gsem, Stata's command for fitting generalized structural equation models (SEMs). It documents the command's syntax, the option groups for model description, groups, latent classes, estimation, reporting, and syntax interpretation, and lists the results stored in e() and r() after estimation.
The remarks section explains how gsem differs from Stata's sem command: gsem allows generalized linear response functions, multilevel models, categorical latent variables, and factor-variable notation, while omitting some features of sem such as standardized coefficients and summary statistic datasets. Advanced topics covered include default normalization constraints, default covariance assumptions, and how to solve convergence problems.
gsem fits generalized structural equation models (generalized SEMs), and it is the command used when you use the Builder in gsem mode.
Go to Statistics > SEM (structural equation modeling) > Model building and estimation.
fweights, iweights, and pweights are allowed; weights are not allowed with the bootstrap prefix, and vce() and weights are not allowed with the svy prefix.
gsem allows generalized linear response functions, multilevel models, categorical latent variables, and Stata's factor-variable notation, none of which sem allows.