MINVAR: Stata module to conduct longitudinal measurement invariance tests

minvar tests whether a latent construct is measured equivalently across repeated measurements (or any set of parallel indicator lists in wide format). It fits a sequence of nested models and compares them: configural same items load on the same factor at each timepoint; all loadings and intercepts free weak (metric) loadings constrained equal across timepoints strong (scalar) loadings and intercepts constrained equal across timepoints strict (residual) optional fourth step (the strict option): each item's residual variance also equated across timepoints There is no limit on the number of indicators, and up to 16 timepoints/groups are accepted. Effects coding is used for identification (Little, Slegers, & Card 2006): each factor's loadings sum to the number of indicators and its intercepts sum to 0, so the latent variables keep the metric of the items and factor means are estimated at every timepoint. The residual of each item is allowed to covary with the residual of the same item at every other timepoint. Estimation is by method(mlmv) (full-information ML under missing data) unless overridden.