Courtesy of Kit Baum, the new package -twostep- by Johannes Giesecke and myself is available on SSC. -twostep- is a bundle of programs to ease multilevel analyses with the two-step approach. The two-step approach to multilevel analysis means to separately estimate a parameter of interest in a unit-level data set (e.g. individuals) for all categories of a cluster-level variable (e.g. countries, states, schools, etc.). The two-step approach is sometimes seen as superior to the more standard one-step approach (see -help mixed-) if the number of observations on the second level is small. Additionally, two-step multilevel analysis may be used for checking the model assumptions of the one-step approach, or for exploratory data analysis of multilevel data.

Most of the programs in the -twostep- bundle are just wrappers. They are designed to perform various tasks of two-step multilevel modelling with very similar commands. The aim was to perform two-step multilevel modelling interactively. We hope you like it.

See the proceedings of year's 2021 German Stata Conference for examples of -twostep.-: https://www.stata.com/meeting/german...1_Giesecke.pdf