TWANG mnps command not recognized

[Jeff Kane]

I'm trying to run a propensity score analysis for a study with 3 treatment groups. I've downloaded the Stata Twang files from RAND and installed them, created a link to R, but when I try to run the mnps command, I get a "command mnps is unrecognized" error. I checked in the ADO folder and there is no mnps.ado file in there, or in the original zip archive containing the files. There is a ps.ado file and I can run the ps analysis for a 2 treatment group design. Apparently, the archive supplied by RAND does not include the mnps ado files, which are necessary to run the multinomial regression to produce the propensity scores for 3 or more treatment groups. Can anyone confirm that this is correct and if so, where I can get the mnps ado files, or tell me what else I might be doing wrong?

[Phil Bromiley]

Welcome to Stata list. When you use a user written procedure, help particularly of a technical nature depends on someone actively answering questions happening to use that procedure. If you can't debug it yourself, you might contact the programs authors.

Note that in recent additions Stata does include a wide variety of treatment models including propensity score analysis. While many user written procedures work very well, often Stata created procedures have been vetted more carefully than user written and often interact better with the rest of Stata.