Factor variables: -mvprobit- was written about 20 years ago, well before factor variables. You'll have to incorporate them "by hand" (creating each one separately before using in regression). Maybe one day I'll get time to add factor variables, but am making zero promises. Interpret in the usual way. (Check out also work by John Mullahy (see his website)
Post-estimation tests: up to you, depending on your research question and analysis and what your discipline expects.
Yes, you can use the same covariate set in each equation.
The estimates of the (transformed) correlations are in the e() macros. You could untransform them using standard expressions (look at the code used in mvprobit.ado)
Reading: I have no additional suggestions beyond what you mention, and any other textbook discussions (any of the Wooldridge books? Cameron & Trivedi?) Look also at published examples of mvprobit analysis
Post-estimation tests: up to you, depending on your research question and analysis and what your discipline expects.
Yes, you can use the same covariate set in each equation.
The estimates of the (transformed) correlations are in the e() macros. You could untransform them using standard expressions (look at the code used in mvprobit.ado)
Reading: I have no additional suggestions beyond what you mention, and any other textbook discussions (any of the Wooldridge books? Cameron & Trivedi?) Look also at published examples of mvprobit analysis
Comment