Dear Stata listers,
Does anyone know why is it that you cannot use robust nor clustered errors in some two-stage process in Stata like -heckman- and -ivprobit-? Those options seem to be available only when ML is allowed.
I was trying to find something on the manual on the specifics, but have not found anything.
Thank you.
Fernando
Does anyone know why is it that you cannot use robust nor clustered errors in some two-stage process in Stata like -heckman- and -ivprobit-? Those options seem to be available only when ML is allowed.
I was trying to find something on the manual on the specifics, but have not found anything.
Thank you.
Fernando