Dear Statalists,
I have a problem which is causing me great grief! I am estimating a
Heckman twostep model which I am quite happy with. However, departing
from this parsimonious model, I want carry out some kind of model
selection. However, my understanding is that *estat ic* (AIC, BIC
etc.) a) is inappropriate and b) does not work after Heckman twostep.
Consequently, my question is what information criteria to apply in
their absence? I have trawled the web without much luck.
Hoping for someone to enlighten me on this issue, pronto, as I am
running out of time...
Cheers,
Lena Fredriksson
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
