Dear Statalist,
I have a query re computing marginal effects from discrete-time duration
model with Gamma distributed random variable for unobserved heterogeneity. I
understand that the gamma variance term needs to be taken into account while
computing marginal effects but not sure how to implement this. The
discrete-time hazard function has the form
hj(Xij)=1-exp{-exp[Xij'B+Vj+log(ei)]}, Vj is the log of the integral of the
baseline hazard over interval j and ei is the unobserved heterogeneity term.
Any help on this is appreciated very much.
<<attachment: winmail.dat>>
