Hi,
My colleague and I want to control for sample selection in a model of two dependent variables (time allocated to activity 1 and time allocated to activity 2) Both activities can be related but individuals need not do either. We would appreciate advice on modelling strategy/estimation.
a) Is it better to estimate, say, a model of total time allocatable to the activities but with two selectivity parameters (lamdas). One for each activity.
b) Estimate a biprobit model that covers each activity and then attempt to estimate the lamdas separately.
Depending on the answer to this question, we would be grateful for advice on either
a) How to estimate the model with two lamdas in STATA (it can be estimated in NLOGIT).
b) If the biprobit route is possible, how would the separate lamdas be estimated?
Many thanks for your help.
Paul.
*
* 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/
