Dear Statalisters,
I am running a regression with several binary endogenous variables and a continuous outcome using the IV approach.
Y = aX1 + bX2 + c X3+ dX2*X1 +eX2*X3 + controls
,where
Y is continuous
X1 is discrete exogenous variable (demographic characteristics)
X2 is discrete endogenous variable
X3 is discrete exogenous variables (e.g demographic characteristics)
Y also suffers from selection bias, which implies that Heckman selection bias method might be necessary also.
Question1: If I were to run this "by hand," what steps would you suggest I should take? I have taken the following steps:
1. X2 = Z + controls
2. predict ghat
3. ivregress = controls (X2 = ghat) >>>> how do I incorporate the interaction terms and correction bias here? Could you advice?
Question 2: ivregress gmm Y controls (X2 X2*X1 = Z Z*X1) >>> would this be the correct way to approach this? how can I take into consideration the selection bias that exists in Y?
Your help and advise would be much appreciated.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
