st: predicted value from fixed and random effects

["Dawit Lidia" <ghibertmk@hotmail.com>]

Dear All

I was trying to estimate the impact of agricultural technology on mean yield 
using the following regressions predicted values. I group my observations 
into those who adopted the technology (con==1) and those that do not 
(con==0). I specified the following exogenous switching regression model

1) Fixed effect
areg x1 x2 if con==1, absorb(hhno) robust
predict u_hat1, xbd
areg x1 x2 if con==0, absorb(hhno) robust
predict u_hat0, xbd

2) Random effects
xtreg x1 x2 if con==1, i(hhno) re
predict u_hat1, xbd
xtreg x1 x2 if con==0, i(hhno) re
predict u_hat0, xbd

My questions are as follow

1) Does it give sense to predict the fitted values including the individual 
effects of the absorbed variable to estimate the mean impact of the 
technology by taking the difference of predicted values (u_hat1-u_hat0)?

2) Do I need to use 'cluster' option in fixed and random effect models if I 
assume observations within household (group) are not independent?

Thanks for your help.

menale

_________________________________________________________________
MSN Messenger http://www.msn.no/computing/messenger - Den korteste veien 
mellom deg og dine venner

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/