Dear All,
I have a question related to Mundlak's approach to capture unobserved effects indicated in Wooldridge’s book (2002, pp 323, 487-489, 508). I am estimating impact of soil conservation technology on productivity using cross-sectional multiple plot observations per household. Following the book I tried to run linear random effects (re) regression model using mean of plot varying variables as additional regressors. In doing so the fraction of variance due to unobserved effects is zero. However, some of the mean of plot varying regressors are significant individually as well as jointly for some of my models, indicating that observed regressors are correlated with unobservables. The question is, can I run Pooled OLS using mean of plot varying variables as additional regressors since the fraction of variance due to unobserved effects is zero. I can not use fixed effects model since some households have one observation.
Thanks for considering this and taking your time.
With Best Regards,
Menale
____________________________________________________________________________________
Sucker-punch spam with award-winning protection.
Try the free Yahoo! Mail Beta.
http://advision.webevents.yahoo.com/mailbeta/features_spam.html
*
* 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/