Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: two fitted values as regressors


From   Pellegrino Gabriele <Gabriele.Pellegrino@unicatt.it>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: two fitted values as regressors
Date   Sat, 14 Apr 2012 20:45:33 +0200

Hi statalisters,

I want to estimate a recursive model made up of 6 equations. More in particular I want to estimate two 2-type tobit models (4 equations: 2 selection equations and two main equations) by using an heckman two step procedure and then a biptobit model (two equations). In order to try to correct for some endogeneity problems I decided to estimate the last step of the model (biprobit) considering as main regressors the fitted values of the dependent variables of the two previous steps. In doing so I have used the command "yexpected" that allow you to obtain the predicted values for all the observations in your sample after an heckman two step estimation. Unfortunately, when I estimate the last step of the model, the coefficients of the two variables (the fitted values of the previous steps) turn out to be quite unstable. This is probably due to some problem of multicollinearity. I tried to repeat the estimations excluding some outliers from the sample , and I tried also to estimat!
 e the model considering different dependent variables, but I could not solve the problem. Does anyone have any useful suggestion to treat this issue?

Thanks,

Gabriele

*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index