> --------------------------------------------------
> We would like to estimate either a fixed effect or random effect model in which 2 explanatory variables are binary choice, dummy variables. For each there are issues of self-selection bias for which we have the following solution.
>
> (1)We apply bivariate probit model to these 2 binary choices.
> (2)Using the estimation results, we calculate the predicted probability
> of each variable.
> (3)Using these predicted probabilities, we estimate the 2nd stage model.
>
> We have the following questions:
> (1) Is this an appropriate strategy for a model with this type of structure? ie. the error terms of the equations for the binary choice models are correlated and the error terms of the binary choice models and the 2nd stage model are correlated.
>
> (2) If NO, what is an appropirate strategy and how should it be modeled in STATA? Are there any good examples, articles or other sources that we can refer to?
>
> (3)If YES, is it possible to calculate the separate predicted probabilities to
> replace the 2 dummy variables in the second stage.
>
> -----------------------------------------------------
>
>
>
--
Eric W. Welch
Associate Professor
Public Administration
University of Illinois at Chicago
http://www.uic.edu/orgs/stresearch
*
* 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/
