Dear Statalisters,
I wonder if any of you could help me on this.
I want to examine the impacts of two variables, say, variable A and
variable B, on variable C. Variable A and B are both subject to
self-selection problem. My proposed solution is this:
First step:
Estimate Probit model with A as the dependent variable, B and other
control variables as independent variable. obtain reversed mills ratio
1;
Similarly, Estimate Probit model with B as the dependent variable, A
and other control variables as independent variable. obtain reversed
mills ratio 2.
Second step:
estimate two equations separately, both with C as the dependent
variable, but one with B, reversed mills ratio 1 and other control
variables as indepdent variables, and another one with C, reversed
mills ratio 2 and other control variables as indepdent variables.
But I am not sure if this is right as someone could argue that A and B
should enter the equation of C simultaneously, rather than separately.
On the other hand, if I have both A and B in the equation of C, how
can I deal with the self-selection problem?
Or probably you have other suggestions?
Many thanks.
Natalie.
*
* 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/
