Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: question about bivariate probit


From   "Sun, Yan (IFPRI)" <Y.SUN@CGIAR.ORG>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: question about bivariate probit
Date   Tue, 22 Jul 2008 17:02:43 -0400

Dear STATA list, 

I am now running bivariate probit model. The command is listed as
following (y1 is suspected endogenous):
"biprobit (y1 x1 x2 x3) (y2 y1 x1 x2 x3)" while both y1 and y2 are 0/1
dummies.

I have three questions:
1) When I run univariate preliminary model "probit y2 y1 x1 x2 x3", the
coefficient of y1 is strongly significant, which meaning y1 does affect
y2; but when I run biprobit model, the coefficient of is y1 is
insignificant and rho is small (0.15) and insignificant. What is the
possible reason?
2) After I run biprobit, I type command "predict xb1, xb1", but I find
95% of the predicted value for y1 (xb1)  are negative, which is strange
because the probability (xb1 is supposed to be probability for y1)
should between 0 and 1, do I misunderstood something here?
3) I try to run sensitivity analysis for different rho ranging from 0.1
to 0.5 to estimate regression, does anybody knows how to do this? 

Thank you for your help. 

Yan 
IFPRI

*
*   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   |   What's new   |   Site index