Dear statalisters,
I have two probit regressions that look like this:
When I run both separately using "probit", X1 is highly statistically significant in equation 1 and has a large marginal effect, which is to be expected and logical given the dataset.
When I use cmp to estimate a bivariate probit model, X1 becomes insignificant whilst all other coefficients stay almost exactly the same. The command I use is:
Am I right in the assumption that I am now unintentionally estimating an IV-probit model? And if so, is there any way (and reason) to estimate this as a bivariate probit model as if X1 was not part of the first equation?
Thank you very much for your help
I have two probit regressions that look like this:
Code:
probit Y1 X1 X2 X3 , robust
Code:
probit X1 = X2 X3 , robust
When I use cmp to estimate a bivariate probit model, X1 becomes insignificant whilst all other coefficients stay almost exactly the same. The command I use is:
Code:
cmp (Y1 = X1 X2 X3) (X1 = X2 X3) , vce(robust) ind($cmp_probit $cmp_probit)
Thank you very much for your help
