Dear Statalist,

I have question related to the calculation of marginal effects in a bivariate probit regression (biprobit). Let's assume I have the following regression model in which A and B are my two binary dependent variables, and X (continuous) and Y (ranges from 1 to 5) are my independent variables:

Code:
biprobit A B X Y
I am now interested in the marginal effect that X has on A, as well as the marginal effect that Y has on A and B. Given that the calculation of marginal effects in non-linear models is not as straightforward as in linear models, I wanted to make sure that I implement the code correctly.

I found this earlier discussion on the same topic:
https://www.stata.com/statalist/arch.../msg00174.html

After reading this post, I was wondering how to calculate the marginal effect of X on A when there is no theoretical link between X and B. Specifically, I was curious whether (or why not), I could use a probit regression to obtain the marginal effect for X on A.

With regards & thanks,

Marvin