I am working with three dependable variables. 2 of them are numerical & one is categorical.
Can I add all these dependable variables only for multivariate probit regression on stata?
Is it okay to do that?
What should be the command for that??
Thank you.
You can use different models. For numeric dependent variable, it's better to apply OLS regression. Then the remaining two categorical variables, you can use probit regression
Just follow stata you tube about the command.
I hope this might help.
Thanks.
Dear Anindita Das Barshan
Could you clarify your question?
Unless you specify a model with several equations, in principle, an econometric model includes only a left-hand-side variable.
Dear Anindita Das Barshan
You cannot include all of your predicted variables in one model. You have to design three different models for three dependent variables.
It is recommended to separate your models with regression equations. Run the regression for similar models using OLS. For probit analysis, watch videos on youtube.
You can refer to studies in the given link:
https://www.researchgate.net/profile/Sitara-Karim/research
Hello, I didn't understand the models you want to estimate. But here are some options. Case 1) You can estimate 3 separate models, 1 for each dependent variable, when the equations are separate and unrelated (neither by functional form nor by simultaneity nor by error terms). Case 2) If the relationship appears only in the error correlation, you can estimate a sur (or non-linear sur). Case 3) when the equations are structurally related, you can use structural equations with latent variables (the gsem command is very useful for this).
Thanks a lot to everyone . You guys were right. The problem solved with OLS regression.
Thanks a ton...😊
- Badges
- Science topic
- Similar topics
- Correction