Dear Professor Cottrell  ;

Thanks so much for your kind and informative response on my question entitled “Help on simultaneous​us equation model”

In this regard I have got sufficent answers on the questions I asked you. Nevertheless, I still want your suggestions on predicted probabilities and their application to reduced form equations and if there is that sort of application in GRETL.

I understand now GRETL does not predict probabilities rather give us fitted or residual values after we estimate an ordered probit model and ask the program to generated predicted values.

Now my question is can I use this predicted value in my second stage ordered probit regression or are these fitted value equivalent to P^ and D^ given original endogenous ordered variables P* and D*?

I really thank you for any assistance which you might be able to provide.

With kind regards,

Tenessa L.

PhD Student
Haremaya University, Ethiopia