Probit or Random effect probit

[zaineb belkhiria]

Hello everyone.
I wish you could help me. I am studying the effect of gender on business failure. I have a binary dependent variable that takes two values: 1 if the company is bankrupt and 0 otherwise. I have panel data (several companies over several years) and I have applied the probit and random effect probit, but I don't know how to choose between the two models or what the most appropriate model is. Is there a test of homogeneity or heterogeneity to do before applying the probit model or random effect probit model( homogeneity test on non-linear regressions)? Also, the Breush-Pagan test, which is normally applied to linear regressions on panel data, can it be applied to nonlinear regressions ?
I look forward to your response and thank you for your time. Best regards.

[Joro Kolev]

Look at the bottom of the output of your -xtprobit-, and you will see something like

LR test of rho=0: chibar2(01) = 5984.32 Prob >= chibar2 = 0.000

This is a test of whether your random effects can be ignored. In the example above, they cannot be ignored and we should choose random effects probit over pooled probit.

[zaineb belkhiria]

I understand, sir. Thank you very much for the clarification. It really helped me.
I have another question about the random effects probit. I can't calculate the "Sensitivity", "Specificity", and the "Correctly classified" percentage. For a simple probit model, I ran the "estat classification" command on Stata and it gave me results.
whereas for a random effects probit model, stata displays "estat classification not valid
r(321);"
do you have an idea what I have to do to get results like for the simple probit model please ?
best regards.