Bivariate probit for non binary variables

[Xavier Rousset]

Do you know the equivalent of a bivariate probit for variables that are not binary. My 2 interrelated variables have values from 1 to 7.

[John Mullahy]

Xavier: Is it appropriate to treat your outcomes as ordered on {1,2,...,7}? If so, then bivariate ordered probit may be suitable.

If you don't have bioprobit installed, then you might try

ssc describe bioprobit

[Xavier Rousset]

Dear Sir, thank you very much for your help, this is the very answer I needed. And this works.
I might ask you more questions further about how to interpretate my results.
Regards,
XR

[Xavier Rousset]

Thank you for your answers. One last question: could you tell me if my mathematical modelization for using biordered probit is correct:

"For each individual perception, we model the frequency of (un)fairness and (non-)ethical perception with a bivariate ordered probit. Formally, if (non-)ethical- and (un)fairness-perception of ADP are the ordered polytomous variables representing the non-ethical and unfairness perception frequency perception of ADP, we have:

Ethical perception of ADP = Alpha X + e
Fairness perception of ADP = Alpha' X' + e'

where X and X' are the regressor matrices, Alpha and Alpha' are the maximum likelihood coefficients, and e and e’ are the residual matrices with
[e follows a probability normal law N ([0 [1 , rho]
e]' 0], [rho, 1 ])

Thanks in advance (matricial script is not easy but I was told not to use attachment).
Regards
XR