Dear Guillaume,

I would like to pose two easy interpretation questions concerning the output produced by reoprob ( just to be on the safe side):

1) is the RHO reported the same in construction like the rho of the xtprobit command ?
( i.e. the proportion of variance explained by the panel-level variance?)

2) Does the LR-test compare a base model consisting of a constant only with the full model ?

And maybe a last question: to your knowledge, is there a restriction for reoprob with regards to the number of observation or time periods, so that the numerical approximation method might not be feasible any more ?

Thanks a lot in advance and I wish you a happy new year !!!!!

Justina