Have a look at the attached script. Notable points: * it seems that the original income variable has to be rescaled by a=20 number close to 10000 to match Greene's results * we don't do RE logit, maybe someone could write a package to do this,=20 but I don't think they're in high demand * in principle, you can obtain unconditional FE logit by adding=20 "dummify(id)" to your list of explanatory variables and dropping all unit= s=20 for which the dependent variable is time-invariant, but it's very=20 computationally demanding and the estimator is inconsistent anyway, so wh= y=20 bother? (It's in the script, though) * the function package "felogit" (by yours truly) does conditional FE=20 logit; download and enjoy * the --random switch to "probit" gives you the ML through Gaussian=20 quadrature; maybe you can play with the --quadrature option * again, FE probit is inconsistent, so nobody uses it. ------------------------------------------------------- Riccardo (Jack) Lucchetti Dipartimento di Scienze Economiche e Sociali (DiSES) Universit=C3=A0 Politecnica delle Marche (formerly known as Universit=C3=A0 di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti