Basic question here, but when collaborating on analysis with a colleague with a different version of stata (16/MP) than I was using (14.2/SE), I was able to get an meologit model to converge and with the exact same dataset and the exact same .do file, his version of the software returned the error message

initial values not feasible
r(1400);
I'm wondering what might be the reason for this - especially since the model will run on the *older* version of stata. Thanks very much!