Roy Levy posted on Wednesday, January 24, 2007 - 10:54 am
I have found that running a path model in which there are two dichotomous outcome variables gives me the same results as running two separate models (each with one outcome) except that coefficicents are mislabeled. That is, the coefficients that should go with the first outcome are listed for the second outcome and vice versa. I'd be happy to provide my input and data files as an example if you wish.
It is true that if both regressions are just-identified, you will get the same results if you run them together or separately. If you want me to look into the alleged label switching, you would need to send your input, data, output, and license number to firstname.lastname@example.org.