Hi,
I am currently comparing the use of binary to ordinal logistic regression in analyzing a screening test (GHQ) for complex survey data, has anyone tried comparing those 2 methods regarding magnitudes of coeficient, and Goodness of fit stat? furthermore, stata does not support model assessment for survey data, are u aware of any software that might prove usefull in such situtation?
Please e-mail me back
Thank you very much for any comment
Khalil
|