|
I looked probabilty function it seemed to me you modeled the problem as a multinomial logit model(discrete choice model). If this is the case, I think a routine in SAS (proc mdc) deals with the problem. But it has limitations for this model to study cross effects(elisticities). But it seems to me the proper model would be a nested logit model according to your document -- which is also available in SAS (proc mdc). Please post if you need an additional help.
|