I've been out of town. Hope my info isn't too late to be of use. The
research I did comparing logistic reg and discrim analysis for my
dissertation revealed that at least some statisticians believe that
logistic regression is more accurate when you have a dichotomous dependent
variable. If you have more than two levels of the DV then discrim analysis
is recommended. I can dig up the reference if you want
Barrie
Barrie Bondurant email: [EMAIL PROTECTED]
Assistant Professor of Psychology phone: 870-698-4217
Lyon College fax: 870-698-4622
PO Box 2317
Batesville, AR 72503
