Is multinomial logit analysis feasible?If by "multinomial logit analysis" you mean logistic regression analysis with a response variable having three or more levels, then this can be done in JMP, which can fit generalized logit to nominal responses and cumulative logit to ordinal models.
