I have four continuous variables which will be my dependent variables but I cannot create an index for them as they cannot be weighted. So if they were categorical variables I would have used multinomial logit which does not require ordering. Is there an equivalent to multinomial logit for continuous variables? I would really appreciate your suggestions. My dependent variables consist for four constructs: percentage of trials done inhouse, % done with domestic CROs, %done with global CROs and % done at foreign affiliates. My independent variables are firm and country level determinants for these four categories. I cannot weigh the four DVs as they cannot be ordered. I would really appreciate your thoughts on this problem.