Dear Statalist,

I conduct a discrete choice experiment. I have seven attributes with 6, 4, 3, 3, 3, 3, and 3 levels All levels are dummy-coded. I chose the "best levels" (i.e., the levels that generate the maximum utilities) as references for the estimations. Each respondent faced two alternatives and could choose to refuse both. I defined the latter as an alternative specific constant. It appears as being another variable in my estimations, similarly to the attributes' levels.
I use the -clogit-, -mixlogit-, and the -lclogit2- commands as choice models.

I want to compute the predicted uptake of treatment using the best case scenario and the worst case scenario (the scenario defined by the attributes' levels that generate the worst desutilies). I know that it is often done in the literature about discrete choice experiment. I tried to use the logit formula but I don't understand how to do that knowing my specifications.

If someone could help me, I would really appreciate !

Thanks, Gabin.