In the data that we are working on, every respondent provides responses for different brands (say 3 to 4 brands). Here, each response corresponds to one particular brand, so each respondent might have 3 to 4 responses depending on the number of brands he/ she has responded for.
The independent variables are factors scores for different drivers and the dependent variable is the Purchase Intent. We are trying to derive utility scores for each driver which can be further used to derive driver importances. However, we are also provided with brand level weights which will have to be included during estimation. Our current methodology allows to include brand level weights during estimation. However, it does not provide us with respondent level importances which is our requirement. Thus, we were exploring HB-Reg and wanted to know if there is a possibility to include brand level (or response level) weights during estimation in Sawtooth's HB-Reg.