[Conjoint Analysis] Measuring (cognitive) cost functions instead of value functions?

Hello :wave:,

usually one uses Conjoint Analysis, like in my case ACBC (Adaptive Choice Based Conjoint) in order to measure preferences and estimate utility or value functions. There are a number of attributes and attribute levels that provide value and a price (usually providing negative value). Thus probands see screens with different choice sets and a question like "which stimuli would you choose?" plus a "none" option. Based on these displayed choices part worth utilities etc. can be calculated.

In my case I would like to estimate different cognitive costs resulting from attributes level combinations. Hence, it is the other way around, the attribute levels provide negative value or coginitive costs and they would receive a discount in return. I am not sure how to best approach this, I could ask something like "What stimuli do you reject" and an "all" option.

At the end I would like to estimate aggregated cost functions for each stimuli, calculate attribute importances (span width) and show that different attribute levels cause different levels of cognitive costs.

It might be really simple to change the conjoint analysis from a method to estimate value / utility functions to a method that estimate cognitive cost functions, because the algebraic signs are inversed (attributes cause costs and the price is a positive discount in my research).
Another view on this is, that it does not matter, because my research question about cognitive costs does not differ from customer value functions. The CBC helps to estimate utility functions and I just change the common set up of several value generating attributes and a cost genereting attribute (price) to several cost generating attributes and a value generating attribute (discount). In this case I would just ask as well, what would you choose and interpret the results similarly. In this regard I am not sure about the interpretation of the part worth utilities...

Does anyone have experience on that or can advice on how to best state the screen questions and interpret the results subsequently?

Thank you very much in advance for your highly appreciated support!