We conducted a single profile, 5-point scale CVA and are now working within our simulator. We are calibrating the none option based on stated share of patients that would prescribe this drug - outside the conjoint. Stated share is 20% - which means, using partial vote, we would have to set the none value at 9. 9 is obviously way above the 1-5 scale but the math makes sense given the partial vote calculation if we need to get it down to 20%. Maybe going back to whole vote makes more sense for CVA to allow for the appropriate sensitivity when a 1-5 scale represents the range? From a rating perspective, there is a huge difference between say a 1 and a 4 yet it would still give the 1 a 20% chance of being chosen.
Thoughts on what people tend to do and if there's a 'more correct' way?