Dear community,
For the probability profiler as used to analyse data in choice experiment analysis, I am wondering how to optimally set the baseline scenario to compare probability profiler data.
Following for instance von Wedel & Hagist (2022), utilize the probability profiler to compare the baseline of the combination of attributes with the lowest marginal utilities. While such an analysis is not wrong perse to compare probability values between different combinations, my initial critique of such an interpretation is that the estimated probability values are not a representation of the probability of success. In order to estimate the probability that an individual in the population will choose a certain combination of attributes (scenario) shouldn't we opt for comparing the hypothetical future scenario to a baseline scenario in which the sum of all utility coefficients is 0.
For my study, in order to present such results, I manually created a business-as-usual scenario where the sum of the baseline utilities ≈ 0 by calculating the exact value of payments to balance out the marginal utilities derived from other attributes.
To clarify, see an attached files: baseline utility close to zero; probability with baseline close to zero; and probability with baseline lowest utility. Specifically my questions are the following:
1) Is such an interpretation to estimate probability profiles correct?
2) And if the answer to the first question is yes, is there a built-in function to set the baseline utility value at 0?
Hope all is clear kind regards,
Declan
