Profile of choice-based conjoint analysis | XM Community
Skip to main content

Hi there, I have many questions,

  1. How can I get the information of the generated profiles in Qualtrics?

  2. Why there are several sets of profile-combination in the generated profiles for different participants? I mean to my knowledge, all participants are required to answer the same questions consisting the same profiles set, rather than some saw one kind of combination of profiles, and others saw another kind of combination.

  3. Why the export of the number of the row in the results of individual utility does not equal to the number of total responses in the survey? Theoretically speaking, they should be the same, because all participants need to finish choice decision and answer questions in the survey block, therefore, the number of individual utility should be equal to that in the responses of the survey.

  4. Can anyone working in Qualtrics tell me the key algorithms in generating those random profiles? I am totally lost!


Hi yon0606! It looks like you were able to get a response from our Support team in regards to your questions! While the Support team is not authorized to provide the code that calculates the individual utilities, which are then used to calculate the preference shares and feature importances, they were able to recommend that you view our Conjoint Analysis White Paper. This page provides all of the information that we release, and goes into detail about how we perform Hierarchical Bayesian estimation to calculate utility scores. Additionally, the reason that the number of the rows in the results of individual utility does not equal to the number of total responses in the survey is because the total responses in the survey will include people who failed to complete the conjoint section or started the survey but never reached the conjoint section. Individual utilities can only be calculated for those who completed the conjoint section of the survey.

Leave a Reply


Terms & Conditions