Utility Estimation Method

Top  Previous  Next

Empirical Bayes and Hierarchical Bayes Analysis


Discover-CBC estimates preference scores (utilities) for each respondent and each level of your attributes using a statistical estimation approach called Empirical Bayes.  Due to the sparseness of data for each individual some information is "borrowed" from the other respondents within the data set to help stabilize and make these individual-level utility scores more accurate. Therefore, an individual respondent’s utility scores will change somewhat if new respondents are added or if respondents are deleted from the data set.  


For more information about Empirical Bayes and how it compares to Hierarchical Bayes, please read Appendix A from Discover-CBC: How and Why It Differs fromLighthouse Studio’s CBC Software.  Those who own the CBC/HB software may export their Discover-CBC data to a .csv file compatible with CBC/HB analysis by clicking Data | Advanced Settings under Collect and then clicking Export CBC Design & Choices.

Page link: https://www.sawtoothsoftware.com/help/discover/manual/index.html?utility-estimation-method.html