Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

HB analysis for data from two cities and 95% confidence intervals

Dear Sawtooth community,

I have an ACBC Lighthouse Studio study distributed in two cities. All survey data collected in the two cities were combined for HB analysis and I segmented the entire data by city. Now I want to analyse one city in detail with regards to this city’s utilities, average importances, segmentation, shares of preferences, etc. Should I run another HB with just data of this specific city’s participants and get individual-level results? Or should I use the individual-level results of that city from the HB of the entire sample? According to my test, the results of the HB of the only one specific city are different from the segmentation results of the HB of the entire sample based on city.

Another question is about 95% confidence interval. I found 95% confidence intervals for each level and for each attribute calculated by SPSS are different from confidence intervals calculated by Analysis Manager-Utility Report in Lighthouse Studio. Why is that? (I think the ones calculated by Lighthouse are correct, but it seems Lighthouse does not do much statistical analysis when I want to get confidence intervals in segmentation).

Thanks for your help!
asked May 1 by Xinlei Hu (340 points)
The correct (Bayesian) way would be to develop the 95% confidence interval: For all draws after covergence, pick out the 2.5th  and 97.5th  percentile values - in between lies your 95% confidence interval.
(Draws can be found in the ...alpha csv file)
Thanks for your answer, danny!

Your solution to the original question

Please only use this to answer the original question. Otherwise please use comments.
Your name to display (optional):
Privacy: Your email address will only be used for sending these notifications.
Anti-spam verification:

To avoid this verification in future, please log in or register.