Have an idea?

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

How to conclude the preferences per segment were significantly different from each other.

How can I conclude, next to significancy of attributes to include this attribute was important for respondent's choice in every segment, that preferences per segment were significantly different from each other?
asked Jun 28, 2017 by SanderM01 (160 points)

1 Answer

0 votes
If you've run hierarchical Bayesian (HB) analysis and you have respondent-level utilities, then the easiest way to do this would be to treat the numbers like any other numbers (e.g. rating scales).  Then you can use the t-test function in SPSS or even in Excel or in online t-test procedures.  

If computing the test by hand you'll need to get the means and standard deviations of each utility for both males and females, again something easily done in Excel.  

If you want to get really fancy you can run Bayesian versions of these tests instead of traditional frequentist tests you learned in school.  Contact me at keith@sawtoothsoftware.com if you want to know more about Bayesian tests.

If you're using an aggregate logit model please contact me and I'll refer you to the relevant research about how to conduct such tests.
answered Jul 2, 2017 by Keith Chrzan Platinum Sawtooth Software, Inc. (66,225 points)