Have an idea?

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

Socio-demographic moderations in HB (covariates?)

Dear forum,

my goal is to detect and consider significant socio-demographic moderations (gender and education) in terms of attributes and levels.

Thus, I am looking for a solution to test if e.g. gender interacts with the formation of preferences (part worth utilities) on any attributes.
If so (significant differences regarding at least one level, compared between men and women), I need to know for which levels the moderation is significant (e.g. significant differences regarding level a but not b).

It is comparable to the example stated in the HBCovariates guide, but I cannot transfer the approach to my study. At the moment it seems to me that covariates in HB estiamtion are rather a means to improve PWU estimations for known moderators, instead of finding moderators.

How would you recomend to detect significant socio-demographic moderations?

Thank you very much for your help!

asked Oct 19, 2018 by Chris Berlin Bronze (570 points)

1 Answer

0 votes
answered Oct 19, 2018 by Keith Chrzan Platinum Sawtooth Software, Inc. (117,375 points)