Have an idea?

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

Standard deviation of Latent Class Attribute Importances


I was just wondering why there are no standard deviation values available for the Latent Class Attribute Importances.

In detail I want to compare the standard deviations of a HB estimation with one segment and the two-segment solution with latent class.

Can anyone please help me?

Thanks a lot!
asked Sep 24, 2014 by Michael

1 Answer

+1 vote
With our conjoint & MaxDiff products, we generally only report standard deviations when individual-level data are involved (standard deviations characterize the degree of heterogeneity across people).  Latent Class parameters are segment-based estimates, not individual-level estimates.  Our Latent Class procedure cannot estimate standard deviations for utilities or importance scores.
answered Sep 24, 2014 by Bryan Orme Platinum Sawtooth Software, Inc. (189,140 points)