Have an idea?

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

Zero centered Diffs vs. Raw utilities


could someone guide me on the choice of outputted utilities to be used (Zero centered Diffs vs. Raw utilities) in case the objectives are:
-re-calculating total utility per winner combination per respondent
-determine preferences using (exp transformation of total utility) respectively to each  combination across respondents ?

Hope these few lines are clear enought to understand the context,

Many thanks in advance for any feedback,
asked Sep 6, 2016 by Rem (200 points)
retagged Sep 6, 2016 by Walter Williams

1 Answer

0 votes

The raw utilities will work for both of these needs.  And if you're using the logit share of preference rule in your simulator, you MUST use the raw logit-scaled utilities for prediction.
answered Sep 6, 2016 by Keith Chrzan Platinum Sawtooth Software, Inc. (60,075 points)
Dear Keith,

thank you very much for your help and quick reply. Highly appreciated. Best,