Have an idea?

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

In HB Reg if a variable is coming significant for one group but not not for another group, then what do we do?

If such a situation happens, then do we leave that variable or consider the group level model with that insignificant variable itself?
asked Feb 21, 2012 by anonymous

1 Answer

0 votes
Not sure if I understand your question.  It is quite possible in HB-Reg for a parameter to be significant for just some individuals, but not for others.  But, that doesn't mean the parameter should be discarded.  It can help predictions overall if it is significant for a significant number of respondents.

This happens in conjoint analysis and HB all the time.  Some attributes are not important to some respondents.  But, we retain those parameters in the model, as it helps predictions to know that for some respondents the attribute has impact.
answered Feb 21, 2012 by Bryan Orme Platinum Sawtooth Software, Inc. (177,015 points)