Have an idea?

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

Obtaining an Average R-Squared with HB-Reg when using a Type 2 data file with only 1 obs per respondent


We recently obtained the HB-Reg package and started playing with it using some old survey data we collected.  Now the data we're using is a fairly straight forward satisfaction study, where an individual rates a single brand on 9 attributes (using a 10 point satisfaction scale - 1=Not Satisfied; 10 = Extremely) and one overall satisfaction questions (again on a 10 point scale).  So we only have 1 observation for each respondent, made up of 10 variables: 9 independent predictors and 1 dependent variable.  

Now we have setup the data as a Type 2 structure, and we can run the HB model ...but the Avg R-Squared always comes out as 0.0. So is there a way to obtains an R-Squared when there is only 1 observation per respondent?

Thanks heaps in advance!
asked Jun 7, 2012 by anonymous

1 Answer

0 votes
This is an inappropriate use of HB (only one observation per respondent).  R-squared should not be meaningful in such an inappropriate case.
answered Jun 8, 2012 by Bryan Orme Platinum Sawtooth Software, Inc. (177,015 points)