Have an idea?

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

How to analyse fixed CBC tasks?

I followed the instructions and created a Holdout CBC (fixed tasks only, 4 per screen) within my ACBC exercise.

In my output the choice of which fixed tasks (1,2,3,4) on each screen was chosen is recorded.

Under the Tab "Analysis-->Calculate CBC Counts/Estimate CBC Utilites" I get an error message "There is no respondent data for CBC".

My questions are:
-Is the analysis not possible because I only included fixed tasks (as they are not included in the utility estimation usually)? Or is there a serious problem with my CBC design?

-Is it enough to have the choices (1,2,3,4) and upload them to the SMRT later on in order to perform the validation tests after calculating the ACBC data?

I am a little worried that the holdout data is not recorded properly and just wanted to make sure before fielding the survey…thank you so much for your answer!
asked Apr 2, 2014 by anonymous

1 Answer

+1 vote
By default, our CBC software just wants to analyze "random" tasks within Counts, logit, latent class and HB.  The fixed tasks are by default unchecked within your task filters settings.  So, if you try to analyze only fixed tasks using Counts without changing the default settings on your task filter, it will say there is no data.  Just go into Settings for Counts and change the task filter to include the holdout task you want to analyze.  But, the best way to analyze fixed holdouts is just by exporting the data to .CSV or SPSS and analyzing with your favorite tab package.
answered Apr 2, 2014 by Bryan Orme Platinum Sawtooth Software, Inc. (179,015 points)