Have an idea?

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

Fixed Task in CBC for validity

my CBC exercise consists out of 13 random tasks and 2 fixed tasks.
I read somewhere that I can use the results of the fixed tasks to compare it with the general part-worth utilities of the 13 random choices to proof the
validity of my study.
How exactly would that work though?
Would I run an individual HB analysis for fixed task 1 and 2 and compare these part worth utilities with the one from the random sample?

Thanks in advance
asked Apr 7 by Charlotte

1 Answer

+1 vote

What folks do is to estimate the utilities from the 13"random" tasks.  Then use those utilities to simulate choices to the holdouts.  Compare the actual shares of the fixed tasks to the simulator-predicted shares for those fixed tasks.

Be worried if your predictions are poor (which can suggest something amiss in your modeling of the 13 random tasks).  We expect predicted and actual shares to be close.
answered Apr 7 by Keith Chrzan Platinum Sawtooth Software, Inc. (74,725 points)