Have an idea?

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

Is there a benefit to doing a conjoint where every task is fixed? How would this impact analysis?

I have always thought that it was best to include 1-2 fixed tasks in a CBC. A colleague has said that he traditionally has fixed all tasks (12 total). Whose approach is better?
asked Jan 24 by Brian

1 Answer

0 votes
It is very common, especially in academia, to use designs with a small numbers of versions - sometimes only one.  In this case one could program the design in our software as all fixed tasks.

These small designs typically make it hard or impossible to estimate interactions.  An interesting academic paper just came out concluding that random designs (even truly random designs using the Random design strategy in our software) are more robust than orthogonal (catalog) designs or efficient designs that many researchers use (and which I used to use up until about 5 years ago).  

The paper is "D-Efficient or deficient?  A robustness analysis of stated choice experimental designs," and it's cite is Walker, J.L., Wang, Y., Thorhauge, M. et al. Theory Decis (2017). https://doi.org/10.1007/s11238-017-9647-3.
answered Jan 24 by Keith Chrzan Platinum Sawtooth Software, Inc. (52,450 points)