Have an idea?

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

Is it possible to have prices as inputs (instead of attributes) in Conjoint Analysis?

I used conjoint analysis as a tool for my research, but instead of using the traditional method where you have prices included as attributes in the profiles, I used prices as inputs (instead of ratings or rankings). So basically I directly asked customers for their WTP for the different profiles. Is it something completely wrong? Can I eventually transform my data in another way and avoid doing everything again?
Thank you
asked Apr 15, 2017 by Valeria

1 Answer

+1 vote

I have seen this done before.  It's just a matter of analyzing the data with a traditional regression-based conjoint analysis such as our CVA package allows.  I think doing it this way can sometimes requires you to do transformations on the price variable to make its relationship with your attributes more linear.
answered Apr 15, 2017 by Keith Chrzan Platinum Sawtooth Software, Inc. (52,325 points)
Thank you! Could you explain more precisely what type of transformations are required? The prices I have as inputs range from 1,000 to 5,000, so they are already in "scale" let's say.

For a linear model you could just leave the raw prices as your dependent variables.  If you look at plots of correlations or look at plots of your regression model's residuals, you may (or may not) see evidence that transforming the price variable would make for a better model, but I would probably not transform price unless I saw some such evidence or unless I had some theory.  For example, if you expect diminishing returns for adding utility, you might want to try a log transform of the price dependent variable.