Have an idea?

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

Multiple Price Functions in One Alternative Specific Design

Hi Team (with special thanks going to Bryan for his ongoing help here),

But something struck me as I was looking to collapse two price functions into one for one concept in an Alternative specific design I am creating, and will be relevant, I believe, for other applications.

I followed Bryan's advice in combining two prices into one (a toll road charge and a daily parking fee for the alternative specified for car) and after reading the instructions listed in the help from the stand alone LC Help Appendix C where it describes the process for creating a continuous price function by exporting the csv file and then altering the levels to be continuously coded and "User-specified" in the LC options for a single run estimation, rather than part-worth estimated and I can see how this is achieved by collapsing two columns together (by using the actual price values and  subtract from each value the mean of the values, which will differ between car with parking with toll and car parking without toll).

However how do I treat the price functions for the other Alternative Specific Concepts in the design? Such as Bus and Tram...


A look at this simple design should make clear what I'm trying to achieve.

What I mean by this is that each alternative specific concept has its price variable in a different column in the design matrix and hence the response file. Should I make all the price functions in the ALT-Spec design continuous (for consistency) or can I just have the car concept estimated as continuous and the other concepts as partworths?

The transformation of one price variable to continuous but I don't know what to do with the others, what would be best practice?
related to an answer for: Multi-Price Alternative Specific Design
asked Feb 26, 2017 by Jasha Bowe Bronze (1,745 points)

1 Answer

0 votes
Well, to my thinking, I would code the price function per alternative in the manner that produces the best fit to the data (without overfitting).  Which means you could mix the independent variable coding across different price terms for different alternatives.  However, depending on  what you intend to do with the data and your intended audience, you might need to make a different decision.
answered Feb 27, 2017 by Bryan Orme Platinum Sawtooth Software, Inc. (152,955 points)