Have an idea?

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

How do I interpret the utilities of main effects and interaction effects with conditional pricing?


In my research (based on an CBC) I applied conditional pricing to reflect a price-class (low, medium and high) at  specific brand (Traditional, Favourite, Top and Signature). So I have three price point for four brand. Meaning that respondents were shown a different low price for example the Traditional brand than for the Favourite brand. For my research I make use of the logit, LCA and HB method to interpret the utilities.     

My question is how I should interpret the part-utility scores of the main effects of brand and price on these methods. I am aware that I have to interpret the utility of a brand part-worth considering the average price that this brand was shown at. So, the brand part-worths are no longer disentagled from the price points.  However I wanted to know how exactly to interpret these part-worths for the main effects of price and brand. In addition I wanted to know how to interpret the part-woth utilities when an interaction effect is present between the brand and price. Do I just add or subtract the part-worth of the interaction effect to the utility at the average price or is this calculated in a different way?
asked Jan 9, 2019 by Rozita Baars

1 Answer

0 votes
It sounds like you are well on your way to interpreting the utilities properly.  As you state, when conditional pricing tables are involved, the main effects also capture the utility difference involved with the average shift in price between levels of the principal attribute (e.g. typically SKU or brand).

To account for the interaction terms, you simply add the main effect for brand, the main effect for price, and the interaction term for that brandxprice combination.
answered Jan 9, 2019 by Bryan Orme Platinum Sawtooth Software, Inc. (198,515 points)
Thans for your answer!

So to clarify. The utilities are calculated as follows:

- Attribute 1: Brand - level 1: Traditional with part-worth of 0.50
- Attribute 2: Price - level 1: Low with part-worth of 0.20
- Significant interaction effect between brand x price with part-worth of 0.10 at Traditional x low price.
Total utility = 0.50 + 0.20 + 0.10 = 0.80
As I understand correctly when a significant interaction effect is present between price x brand I can interpret the utilties for the prices I used in the conditional pricing table at the specific brand. For example the Traditional Brand at a low price. Or do I have to mention that the main effect is based on the average price?
Thanks in advance.

Kind regards.