Abstract:

 

A persistent dilemma in comparing products is to know when the difference between them becomes consumer relevant. Since the probability that any two products are exactly the same is zero, rejecting the null hypothesis that they are identical may not be that informative; an analysis will always result in such rejection provided there is a sufficient sample size. What is more important to know is whether the products differ by enough for their difference to be consumer relevant - establishing this knowledge is an outstanding problem in the sensory and consumer research fields and various methods have been suggested. One method is to benchmark product differences from past tests where it is known that consumers continued to purchase the product even in the presence of variability or change. This method can be employed when products are routinely made in different factories or when blend and flavor modifications have already been introduced without any appreciable loss of sales. Another approach is to link internal panel measurements with consumer hedonic response to the set of differences. But ideally, it would be useful to have a consumer-based estimate of the average criterion that consumers use to decide whether products are the same or different. In this report we discuss a way of developing this method through the use of same/different judgments.

​