Abstract:

 

Sensory difference testing is a business relevant topic, as evidenced by its use in cost reductions, ingredient and process changes, compliance with health initiatives, and quality control. In recent years there has been an increasing emphasis on measuring the size of possible inter-product differences and making confidence statements about them, as opposed to conducting hypothesis tests to detect possible differences. This is a valuable shift that has important implications for how we view difference testing for the objectives listed above. In current practice, sensory difference testing continues to rely on statistical techniques that in many ways fail to correspond to the issues difference tests are conducted to address. In this report, we discuss a more appropriate statistical framework to address more directly the objectives of difference tests and the variability in the measurements upon which the framework rests.

​