Statistics for Use in Discrimination Testing

This chapter considers statistics used in sensory difference testing from the perspective of business objectives and risks. Selection of an appropriate test method and experimental design must be aligned such that data analysis is able to provide meaningful information that relates back to the business objective and, where applicable, provides an actionable result. Various test methods are discussed and classified. Consideration is given to implications of analysis for unreplicated and replicated data, to statistical power, and to statistical tests for difference and equivalence. The link between experimental design and statistical analysis is stressed.