Equivalence tests – A review

Equivalence tests are becoming increasingly popular in many application areas including sensory sciences. They should be applied whenever the aim of the study is not to show differences, but to conclude similarity.There has been quite some debate about pros and cons of different approaches, reaching the sensory and sensometrics community in recent years. Parts of the sometimes heated debate are, in our opinion, due to mutual misunderstandings, and to different objectives and hence selection criteria for an “optimal” test, boiling down to the question whether power is the only optimality criterion to apply, or whether somehow vague criteria like intuition should be taken into account as well.This review intends to give an introduction into equivalence tests, starting with some general considerations on the statistical testing of hypotheses. We will subsequently give an overview over the most common approaches, some of which are shown to be inappropriate (e.g. the power approach). Some valid and relatively simple methods will be introduced and their correspondence to confidence intervals clarified, while we will skip the mathematical details of some more recent tests. Instead, the pros and cons of different approaches will be discussed and recommendations given.

► Equivalence (similarity/parity) tests re-visited.
► Conceptual description and motivation of the general approach.
► Thorough discussion of several frequently used approaches.
► Guidance for practitioners on how to choose an appropriate approach for equivalence testing.