Olive oil sensory evaluation: An alternative to the robust coefficient of variation (CVr %) for measuring panel group performance in official tasting sessions

Sensory evaluation of virgin olive oil has become an issue since its first inclusion in law requirements, in 1991. After more than 25 years of application that greatly improved virgin olive oil quality, it is still criticised for its alleged unreliability and may require further reinforcement to improve its reliability. Nevertheless, many panel groups have begun their activity everywhere olive oil consumption is noticeable. The official method considers the way of evaluating panel group performance through different tests regularly carried on, both on single assessors and on panel group in its whole. During official tasting sessions, attribute intensity evaluations are valid if robust coefficient of variation (CVr %) of judgements is <= 20.0%. When median is very close to zero as it is for very light perceptions, CVr % greatly Increases, often overcoming the limit of 20.0. On the opposite extreme, for very high perceptions, even with unsatisfactory robust standard deviations, CVr % can accomplish that limit. The aim of this paper is to provide an alternative approach to the official method, which has the limitations mentioned above. It makes use of the robust standard deviation (s*), already used for calculating the robust coefficient of variation (CVr %), instead of CVr % itself and, furthermore, the worst acceptable candidate performance during selection carried on as required by the official method. This strategy avoids those drawbacks from using CVr %, and makes panel performance evaluation more feasible and balanced respect to the entire range of attribute intensities.