Abstract

A set of 334 commercial virgin olive oil (VOO) samples were evaluated by six sensory panels during the H2020 OLEUM project. Sensory data were elaborated with two main objectives: i) to classify and characterize samples in order to use them for possible correlations with physical-chemical data and ii) to monitor and improve the performance of panels. After revision of the IOC guidelines in 2018, this work represents the first published attempt to verify some of the recommended quality control tools, in order to increase harmonization among panels. Specifically, a new "decision tree" scheme was developed, and some IOC quality control procedures, were applied. The adoption of these tools allowed for reliable classification of 289 of 334 VOOs; for the remaining 45, misalignments between panels of first (on the category, 21 cases) or second type (on the main perceived defect, 24 cases) occurred. In these cases, a “formative reassessment” was necessary. At the end, 329 of 334 VOOs (98.5%) were classified, thus confirming the effectiveness of this approach to achieve a better proficiency. Panels showed good performance, but the need to adopt new reference materials that are stable and reproducible to improve the panel’s skills and agreement also emerged.