Confronto di panel addestrati e inesperti nell'analisi descrittiva del formaggio pecorino

Many studies have been addressed to investigate whether Descriptive Analysis panels independently trained provide reproducible and meaningful results. Other authors have investigated the relationships between the data provided by trained panellists and the preferences and perception of consumers. Few papers have compared the evaluation of sensory attributes intensities by trained and semi-trained assessors. This paper describes a profiling study aimed to compare samples of Pecorino cheese (obtained with 2 different cheesemaking techniques and from animals fed 2 different diets) evaluated by two panels that represented different types and levels of training: the first was very experienced in Descriptive Analysis for cheese, whereas the other had no previous experience. Before performing the analyses each panel independently participated to some training sessions on the same set of samples. Nineteen and nine variables, respectively, were chosen by the trained and semi-trained panel to describe the cheese. Data were analysed using Generalised Procrustes Analysis (GPA) and Partial Least Square Regression (PLS) to provide an interpretation of the sample differences and panels performance. Results showed a good separation of the cheese samples between the 2 cheesemaking techniques whereas the dietary treatment did not have significant effects. The two panels disagreed on the relative size of differences in intensity for most attributes and for the direction of differences in some cases. Nevertheless they provided analogous distribution of the sample spaces.