Panel consonance in the sensory evaluation of apple attributes: influence of mealiness on sweetness perception

Abstract

Seventeen apple samples from nine different cultivars were evaluated by a trained panel of nine panelists. The first dimension of the raw principal component analysis models for each attribute (sweetness, sourness, juiciness, firmness, mealiness and crispness) accounted for more than 57% of the variance in sensory scores, except in the case of sweetness, for which the degree of explained variance was only 40%. A generalized procrustes analysis (GPA) was then applied to maximize the explained variance between the panelists. The final consonance results can be understood by looking at the interactions among the attributes: analysis of the GPA rotation matrix indicated that there was a rotation interaction between sweetness and mealiness that could explain the lower degree of consensus in the panel when evaluating sweetness. Conversely, the highest degree of consensus was observed for sourness, which showed the lowest correlation with the rest of the attributes.

This work was supported through project RTA2005-00050-00-00, financed by the Instituto Nacional de Investigación y Tecnologia Agraria y Alimentaria (INIA) of Spain’s Ministerio de Educación y Ciencia