Sensory profiling data studied by partial least squares regression

The statistical analysis of a descriptive sensory profiling data set distri buted at the sensometrics meeting is presented. The data set is analysed wi th focus on the sensory differences between products (cooked potatoes). The data analytical strategy involves a descriptive statistical analysis to ob tain an overview of the distribution and standard deviations of the scores for each sensory attribute. Subsequently, three-way analysis of variance (A VOVA) of the data gives a statistical measure of the reliability of the sen sory attributes supplemented by principal component analysis, which visuali se the main tendencies of systematic variation. Discriminant and ANOVA part ial least squares regressions are used to relate the sensory structure to p roduct design structure and vice versa. Statistical reliability and predict ive validity of the product differences are obtained by ANOVA and cross-val idation. Similar data structures are observed in the various multivariate m odels. Texture, taste and flavour attributes differentiated the potato samp les, with the texture attributes being most reliable. It is emphasised that an appropriate interpretation of the profiling data should also include kn owledge of the experimental background. (C) 1999 Elsevier Science Ltd. All rights reserved.