12-5 |
Use of discriminative and descriptive methods to characterize sensory differences among sparkling natural mineral waters with different levels of carbonation |
V. BRAUN1, M. A. O'Mahony2, J. M. Sieffermann1, M. Danzart1, M. Rogeaux3, N. Schneid3, and B. Rousseau4. (1) Département Science de l'Aliment, Ecole Nationale Supérieure des Industries Agricoles et Alimentaires, 1 avenue des Olympiades, Massy, 91744, France, (2) Dept. of Food Science & Technology, Univ. of California, Davis, Sensory & Psychophysics Lab., 1 Shields Ave., Davis, CA 95616, (3) Danone Vitapole, 15 avenue Gallilée, Le Plessis Robinson, 92350, France, (4) Institute for Perception, 2306 Anza Ave., Davis, CA 95616 Both discrimination testing and descriptive analysis seek the characterization of sensory differences among products. However, because they do not employ the same testing or data analysis protocols, they are usually seen as providing independent information. The objectives of this research were to validate the use of these two methodologies on sparkling waters and to compare the discriminations obtained on waters with three different levels of carbonation. For discrimination testing, twelve judges performed two protocols over thirteen sessions and three stimuli: the 2-Alternative Forced Choice (2-AFC) and 2-Alternative Choice (2-AC i.e. 2-AFC with "no difference" option) protocols. Degrees of difference (d' values) among products were estimated. Descriptive analysis was performed using a comparative Free Choice Profiling method derivative (Flash Profiling). Six judges performed three sessions in which they ranked the same three stimuli plus two alternatives on individual descriptors. Data were analyzed using Generalized Procrustean Analysis and Principal Components Analysis. Both methodologies allowed appropriate discrimination among products and provided results in agreement with the differences in level of carbonation. The d' values obtained for 2-AFC and 2-AC protocols were not significantly different, confirming Thurstonian predictions. Relative positioning of the products obtained by Flash Profiling provided a discrimination mainly correlated to simple "carbonated sensations": tingly, fizzy, bubbly... This research demonstrated that seemingly different methodologies such as discrimination testing and descriptive analysis can generate results that are consistent between them. The 2-AFC and 2-AC testing provide similar information, as indicated by the identity of their d' values, and shows that Thurstonian models can be applied to stimuli involving chemical irritation such as sparkling waters. In addition to a quick discriminative positioning of the products, Flash Profiling provides a basic verbal explanation of the sensory characteristics on which the products are differentiated. It suitably complements results from difference tests.
Session 12, Sensory Evaluation: Sensory techniques - Descriptive analysis
|