Comparison of preference mapping methods: a case study on flavored kefirs

Abstract

The following preference mapping methods applied in sensory analysis were compared with the generally used internal preference mapping method (MDPREF): categorical principal component analysis (CATPCA), correspondence analysis (CAN), and penalty analysis (PAN). The aims, criteria, advantages, and disadvantages of the different methods are shown on the example of flavored kefir products. The pair correlation method (a robust variable selector method) has been applied for sensory data for the first time. It was also used for assessing the importance of “just about right” attributes of kefir products at first. The method is fast and easy to interpret and helps to identify product attributes having the highest impact on overall liking. Overall liking of the products was influenced mostly by the kefir flavor and fruit flavor intensity. Key factors when creating consumer questionnaires are simplicity, briefness, and comprehensibility. Instead of the long, one‐by‐one evaluation of the products, the consumers need to answer a simple ranking question. For such data, CATPCA produced very similar results to MDPREF. PAN gave more sensitive results about product discrimination compared with the MDPREF. A new way of applying CAN is the evaluation of a blind identification test. Results showed that two levels of flavoring could not be differentiated using MDPREF, CATPCA, and CAN, but PAN and pair correlation method reveals the differences. The first three methods are good choice for uncovering consumers' preference, while the last two are highly important for product developers. Combination of the proposed methodologies helps to uncover a higher degree of information in product development of flavored kefir products. Copyright © 2014 John Wiley & Sons, Ltd.