In the Food and Technology Science is important to improve products and methods and this work aimed to present an analysis strategy for sensory data, using the ordinal nature of the response variables, that corresponds the hedonic scale. The data are from a study with prebiotic beverages made from cashew nut added to grape juice to evaluate its sensory characteristics and the experimental study followed a balanced incomplete block design, in which each panelist evaluated 4 of the 13 proposed beverage formulations. As statistical methods were used correspondence analysis techniques and proportional odds mixed models. Three better beverage formulations were selected: 8% sugar and 40% grape juice, 6% sugar and 44% grape juice and 9% sugar and 30% grape juice. It was verified that the correspondence analysis as well as the ordinal mixed models were useful for sensory data analysis, contributing to methods in the area.
