MULTIDIMENSIONAL-SCALING OF SORTING DATA APPLIED TO CHEESE PERCEPTION

A series of experiments examined the feasibility of using sorting data as input to multidimensional scaling (MDS) to create perceptual maps of cheeses. One study used a fairly diverse set of cheeses. Cheese names were also sorted for conceptual mapping. In a second study, only blue-veined cheeses were sorted. After sorting cheeses, intuitively sensible perceptual maps were obtained from MDS with low stress. Duplicate cheeses and those with highly similar characteristics were plotted close together Regression of attribute scab data into the space revealed major dimensions corresponding to flavor and texture attributes (e.g. smooth and soft vs. firm and crumbly). Two consumer groups, one experienced and knowledgeable about cheeses and the other a group of cheese consumers with no special knowledge, training or expertise, participated. Data from the two knowledge groups were similar except that the more knowledgeable group had a larger number of significant attributes when regressed through the MDS space for blue cheeses. The sorting task was rapid and simple for participants. The overall result substantiates the applicability of the MDS/sorting technique for perceptual modeling of food categories.