Abstract
Two groups of subjects sorted a set of fragrance materials representing woody, citrus and ambiguous (woody/citrus) odors. One group divided the set into as many categories as they wished, while a second group was restricted to sorting into only two categories. The number of times a given odor pair was sorted into the same group was counted as an index of similarity and submitted to multidimensional scaling (MDS). The restricted group produced an MDS configuration with only two clusters (woody and citrus), with ambiguous odors on the edges of each cluster. The unrestricted group produced four clusters, with lime fragrances and ambiguous odors falling into a central cluster between woody and citrus groups. The sorting task offers a potential shortcut in generating data for MDS analysis that can be used for perceptual mapping of odor classes. However, the complexity of the resulting configurations depends upon the restrictions imposed on subjects during sorting.

This publication has 4 references indexed in Scilit: