Descriptive profiling versus direct similarity assessments of soft drinks

A set of seven soft drinks was profiled using a check list of descriptors and the same set of products was assessed for similarity by a panel of ten assessors. The profile data were used to generate a matrix of interstimulus distances, and this and the dissimilarity matrix obtained directly from the similarity assessment were examined using multidimensional scaling approaches (MDS/INDSCAL). The two methods yielded broadly similar results in terms of the spatial representation of the stimuli. This was further confirmed on analysis of the data using a multidimensional unfolding program which aimed to highlight any differences between the two sets of dissimilarity measures.