Capturing Images in a Net: Perceptual Modelling of Product Descriptors using Sorting Data

One task facing market researchers is that of profiling a product, by taking a predefined list of descriptors and eliciting the relationship between the product and each descriptor in turn (“applicable” or “not applicable”; “good description” or “bad description”). Accuracy is gained if in addition the relationships among the descriptors are known: they become a meaningful nomological net, which can be depicted as a geometrical model. In this study, a large, extensive net is constructed, using an innovative split-deck sorting procedure to collect the data. It is argued that in comparison with the conventional treatment of sorting data, the custom-designed algorithms involved in the construction of this net are less susceptible to artefacts.