MAPPING IN PRODUCT TESTING AND SENSORY ANALYSIS: A WELL LIT PATH OR A DARK STATISTICAL LABYRINTH?

Mapping is a popularly accepted statistical method, currently used in many applications in order to reveal and perhaps understand relations between stimuli. This paper discusses some of the shortcomings of mapping as a procedure, specifically those having to do with the 'actionability' of the data. The paper recommends that mapping be used to embed the stimuli in a geometric space, but then the coordinates of that space be used as independent variables in a model. In this way the geometrical representation does not stop at being a 'pretty picture' without further utility, but instead incorporates within it the necessary tools to allow researchers to go from any location in the space to any other location.