new sensory findings
books & publications
contact us

ESN Seminar 2007
ESN Seminar 2005
>Programme and Abstracts
>Our speakers
>Press release May 30
>Press release May 18, 2005
>Nota de Prensa Mayo 26, 2005
>Nota de Prensa Mayo 25, 2005
>Nota de Prensa Mayo 18, 2005

Preference mapping: methods and applications


Pascal Schlich

Laboratory of Interface between Research, Industry and Sensometry (LIRIS)

at the European Centre for Taste Sciences (CESG)

DijoN (France)


Preference mapping is a set of statistical techniques aiming to explain consumer preferences for a set of competitive products by the sensory profiles of these products drawn from a trained panel of assessors. It has been widely used in the food industry to help formulating new products tailored to specific consumer segments.


The paper will present the basics of preference mapping (multiple polynomial regressions), discuss some limitations of the classical techniques and thus propose the PrefMaX technique to overcome them. Two PrefMaX applications of peculiar interest for this seminar will be described. The first one aimed to optimize the feeling of the efficiency of the braking system of a car by its driver and was, to our knowledge, the first application of sensory analysis in the car industry.


The second application aimed to draw a segmentation of sensory preferences for tomato in France. Preference mapping of such fresh products leads to specific problems which will be discussed. Among them is the use of incomplete design where each consumer does not rate every product. This is possible using specific design based on Mutually Orthogonal Latin Square (MOLS). The tomato study was first run with a complete design in 1999, then using an incomplete one in 2001. The segmentations found were consistent over the two years in terms of both their sensory meaning and their socio-demographic and attitudinal interpretation.


 Slides                                                                                         Back to programme

  top of the page