A COMPARISON OF SEVEN PREFERENCE MAPPING TECHNIQUES USING FOUR SOFTWARE PROGRAMS

作者：Yenket Renoo; Chambers Edgar^*; Adhikari Koushik

来源：Journal of Sensory Studies, 2011, 26(2): 135-150.

DOI：10.1111/j.1745-459X.2011.00330.x

Various methods are used to create a preference map based on different theories to model and analyze relationships between product descriptors and consumer preferences. Several programs offer solutions to understand influences and relationships between a descriptive sensory profile and consumer liking, but researchers need to know the (dis)advantages and (dis)similarities of these programs. This study compares the advantages and disadvantages of four statistical software programs and seven multivariate techniques for three empirical studies: milk, paint and fragrance. Internal and external preference mapping are included in this study. No multivariate method consistently generated a high percentage of consumers who mapped closest to their most-liked products. Also, neither the variance nor the complete solution was explained among descriptors and consumers. Of particular interest is that the high percent variance explained in descriptors and/or consumers did not guarantee a preference mean vector that was close to the most-liked product. For uncomplicated data, researchers can use any method/program to create a preference map when consumer data are highly homogenous in product liking patterns. XLSTAT PLS2 (Addinsoft, New York, NY) and Unscrambler PLS2 (passified; CAMO Software Inc., Woodbridge, NJ) are recommended for less homogenous consumer data. For complex heterogeneous data, multidimensional preference analysis is recommended for understanding consumer preference.
PRACTICAL APPLICATIONS
This research provides comparisons of various perceptual mapping techniques and statistical computer packages for creating preference maps. Results from the study show that when the interpretation of maps is implemented visually without supportive underlying findings from the original data and basic analyses, preference maps may be misleading. Such use appears common in the literature suggesting that analysis using preference mapping techniques is either overused or inappropriately used in some cases.

• 出版日期2011-4