In this study, the authors agree with the assumption that consumers are able to describe products and specify their ideal. They
compare two different methodologies that can be used to analyze
JAR or Ideal Profile data, and which give guidance for product
improvement, based on the deviations from the ideal levels and
the relative importance of each attribute for liking. For the analysis
of JAR data, PLS on dummy variables can be used (Xiong & Meullenet, 2006). For the analysis of Ideal Profile data, a method based on
regression on principal components can be used (the Fishbone
method).