Free JAR experiment: Data analysis and comparison with JAR task

Abstract

The Free JAR protocol is introduced as a methodology for highlighting product improvement keys without imposing specific attributes to the respondents. Being relatively new, there is a need to demonstrate its relevance compared to an established sensory methodology such as the JAR procedure. The first aim of this study is to outline a strategy for the statistical analysis of Free JAR comments. This consists in extracting all the attributes assessed according to modalities such as “not enough”, “JAR” and “too much”, and recoding the Free JAR comments of each respondent into a products × attributes matrix. At a product level, an analysis is performed with the aim of highlighting the attributes that characterize the product under study. Thereafter, a multivariate descriptive analysis is presented. It consists in calculating a dissimilarity measure for each pair of products. Multidimensional Scaling (MDS) is performed on the dissimilarity matrix thus formed in order to obtain a graphical display that depicts the relationships among the products. For a comparison purpose, the same analyses are applied to JAR data. The findings based on a case study pertaining to cheeses make it possible to assess the respective shortcomings and benefits of Free JAR and JAR procedures.