Establishing Metrics for Kansei Responses: An Approach Using the Rasch Model.

Although qualitative comparisons are necessarily part of the process to elicit users’ kansei responses to products, they are insufficient to provide a more fine-grained interpretation of users’ interaction, which can require a measurement system. However, kansei variables cannot be measured directly. Data obtained from kansei responses need to be transformed by statistical methods and meet measurement assumptions. An approach to validate the quantitative structure of kansei scales is the application of Rasch measurement theory. The Rasch model, which is referred to as a family of probabilistic models, provides mechanisms to test the hypothesis that the observations meet the assumptions for establishing a quantitative structure. In this paper a number of procedures in Rasch modelling are outlined. Different examples from empirical applications using some techniques of kansei engineering show that the establishment of measures for comparisons between individuals and between stimulus objects is not a trivial matter.