Development of EsSense25, a shorter version of the EsSense Profile®

Abstract

Emotion measurement has received increasing attention as reflected in the growing number of methods for measuring emotions in a product development context, including questionnaires. One such emotion profile, the EsSense Profile® (King & Meiselman, 2010) contains 39 emotions and has had increased use in product research. One of the practical challenges to its wider adoption is the length of the profile. We present results from 4 sets of studies designed to evaluate the semantic structure of the EsSense Profile®, to develop a shorter version of it, and to validate the shorter version, both conceptually and practically. A sorting methodology was used to sort the 39 EsSense Profile® emotions (printed on cards) into groups based on a similarity criterion, using two subject populations (n=121). A unique modification was implemented where subjects indicated from each sorted pile which word in that pile best described the sorted group. Hierarchical clustering was used to highlight the structure, and candidate words for removal were chosen, reducing the list to 25 words. Using both the EsSense Profile® and EsSense25 lists, a validation sorting study was performed (n=87). Confirmation studies conducted online (n=1728) and in CLTs (n=306) used the questionnaires in real-world scenarios, including food-name evaluations, brand evaluations and product tasting. Overall both lists performed similarly, but analyses revealed that there may be important context effects in which the meaning of words may change across product categories and across emotion lists. We finish with a discussion of why this might be and what this means to the sensory practitioner.