Comment analysis of consumer's likes and dislikes as an alternative tool to preference mapping: A case study on apples

Abstract

This study compares the analysis of consumer's comments resulting from a hedonic test as an alternative to the traditional internal preference mapping. During a consumer test, 87 apple consumers first evaluated six different Golden apple batches on a hedonic scale and then answered to the non-mandatory open-ended questions stating separately what they liked and disliked from each batch. In parallel, an expert panel described the sensory profiles of the studied products. To compare the results obtained by the two studied methods the RV coefficient was calculated and was found to be 0.8656 (p= 0.011). Therefore, the information obtained by the comment analysis of likes and dislikes was similar to that resulting from sensory characterization done by the trained panel. With both methods, crunchiness and sweetness appeared as main sensory preference key drivers, while mealiness was not appreciated. At the same time, some characteristics such as juiciness appeared important for consumers but it was not a significant discriminant attribute for the trained panel. A new method, the Chi-square per cell, was used to deeply analyze the contingency table of the main modalities used by consumers allowing the identification of the significant modalities which described each apple liking. Finally, the distinction between likes and dislikes made the transcription of consumers' opinions easier, without a need of interpretation on behalf of the transcoder.