Comparison of different clustering methods for investigating individual differences using choice experiments

Different strategies for investigating individual differences among consumers in choice experiments using are compared. The study is based on a consumer study of iced coffees in Norway. Consumers performed a choice a choice experiment using iced coffee profiles varying in coffee type, production origin, calorie content and price. Choice data will be analysed using two different clustering strategies: the Latent Class Model, and Mixed Logit Model combined with Principal Component Analysis for visual segmentation or with automatic clustering detection using Fuzzy Clustering. The different approaches are compared in terms of data analysis methodologies, outcomes, flexibility, user friendliness and interpretation.