Sammendrag
This master thesis reviews and discusses the literature on discrete choice methods. Discrete choice modelling is an effective tool for predicting choice behavior, analyzing consumer preferences, and determining consumption patterns and market trends. Nevertheless, increasing diversity in markets, customer preferences, and available products require more customized models that have high predictive power in different choice situations and behaviors. This also applies to the study of food choice. This thesis assesses the available methods from the point of view of statistics. It discusses the usefulness, strengths, and weaknesses of some of the most widely used discrete choice models, with a special focus on seafood preferences and food choice. We use survey data on the French seafood market for empirical assessments of the models, in order to determine the best model for predicting choice probabilities. By doing so, this thesis aims to improve the modelling of seafood preferences in particular, and of food choice in general.