Choice models with mixtures: An application to a cocktail experiment

Abstract

Data from choice experiments are analyzed using choice models. Proper modeling allows a good understanding of the consumer preferences, the identification of consumer segments and the optimization of products, and, eventually, offers opportunities for competitive advantage. However, in the analysis of choice data, products that are mixtures of ingredients have been largely overlooked. In this paper, we combine traditional choice models and traditional mixture models. We apply the resulting mixture choice models to data from a real-life experiment in which consumers made pair-wise comparisons involving seven cocktails. For the choice models, we first assume consumer homogeneity. Next, we allow for heterogeneity among individuals. Therefore, we discuss the multinomial logit model, the mixed logit model and the latent class model. For identifying segments, besides the latent class model, we explore various two-stage segmentation approaches in which forces, hierarchical Bayes estimates and Firth individual-level estimates are used as input for a cluster analysis. The results show that the mixed logit model and the latent class model describe the data better than the multinomial logit model. The optimal cocktails for the segments we identified using the various methods are not identical, but show many similarities.

Bayesian I-optimal designs for choice experiments with mixtures

“…Third, we focused on the multinomial logit model, which assumes that there is homogeneity in the preferences of the respondents. However, as demonstrated by Courcoux and Séménou [15] and Goos and Hamidouche [16], this might be an unrealistic assumption. Therefore, it would make sense to extend the models and algorithms presented here to other types of models that take into account the possible presence of consumer heterogeneity.…”

“…Examples of mixture ingredients include the chemicals that are used to create a pesticide [11]; media used in advertising campaigns [12]; components of a mobility budget such as car with fuel card and public transport card [3]; cement, water, and sand to make concrete [13]; the wheat varieties used to bake bread [14]; and the ingredients used to make a drink such as mango juice, lime juice, and blackcurrant syrup [15,16]. In mixture experiments, the products and services under investigation are expressed as combinations of proportions of ingredients and the researchers' interest is generally in one or more characteristics of the mixture.…”

“…This means that the choice experiment comprised 60 × 8 = 480 choice sets, each of size 2. Goos and Hamidouche [16] described how to combine Scheffé models for data from mixture experiments with the logit type models typically used for choice experiments, and demonstrated the usefulness of the resulting combined models using the data from Courcoux and Séménou [15].…”

Research on Real Purchasing Behavior Analysis of Electric Cars in Beijing Based on Structural Equation Modeling and Multinomial Logit Model

“…In this paper, we distinguish the observed variables into positive and negative attributes. Through the establishment of the multinomial logit model [49][50][51][52][53][54][55], we further study the influence of various attribute factors on the actual purchasing behavior of electric cars.…”

Gendered fertility intentions and child schooling: insights on the quantity–quality trade-off from Ethiopia

“…The DCE is designed using a mixture-amount model. This model is highly appropriate to jointly disentangle the intention for different types of education (i.e., the mixture or quality) and the preferred family size (i.e., the amount or quantity), and is based on pioneering work of some authors [Raghavarao and Wiley (2009), Khademi and Timmermans (2012), Ruseckaite et al (2017), Goos and Hamidouche (2019), Zijlstra et al (2019)].…”