Hybrid choice models (HCM) incorporate latent psychosocial variables into choice modeling, offering a deeper understanding of consumer preferences. Unlike alternative two-step methods, HCMs estimate latent factor scores, representing psychosocial beliefs, along with preference parameters simultaneously. However, existing HCMs often overlook scale and preference heterogeneity, limiting their ability to capture the full complexity of consumer decision-making. This article introduces a Bayesian Hybrid Generalized Multinomial Logit (BHGMNL) model that addresses these gaps by simultaneously incorporating latent factors and accounting for both scale and preference heterogeneity.
The model is applied to consumer preference data for single-use eating-ware products, revealing its ability to capture behavioral insights.
The findings demonstrate that the BHGMNL model is a robust and flexible framework that offers significant advances in consumer choice modeling.
We proposed a new method, which helps with the choice experiment data.
