Modeling of Relationship between Mental Accounting and Purchase Behaviors: Proposal of Hierarchical Bayesian Threshold Poisson Regression Model

Abstract

Our research aims to model mechanisms to describe consumer purchase quantity on a single shopping trip with consideration of consumer sentiment called mental accounting. The model consists of two components; (1) consumers mental loading that changes at each shopping trip, and (2) consumers purchase quantity generating mechanism, which switches regimes depending on degrees of the mental loading and threshold parameter individually estimated. The model is derived based on threshold Poisson regression model framework, and the total model parameters sets are estimated using Markov Chain Monte Carlo (MCMC) mehods. Empirical studies have been applied with ID-POS data from a retailer shop. Results indicate that our model outperforms by having consumer sentiment considered such as mental loading when they make purchase decisions. In addition, the model reveals individual payday as by-product through estimation of each mental loading conditions.