Testing Validity of Probability-One Assumption in MS-Based Brand Choice Model
摘要
The Mover-Stayer (MS) model, originally introduced by (Blumen et al. 1955), has served as a foundational statistical framework to achieve several research purposes in various academic disciplines. This study examines the validity of the Probability-One assumption embedded in the Mover-Stayer (MS) model and its impact on brand choice analysis. Probability-One assumption, which fixes the transition probability of stayers at zero, leads to biased parameter estimates by underestimating potential transitions and overfitting movers’ switching behaviors. By challenging this assumption, our findings highlight the need for a more flexible framework that better accounts for consumer heterogeneity and real-world purchasing dynamics. The results contribute to advancing MS-based brand choice models by improving their capacity to reflect actual consumer behavior.