Abstract
In this study, we deal with a method to specify simultaneously both the whole Asian path diagram and the diagrams which express the differences between the countries with regard to the model of second-order factor analysis for brand evaluation. Utilizing “Brand Asia”, a survey project investigating brand images of 60 global brands over 12 countries in Asia, we also verify its efficacy. Although the conventional methods allow us only to analyze either the overall Asia or the differences between the countries, the Bayesian approach enables us to analyze both simultaneously. Besides, by interpreting the hyperparameters of prior distribution aggressively, it also enables us to clarify the proportion of the university overall Asia to the diversity in each country and the unstable indexes for the model.
