제품속성간의 계층구조와 연관성 파악을 위한 기법 = 비대칭형 다차원척도법의 응용

다국어 초록 (Multilingual Abstract)

Product attributes are hierarchically structured. The concrete ones are at the bottom, and the more abstract at the top. Successful positioning of a product requires the knowledge about how the concrete attributes inherent in the product associate with the abstract values that consumers seek. For this, marketers have used the laddering or hierarchical clustering techniques so as to identify the hierarchical structure of attributes. Since the analysis using laddering techniques depends on subjective judgment and heuristics, however, its outcome is not objective enough for general acceptance. Hierarchical clustering approach also mislays the important information about causal relationship among attributes in the process of data transformation. In this paper, we developed an asymmetric multidimensional scaling method which can simultaneously identify the horizontal and vertical structure of attributes. We applied this method to analyze the underlying structures of beer attributes, and compared its outcome with those of the conventional laddering and clustering techniques. This new method would be superior to the traditional method in systematically and objectively investigating the horizontal and vertical relationship among attributes.