The questionnaire test using the rating scale data is applied effectively, since it features relatively high reliability and the simple processing. the semantic structure analysis (SS analysis) is proposed as the method of structural analysis for the rating scale data. In the SS analysis, the order of the items is considered, and the human behavior or the psychological aspects are interpreted from the ordering structure of the items. In the conventional questionnaire test using the rating scale data, the evaluation is possible in many cases even if the data are composed of a discrete single point. Depending on the questionnaire items, however, it may happen that the evaluator is not convinced of a discrete single point. From such a viewpoint, this paper proposes a structural analysis considering the ordering among items which is not a discrete point evaluation to evaluate by a single discrete point, but is an interval evaluation to evaluate by an interval on the rating scale. For this purpose, the difference between the ratings of two items is quantified anew based on the interval rating data. the inversely ordered rating ratio and the ordering coefficient are defined as the indices for the ordering. It is shown that the method proposed in this paper is an extension of the SS analysis. Furthemore, as the structural analysis, the ordering structure graph is constructed for the set of items. A method of constructing the ordering structure graph is presented, and the application process is shown through an actual example.
|ジャーナル||Electronics and Communications in Japan (Part III: Fundamental Electronic Science)|
|出版ステータス||Published - 1995 10|
ASJC Scopus subject areas