サブグループ同定法による比較調査データの相違点に対する影響要因の探索

-東京と成都の街路景観満足度評価データへの応用-

抄録

To evaluate differences between countries/regions, the classical statistical method is by evaluation of individual question items using hypothesis tests. They cannot be used, however, in surveys in that have a response (outcome). Regression analysis is another statistical approach, but cannot be used to evaluate the influencing factors in the difference in response between countries/regions. Here, we focus on statistical methods for finding differences in international comparison data, namely subgroup identification methods. We examine the usefulness of subgroup identification methods using data on streetscape satisfaction from surveys conducted in Tokyo, Japan and Chengdu City, China.