In survey analysis with many questions, the complex correlations between items make it difficult to determine which items should be prioritized after the analysis. This study proposes a method for selecting important items using composite factor loadings to address this challenge. Multiple outcome-related items are summarized using principal component analysis, with the first principal component (ZY1) used as the outcome variable. On the other hand, since there are many causal questions, they are grouped using cluster analysis based on correlation structure. Next, multiple regression analysis of variable selection, with ZY1 as the dependent variable and the principal components of each group as independent variables, extracts principal components that have a significant impact on the outcome variable. However, it is not possible to directly address the principal components themselves, and it is difficult to determine important items when multiple principal components are selected from the same group. Therefore, this study presents a method for selecting important items using composite factor loadings that integrate the information of the selected principal components. All of the analytical methods used in this study, including principal component analysis, cluster analysis, and multiple regression analysis, can be executed using JMP. This method can be used as a practical approach to support priority-oriented selection of important items in survey analysis with many questions.
