S. Schwartz の概念枠組みにもとづく価値観の国際比較：ドイツと日本における「大学生調査」のデータ分析

This paper is a sequel to our previous paper with the same title (Kwansei Gakuin University School of Sociology Journal No.133). In the previous paper, we reported the descriptive results of student surveys in Germany and Japan focusing on frequency distributions for value question items. Based on those results, the current data analysis explores the following: 1. The internal reliability (or consistency) of the question items in each value type (or scale). We used Cronbach’s alpha to determine the extent to which students answered the question items in the same way. 2. The measurement equivalence (or invariance) of each value scale. Measurement equivalence implies that using the same value scale in different countries measures the same value type in the same way. (1) The most commonly used technique for testing cross-national measurement equivalence is multi-group confirmatory factor analysis (MGCFA). (2) Like all scientific techniques, MGCFA has limitations. Therefore, we supplement MGCFA with smallest space analysis (SSA). The results of our data analyses are as follows: 1. If we set Cronbach’s alpha to 0.6 as the lower limit for an acceptable reliability, two value scales in Japan -Achievement (0.396) and Humility (0.387)-remain far below the limit. Apart from these two cases, the remaining Japanese scales and all German scales reach the acceptable level of reliability. 2. The results of MGCFA show that four of the value types ―Tradition, Universalism Concern, Face, and Benevolence Dependability-reach the level of full scalar equivalence, and the remaining value types reach the level of partial scalar equivalence. 3. We tried to confirm the comparability-i.e., measurement equivalence-of the value items and scales using SSA. The German SSA map shows a better fit with the Schwartz value model than the Japanese one.