2012 Fiscal Year Final Research Report
Multivariate analysis methods for frequency-valued data
Project/Area Number |
23653233
|
Research Category |
Grant-in-Aid for Challenging Exploratory Research
|
Allocation Type | Multi-year Fund |
Research Field |
Experimental psychology
|
Research Institution | The National Center for University Entrance Examinations |
Principal Investigator |
MIYANO Hisao 独立行政法人大学入試センター, 研究開発部, 教授 (90200196)
|
Co-Investigator(Kenkyū-buntansha) |
ARAI Sayaka 大学入試センター, 研究開発部, 助教 (00561036)
|
Project Period (FY) |
2011 – 2012
|
Keywords | 多変量解析 / 尺度化 / 対応分析 / 多重対応分析 / 区間値データ |
Research Abstract |
In psychological studies, the mean values of responses are often treated as a main concern of statistical data analysis, neglecting the variations of responses. In this study, we addressed the issue of multivariate analysis methods for analyzing the frequency distribution data of responses, and, utilizing the idea of mixture distribution, developed a principal component analysis method for interval-valued data, and a scaling method of successive categories for contingency table. Also the latter method was extended to contingency table with more than two variables In psychological studies, the mean values of responses are often treated as a main concern of statistical data analysis, neglecting the variations of responses. In this study, we addressed the issue of multivariate analysis methods for analyzing the frequency distribution data of responses, and, utilizing the idea of mixture distribution, developed a principal component analysis method for interval-valued data, and a scaling method of successive categories for contingency table. Also the latter method was extended to contingency table with more than two variables
|