2017 Fiscal Year Final Research Report
Psychometric study on perfect simple structure principal component analysis and its applications to scale construction for measuring individual diffrences
| Project/Area Number |
15K04197
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Experimental psychology
|
|---|
| Research Institution | Chukyo University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2015-04-01 – 2018-03-31
|
| Keywords | 主成分分析 / 直交回転 / 斜交回転 / 多重対応分析 / クラスター分析 / 単純構造 / 心理測定尺度 / 調査データ |
|---|
| Outline of Final Research Achievements |
Multiple correspondence analysis is often called principal component analysis for categorical data. This project was aiming at making the statement hold literally. More concretely, a method named orthogonal polynomial principal component analysis was developed, where each category was quantified using number of categories minus one centered and orthonormal weight vectors. Astonishingly enough, it was shown that the weight vectors are entirely arbitrary, Owing yo the arbitrariness, two ways of procedures was employed; one is the coding of categories by orthogonal polynomials, the other is the quartimax rotations of the weight matrix from both sides. The method was applied to various psyhometirc and survey data, and several new findings have been obtained.
|
| Free Research Field |
計量心理学
|
