A method for determining the number of common factors by cross validation

Research output: Contribution to journalArticle

Abstract

The purpose of this study is to propose a criterion for determining the number of common factors. This criterion is called MSCV (Model Selector by Cross Validation), which is an estimator of expected log-likelihood. Through the results of computer simulations, the criterion is shown to be effective for determining the number of common factors in factor analysis. The practical application with such criterion is demonstrated by using standardized data of Japanese WISC-R, made of three common factors.

Original languageEnglish
Pages (from-to)252-256
Number of pages5
JournalJapanese Journal of Educational Psychology
Volume34
Issue number3
DOIs
Publication statusPublished - 1986
Externally publishedYes

    Fingerprint

Keywords

  • Computer Simulation
  • Cross Validation
  • Expected Log-Likelihood
  • Factor Analysis
  • MSCV

ASJC Scopus subject areas

  • Education
  • Developmental and Educational Psychology

Cite this