コンピュータ・シミュレーションによる内田・クレペリン作業曲線の合成と因子分析による分解

東村 高良; 辻岡 美延

doi:10.4992/jjpsy.45.198

FACTORIAL SYNTHESIS AND ANALYSIS OF THE UCHIDA-KRAEPELIN PSYCHODIAGNOSTIC TEST BY COMPUTER SIMULATION

TAKAYOSHI HIGASHIMURA, BIEN TSUJIOKA

Author information

JOURNAL FREE ACCESS

1974 Volume 45 Issue 4 Pages 198-208

DOI https://doi.org/10.4992/jjpsy.45.198

Details

Published: October 10, 1974 Received: May 10, 1974 Available on J-STAGE: July 16, 2010 Accepted: - Advance online publication: - Revised: -
Correction information
Date of correction: July 16, 2010 Reason for correction: - Correction: ABSTRACT Details: Wrong : The first purpose of this study is to check the isomorphism between the observed and the simulated data of the Uchida-Kraepelin Psychodiagnostic Test. The simulated data consists of 11 groups of 300 Ss, while the observed data is derived from the same number of college students, as reported by Tsujioka and Higashimura (1974a).
Synthetic scores are generated by computer, using both normal random deviates of the zero mean and the unit variance of factor-scores of five state factors (i.e., General Energy Level, Fatigue, Excitation, Periodical Rhythm and Volitional Tension) and those of thirty error factors in the Formula 1. However, the standard deviation of the factor-score of Calculation Speed was reduced in size from 1.00 to 0.00 at the interval of 0.10 to generate 11 variations, as shown partly in Tables 2, 3 and 4.
Eleven correlation matrices were analyzed by the same method as in the analysis of the observed data. That is, after determining the number of factors to be extracted by Cattell's Scree Test, six centroid factors were obtained, using the iterative communalities as shown in Table 4. The patterns of factor configuration are found very similar across the corresponding factors, as shown in Fig. 4.
The second purpose of the study is to examine validity of the criteria for the correct number of factors to be extracted in factor analysis. Since the number of contributing factors in our simulated data was determined in advance, it is clearly shown which criterion, the Cattell's or Guttman's, is superior, as displayed in Fig. 3.
The absolute size of eigenvalues is relative to the contributions of common factors, but graphs of the eigenvalues ranked in order of the size do not show changes, the error factors being characterized as a straight line as displayed in Fig. 3. It appears clear that the results of our study strongly support the criterion suggested by Cattell.
Right : 11 simulated correlation matrices of 30 composites, which were equivalent to the observed 30-minutes' performances of the Uchida-Kraepelin Test, synthesized by substituting 36 sets of normal random deviates of 0 mean and unit variance into 6 common factors and 30 error factors of the factor-analytic model of the test (Formula No. 1), were re-factoranalyzed. Complete isomorphism between the observed and the simulated data was found in the means, standard deviations, correlations, eigenvalues and factor-loadings and also in the weights for factor-scores. The results seem to give a clear-cut answer to the determination of the number of factors. Cattell's Scree Test is found superior to Guttman's criterion which uses eigenvalues of the correlation matrix with unities in the diagonal.

Abstract

11 simulated correlation matrices of 30 composites, which were equivalent to the observed 30-minutes' performances of the Uchida-Kraepelin Test, synthesized by substituting 36 sets of normal random deviates of 0 mean and unit variance into 6 common factors and 30 error factors of the factor-analytic model of the test (Formula No. 1), were re-factoranalyzed. Complete isomorphism between the observed and the simulated data was found in the means, standard deviations, correlations, eigenvalues and factor-loadings and also in the weights for factor-scores. The results seem to give a clear-cut answer to the determination of the number of factors. Cattell's Scree Test is found superior to Guttman's criterion which uses eigenvalues of the correlation matrix with unities in the diagonal.

Corresponding author

Correction information

Register with J-STAGE for free!