内田・クレペリン検査の因子得点判定法の因子的妥当性

東村 高良

doi:10.4992/jjpsy.47.30

FACTORIAL VALIDITY OF THE FACTOR SCORE METHOD IN THE UCHIDA-KRAEPELIN PSYCHODIAGNOSTIC TEST

TAKAYOSHI HIGASHIMURA

Author information

JOURNAL FREE ACCESS

1976 Volume 47 Issue 1 Pages 30-39

DOI https://doi.org/10.4992/jjpsy.47.30

Details

Published: May 20, 1976 Received: July 14, 1975 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 : PURPOSE: The purpose of this study was to reveal the factorial structure of the Uchida-Kraepelin Psychodiagnostic Test, in which testee's speed of addition (per min) through 30min's work period with 5-min's intermislion in the middle is plotted as a work curve or a profile. Tsujioka and Higashimura (1968, 1974a, b, c, d) had established a factor score method to evaluate the functional units in this test according to the factor structure consisting of a trait factor (Calculation Speed) and five state factors (General Energy Level, Fatigue, Excitation, Periodic Rhythm (of 10min for a periodic time) and Volitional Tension).
METHOD AND RESULTS: 11 Factor analyses were performed using the same technique as by Tsujioka (1968). 11 correlation matrixes of order 30 based on 11 samples in Table 3, in which every min's speed of performance (the number of additions) was correlated with every other, were analyzed into 11 centroid factor matrixes of 30×6. The number of factors to be extracted in each factor analysis was determined by the Scree Test, the Scree Graph and the Strata Graph designed by Tsujioka and Higashimura (1975) and the communalities were estimated by the iterative centroid method. The centroid method instead of the principal factor method was used in order to extract the “State Factors” placing equal weight upon each min's performance. As the result from the largest sample (No. 8 in Table 3: N=2931) showed the same sharp break on the scree between the 6th and the 7th eigenvalue as in Tsujioka's analysis (1968), although some of other samples did not indicate clear breaks, all correlation matrixes were analyzed using the 6-factor model. Only two of them are shown in Tables 1 and 2. The coefficients of congruence among 72 factor-loading vectors (6 factors by 11 analyses plus Tsujioka's original 6 factors) were calculated by Formula (4) but only the coefficients corresponding to Tsujioka's factors are shown in Table 3 where the factor matchings are very clear except for some cases of Periodic Rhythm (the fifth factor). Resides congruence coefficients, the correlations among 72 factor-estimates were calculated from the total sample (N=2931) but only the correlations corresponding to Tsujioka's are shown in italic in Table 3 (each lower rows) which indicate the similar pattern of factor matching as in the coefficients of factor congruence. The principal component analysis with Varimax Rotation was applied to the correlation among the factor-estimates based on different evaluation procedures (using mean and standard deviation vectors and also weight matrix of each sample), and the result showed clearly the clusters of the matching factor-estimates. In Fig. 2 the configuration of the factor-estimates in this Varimax Solution was shown.
DISCUSSION: The feasibility of transforming Tsujioka's 6-factor model in to another factor model by Procrustes Rotation was discussed (see Table 4). Next, the relationship between Tsujioka's factor-analytic model and the spectrum model of Fourier analysis by Kashiwagi, Nakatsuka and Okumoto was considered. Finally, the principal component analysis followed by Varimax of the correlation among the factor-estimates obtained by different evaluation procedures was discussed as a useful tool in the future study for testing identity or transferability of faitor scores.
Right : The factorial structure consisting of a trait factor and 5 state factors (Calculation Speed, Energy Level, Fatigue, Excitation, Periodic Rhythm and Volitional Tension) found by Tsujioka (1968) in the work curves of the Uchida-Kraepelin Psychodiagnostic Test was tested by the factor analyses of 11 correlation matrixes based on various subject samples (Table 3), in terms of coefficients of congruence among the factor loadings by Formula (4) and correlations among the factor-estimates by Formula (1); and also by the principal component analysis with Varimax rotation of the correlation among the factor-estimates which were obtained by various evaluation procedures. The results revealed the remarkable similarity with the factors found by Tsujioka, except for a partial inconsistency in Periodic Rhythm Factor. The transformation from Tsujioka's 6-factor model to the spectrum model of Fourier analysis by Kashiwagi, Nakatsuka and Okumoto was discussed.

Abstract

The factorial structure consisting of a trait factor and 5 state factors (Calculation Speed, Energy Level, Fatigue, Excitation, Periodic Rhythm and Volitional Tension) found by Tsujioka (1968) in the work curves of the Uchida-Kraepelin Psychodiagnostic Test was tested by the factor analyses of 11 correlation matrixes based on various subject samples (Table 3), in terms of coefficients of congruence among the factor loadings by Formula (4) and correlations among the factor-estimates by Formula (1); and also by the principal component analysis with Varimax rotation of the correlation among the factor-estimates which were obtained by various evaluation procedures. The results revealed the remarkable similarity with the factors found by Tsujioka, except for a partial inconsistency in Periodic Rhythm Factor. The transformation from Tsujioka's 6-factor model to the spectrum model of Fourier analysis by Kashiwagi, Nakatsuka and Okumoto was discussed.

Corresponding author

Correction information

Register with J-STAGE for free!