内田・クレペリン検査の因子得点判定法と直観判定法との併存的妥当性

辻岡 美延; 東村 高良

doi:10.4992/jjpsy.46.333

CONCURRENT VALIDITY OF THE FACTOR SCORE METHOD WITH THE INTUITIVE RATING OF THE UCHIDA-KRAEPELIN PSYCHODIAGNOSTIC TEST

BIEN TSUJIOKA, TAKAYOSHI HIGASHIMURA

Author information

JOURNAL FREE ACCESS

1976 Volume 46 Issue 6 Pages 333-342

DOI https://doi.org/10.4992/jjpsy.46.333

Details

Published: March 31, 1976 Received: July 30, 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: Concurrent validity between an objective analytical and an intuitive rating method is examined for individual work curves of the Uchida-Kraepelin Psychodiagnostic Test. The objective method was designed by Tsujioka on the basis of his past factor analytic studies (Tsujioka, 1952, 1968; Tsujioka & Higashimura, 1974a, b, c; Higashimura & Tsujioka, 1974). In the present study, a total of 8 variables-6 factor scores (Calculation Speed, General Energy Level, Fatigue, Excitation, Periodic Rhythm and Volitional Tension) and either the Model Score (MS) and the Anti-Model Score (AMS), or the Expected Model Score (EMS) and the Expected Anti-Model Score (EAMS)-were used so as to evaluate the degree of deviation from “the Normal State Curve” defined by Uchida who established the intuitive rating method.
METHOD: The present authors distinguished between two kinds of approach. One is an inductive method in which the multiple correlation method was used to predict Uchida's intuitive rating scores [a criterion variable assigned by 1 (the best normal state curve) through by 5 (the worst abnormal state curve)] from the 8 predicting variables (in addition to the 6 factor scores, either MS & AMS, or EMS & AEMS). The other is a deductive method of validity, by which coefficients of concurrent validity would be obtained after setting three principles concerning the amount of performance, the level of the state factor scores and that of the Anti-Model Score for evaluating the individual curves, according to the psychological hypotheses on the work process of this test.
RESULT: First, in the multiple correlation method, it was revealed that coefficients of concurrent validity ranged from 0.754 to 0.785 and that in the cross-validation studies where the weight vector from one sample was applied to the other sample, the reduction of coefficients of concurrent validity did not occur as Shown in Fig. 1.
Secondly, the weight vectors in the multiple correlation analyses revealed that the “Normal State Curve” intuitively defined by the Uchida's school tends to be the curve of higher score in Calculation Speed, higher score in General Energy Level, higher in Fatigue and higher in Volitional Tension than each average. This means that Uchida's Normal State Curve is not the average performance curve from the view points of both the amount of performance and the state factors. This fact results the reduction of the concurrent validity in our deductive approach as shown in Table 5.
DISCUSSION: Finally, use of Mahalanobis' generalized distance of the above 8 measures was proposed for evaluating the individual curves in the future study in connection with Tsujioka's Normal State or Uchida's Normal State, because the individual test results of the Uchida-Kraepelin Psychodiagnostic Test should be evaluated according to the probability of occurence from the Normal State or other certain points in this generalized distance space which would be defined by the studies of practical validities much more vital to the validity of the test itself.
Right : Concurrent validity was examined between an objective analytical method and Uchida's intuitive rating. To evaluate objectively the degree of deviation from “the Normal State Curve” defined by Uchida, the present authors used the two pairs of scores: “Model, Anti-Model or Expected Model and Expected Anti-Model Scores” which were defined by Tsujioka on the basis of his factor analytic model. The concurrent validity between the 2 methods was verified by two approaches: Inductive and deductive. The former was the multiple correlation method while the latter was a method by means of which the coordinates of individual curves plotted on a 2-dimensional orthogonal axes were categorized into several groups. The results show a relatively high degree of concurrent validity.

Abstract

Concurrent validity was examined between an objective analytical method and Uchida's intuitive rating. To evaluate objectively the degree of deviation from “the Normal State Curve” defined by Uchida, the present authors used the two pairs of scores: “Model, Anti-Model or Expected Model and Expected Anti-Model Scores” which were defined by Tsujioka on the basis of his factor analytic model. The concurrent validity between the 2 methods was verified by two approaches: Inductive and deductive. The former was the multiple correlation method while the latter was a method by means of which the coordinates of individual curves plotted on a 2-dimensional orthogonal axes were categorized into several groups. The results show a relatively high degree of concurrent validity.

Corresponding author

Correction information

Register with J-STAGE for free!