情報量にもとづく“層別因子構造”の提案

村上 宣寛

doi:10.4992/jjpsy.47.308

A PROPOSAL OF “STRATUM FACTOR STRUCTURE” ON INFORMATION MEASURES

YOSHIHIRO MURAKAMI

Author information

JOURNAL FREE ACCESS

1977 Volume 47 Issue 6 Pages 308-315

DOI https://doi.org/10.4992/jjpsy.47.308

Details

Published: March 20, 1977 Received: February 04, 1976 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 purpose of this study is to validate the hypotheses of “stratum factor structure” in personality ratings, which is a new concept proposed by the author and defined as follows: The average amount of information (bit) of every scale for each stimulus is calculated and converted into means for each scale. Next, poinits (point unit) 2^H are obtained. Then the scales are put in order according to the largeness of entropy, and can be divided into n groups. They are analyzed by design 2 (Heise, 1969) and n factor structures are obtained. The ordered assembly of this structure S={S₁, S₂, …, S_n} is called by the name of the “stratum factor structure”. The basic presupposition of this concept is that the scales of smaller between-subjects variance (entropy) represent more general and affective sides of the meaning and that the scales of larger variance represent less general and denotative sides of the meaning. And it will be expected that the lower levels of the “stratum factor structure” must be the stimuli-content which is represented by the Osgoodian E-P-A structure and that the higher levels must be the subjects-content which was called the denotative structure by Kuusinen (1969). Much of the labyrinthes of the SD results were solved by this concept.
TAT No. 1 was rated by 167 subjects on 132 monopolar scales, in which 6 groups were distinguished by the poinit numbers and principal component analysis and geomax rotation were applied. Main results were indicated in Table 1 and the interpretations were done in frame of E-P-A structure, Cattell's 16 PF test and Y-G personality questionnaire. In level 1 (4.2 poinit), the included scale is “talkative” only and the scale indicates the general agreement that he is not so talkative. In level 2 (4.6-4.8 poinit), 5 scales were included and the “thoughtful-thoughtless” (E′) factor was obtained. In level 3 (5.1-5.5poinit), 22 scales were included and “impulsiveness” (E-), “high energy level” (P), and “activity” (A) factors were obtained. In level 4 (5.6-6.0 poinit), 39 scales were included and “schizothymia”, “affectothymia” (A′), “dependent behavior”, “verbal activity” (A″), and “self-sufficiency” factors were obtained. In level 5 (6.1-6.5 poinit), 51 scales were included and “superego strength”, “deficiency of mental sensivity”, “tough-minded vs. tender-minded”, “lower ego-strength”, and “self-confidence” factors were obtained. In level 6 (6.6-6.8 poinit), 14 scales were included, and “timid-carefree”, “radical-real”, and “intuitive-careless” factors were obtained.
It is clear that the hypotheses derived from the “stratum factor structure” were supported, because in lower levels (level 2-3) the typical E-P-A structure appeared, though the included scales were not the typical SD scales, and because in higher levels (level 4-6) the denotative structures in personality ratings mentioned by Kuusinen emerged. These exciting results evoke the world-wide disputes on semantic differentia lmethod as pseudo-problems, which had originated from the ignorance of between-subjects variances. Thus, it is concluded that the E-P-A structure is general since the included scales were ordinarily small in between-subjects variances or in entropies that is the generality of this structure is due to the smallness of scale variances, and that the re-examinations of the personality factors is surely needed.
Right : The concept of “stratum factor structure” proposed by the author revealed the labyrinthes of SD results and predicted that in lower levels (structures by less entropy scales) the typical E-P-A structure will appear and in higher levels the denotative structures will appear. This hypothesis was validated by using TAT No. 1 which was rated by 167 Ss on 132 monopolar scales. The E-P-A structure appeared in level 2-3, and the denotative structures in level 4-6. Much of the personality factors corresponded with Cattell's 16 PFT.

Abstract

The concept of “stratum factor structure” proposed by the author revealed the labyrinthes of SD results and predicted that in lower levels (structures by less entropy scales) the typical E-P-A structure will appear and in higher levels the denotative structures will appear. This hypothesis was validated by using TAT No. 1 which was rated by 167 Ss on 132 monopolar scales. The E-P-A structure appeared in level 2-3, and the denotative structures in level 4-6. Much of the personality factors corresponded with Cattell's 16 PFT.

Corresponding author

Correction information

Register with J-STAGE for free!