The present study was designed (1) to classify high school students according to the weighted scores of a 25 aptitude test battery and their achievements on the seven subjects, by the method of principal component analysis, and (2) to investigate whether or not high and low achievers could be discriminated on the basis of the linear discriminant function of 10 aptitude test scores in this battery.
Twenty hundred and three high school students of technology answered the items of the aptitude test battery which consisted of following subtests:(1) general and vocational aptitude test (10 scales), 2) interest test (8 scales),(3) extroversion v. s. introversion test (6 scales),(4) work aptitude test (1 scale). In addition to these items, their achievement records on the following seven subjects were included: Japanese Language, Foreign Language, Social Science, Natural Science, Mathematics, Professional Course, Drafting.
In the first part of the study, the principal component analysis was applied to these data (32 variables about 203 S's, in total). As a result, six main components were extracted which represented about 60% of the total variation of the original data. The meanings of the first three were scientific thinking, general aptitude, and scientific v. s. literary interest, respectively. Based upon each of these component scores, classification and total evaluation of the individual student were established.
In the second part of the study, ten aptitude variables (verbal and mathematical ability, scientific v. s. literary interest, work aptitude, and six extroversion v. s. introversion scales) were chosen as predicton from the original item pool which seemed to have positive relations with the seven achievement variables as a result of the principal component analysis, and 203 S's were grouped into the upper and lower achiever classes based upon their ratings of each seven achievement variable.
Applying the linear discriminant analysis to these data (ten aptitude variables, two achiever groups in each seven subject), the following results were confirmed:(1) the discriminant function obtained had high predictive power for the achivements on Mathematics and Professional Subject, but had lower discriminability for Japanese Language and Social Science, and (2) the direction of discriminant coefficients indicated that extroversion v. s. introversion variable had minus (-) weight, and work aptitude had plus (+) weight in all subjects.

View full abstract

The first purpose of this study was to analyze the evaluative contamination in semantic differential (SD) scale. Osgood referred the dominance of evaluative factor of semantic space dimensions. It seemed that this was mainly due to the evaluative property of SD scale. Every qualifier often used in the description of certain personality traits has not only attributive property but evaluative property, so SD scale obtained by making a pair out of these qualifiers has also evaluative property. Thus two basic dimensions- & lsquo;evaluation & rsquo; and each & lsquo;attribute & rsquo;(all that are independent of & lsquo;evaluation & rsquo;)-were assumed in the semantic structure of the qualifiers. Then three different types of Scales A, B and C as in Fig. 1 were derived. Now Scale B is most frequently used as SD scale, but Scale C is conceivable on each Scale B. 24 sets of scales analogous to these three scales were collected. These scales were used for measuring cognitive component of attitude. Rated persons were liked and disliked persons. This made it possible to, examine the consistency between two attitudinal components-cognitive and affective-.(The second purpose of this study)
53 high school boys were instructed to mention the names of three boys each whom they liked and disliked, then to rate each of them on the 72 seven-point scales.(3 & times;24)
Hypothesis (Hy) 1: Scale As are pure evaluation type of scales, Scale Cs are pure attribute type of scales, Scale Bs are compound scales of both evaluation and attribute.
Hy 2: When rated persons are classified into two groups, liked group (L-gr) and disliked group (DL-gr):
(1) On Scale As L-gr will have positive evaluation, while DL-gr will have negative evaluation.(affective-congnitive consistency)
(2) L-gr will have higher means on Scale Bs than on Scale Cs, while in case of DL-gr the reverse will be true.(affective-congnitive consistency, that is distortions in attribute congnition by affective component)
(3) The rating profiles of L-gr and DL-gr on Scale As, Bs and Cs will be as shown in Fig. 2. The ratings of 139 liked persons (L-gr) and 103 disliked persons (DL-gr) were finally adopted and factor-analyzed by the group principal varimax method to verify Hy 1. Then only on the sets of scales which supported Hy 1, discrepancy measures between means of L-gr and DL-gr and on each group among Scale As, Bs and Cs were calculated and t-tested to verify Hy 2.
The results were as follows:
1) 14 out of 24 sets supported Hy 1. This suggests the evaluative contamination in SD scale.
2)(1) and (2) of Hy 2 were supported,(3) was supported in case of DL-gr. As for L-gr means were higher on Scale Bs rather than on Scale As. This may suggest that cue of attribute makes the rating more positively evaluative than in case there was only cue of evaluation.
As byproduct of this study, Osgood's dynamism factor, coalescence of potency and activity factor, was extracted in person perception.

View full abstract

The purpose of this study is to examine what degree of retardation and acceleration the feebleminded children show in figure cognition compared with the normal children. Particularly, the following three points are comparatively investigated.
(1) The difference in euclidean expression when reproducting diamond.
(2) The difference in the form of direction cognition in the comparison of similarity.
(3) The difference in the ability of figure articulation.
The four sides of the diamond are each 5 Cms, and the upper and lower angles are each 60 degrees. The test for direction cognition in the comparison of similarity consists of a standard figure and seven comparison figures. The test for figure articulation consists of 17 questions of overlapping and embeded figures.
The mental age of the feeble-minded children is 5 or 6 years. The chronological age of the normal children is 4 to 8 years.
(1) In the euclidean experession of diamond reproduction, the feeble-minded of mental age of 6 years are equal in their results to the normal of chronological age of 6 years and clearly superior to the normal of chronological age of 5 years.
(2) In the direction cognition in the comparison of similarity, the feeble-minded of mental age of 6 years are clearly inferior to the normal of chronological age of 6 years, and tend to be behind the normal of chronogical age of 5 years.
(3) In the figure articulation of overlapping realistic figures, the feeble-minded of mental age of 5 years tend to be inferior to the normal of chronological age of 5 years, and to be superior to the normal of 4 years. In the case of overlapping geometric figures, the feeble-minded of mental age of 5 years are clearly inferior to the normal of chronological age of 5 years, and show no difference in comparison with the normal of chronological age of 4 years. In the embeded figures, it is found that the feeble-minded of mental age of 5 years are clearly inferior to the normal of chronological age of 4 and 5 years.

View full abstract