Nowadays, there is an increasing tendency in delinquency of preadolescent children. The energy spent to prevent delinquency is to control the particular environmental conditions which are considered to lead to delinquency. The first purposse of this study was to find out the dimensions of such environmental factors by means of comparative studies on problem and nonproblem children. The subjects were 513 5th grade boys in a city of Osaka Pref. They passed a questionnaire about their attitudes toward their school, classroom teacher, friends, parents, leisure, and so on. Based on the classroom teacher's evaluation on the pupils' behavioral traits, subjects were divided into problem and non-problem groups. The problem group (70 boys) and the non-problem group (42 boys) were compared through their responses to the questionnaire. Consequently, 23 items were found significant between the two groups. Through the factor analysis, using 23 items, 4 factors were abstracted, and each of them was named as follows: school adjustment-maladjustment, emotional conflict between parent-child relationship, leniency-severity by the father, and thrilling play. It was found that these were the very dimensions of environmental conditions leading to delinquency. The second purpose was the classification of typical problem children by applying Hayashi's quantification III. Using the above 23 significant items and 57 categories, 17 typical problem children were classified into five classes. And it was suggested that each class represented the degree of potential delinquency.
A stratified adaptive test of verbal ability has been built by use of 280 vocabulary items with known parameters. The test was designed to measure ability in broad range. Administrations of the test and an additional questionnaire on the test indicate that the test was appropriate for 6th graders through graduate students. An ability score of each subject was estimated by the maximum likelihood method. An information curve for the set of items answered by each subject showed that appropriate items were selected for the subject. The curve of the amount of information has maximum value at the neighborhood of the estimated level of his ability. All these results demonstrated that this stratified adaptive test could be used for measuring broad range of verbal ability.