一般能力検査(SAT)の追跡研究 : その1 変遷と現状(1961〜1977)

概要

論文の詳細を見る
This report is a summary of our joint study under the aforementioned title, conducted at the Institute for Educational Research and Service in ICU during 1976-1978. The study aimed at to self-evaluate our present SAT and to invite criticisms and suggestions from our colleagues for its further improvement. 1. Purpose and Characterisitics of SAT. As being described in the Manual of ICU Admissions Tests, our SAT intends to measure a part of basic learning ability, i.e. scholarstic aptitude, throgh various tests of word and concept comprehensions, understanding of word relations, e.g. synonym and antonym, analogies with mathematical and figural materials, and logical reasonings. Although the ultimate goal of SAT is same to other learning ability tests for admissions, the former emphasizes more or less a phase of quick judgement and accurate inference, which are supporsed to be essential for our clear, flexible and creative thinking in advanced studies. 2. Original Saurces and Revising Procedures. Several members of the IERS were asked to construct a test for the substitution of College Entrance Aptitude Test, which had been administered by the Ministry of Education but became no longer availabe in late 1950s. Thus, under the leadership of late Professor Yataro Okabe and Dr. Maurice E. Troyer, and modeling after the Miller Analogies Test of the Psychological Corporation in New York, a new form of SAT has been constructed every year based on the through scrutiny of those questions employed in previous years. Item analysis and factor analysis were two main techniques and the saurces of information for revision. The present form consisting of 100 items can be grouped into 14 subtests. 3. Distribution of Scores and Reliability of Measurement. By vertue of item anaysis and our effort for standerdization, all editions so far constructed have had remarkably similar distributions close to ideal and relatively high reliabilities. Ranges of the standard error of measurement were estimated to assist examiners to evaluate raw scores with necessary caution. 4. Factorial Structure. Correlational matrices calculated from the 14 subtests were factorially analyzed for 4 successive years since 1974. The application of centroid method revealed that all 14 subtests were highly loaded with a general factor. The second factor further differentiated them into verbal and nonverbal aptitudes. Results obtained by varimax method resembled to the above. Interestingly enough, however, the third factor extracted by this method reflected differences of cognitive styles or strategies required for problem. solvings. 5. Correlations among High School Records, Admissions Test Scores and College Grades. Correlation Coefficiencies among grades of various subject matters at high school, admissions test scores and gradess obtained from the College of Liberal Arts at ICU were factor analyzed by varimax method. The highest loadings were observed in 5 major subject matters of high school and all college grades. The second factor included high school English, English tests for admissions, Freshman English Program and grades in majoring areas. The third factor related mostly to SAT. The fourth factor seemed to influence college grades, among which physical education had the highest loading. From these results, it was revealed that the college grades would never be predicted only from high school performances or admissions tests, but students' aptitudes toward their fields of concentration as well as their motivational levels have significant roles to determine their undergraduate academic performances.

一般能力検査(SAT)の追跡研究 : その1 変遷と現状(1961〜1977)

スポンサーリンク

概要

著者

関連論文

スポンサーリンク