大學入學考査의 測定理論的 分析硏究 = An Analysis of College Entrance Examination by Measurement Theory

다국어 초록 (Multilingual Abstract)

A. Summary
In the situation of Korea as a developing country, competions for entering colleges are an unavoidable problem, and harmful after-effects from it are occurring as educational or social problems in spite of the efforts to supplememt it in all phases. Therefore, this kind of study must be continued so that the entrance exams can be functioned perfectly.
This study has three purposes of analysis:
1. Can the entrance exams be valid and reliable?
2. Can the entrance exams function naturally?
3. Can the tests for entering colleges have proper item difficuty, high item discrimination, and proper option responses?
In this study the results of entrance exams of Jeonbug National University executed from 1970 till 1980 were analyzed and compared with each other.
Score distribution, central tendency, variability, validity, reliability, item difficulty, item discrimination, and option responses were studied analytically. Most of all were analyzed according to the theory of measurement, and the results of consistent statistical study about them from 1970 till 1980 were borrowed, compared, and analyzed.
When the result of this dealing was analyzed, the score distribution was abnormal. And central tendency (Mean) and variability (standard deviation) were a little more or less according to each subject. And as to the total scores all the numerical values were very close to the ideal according to the theory of measurement.
On the aspect of making items regarding behaviors, items about application, analysis, synthesis, and evaluation were few and in some subject, not at all. Correlations between total scores and subjects and subjects and subject, were a desirable result, and so the validity of all subjects can be proved to be good. Correlation of IQ and the scores was somewhat high in the intellect-centered subjects.
Reliability was somewhat low because of the methodical improperiety of making items. More efforts to better it are still needed.
Difficulty is somewhat low. But in case of Jeonbug National University items must be easier than this year. For example the items of mathematics in 1979 were very difficult.
Discrimination is ordinary, but taking into account of selective examination, that extent must be raised.
As a result of analysis about option responses more attention is technically needed in making multiple-choice items.
As to the high school records, entrance examination records, and college records are correlated, high school records and entrance examination must be study continuelly.
B. Suggestion
The method of entrance examinations has been studied and improved as years go on. Some methods as almost perfect now but others meed supplements. The concrete suggestions for this are as follows:
1. In the study of validity predictive validity be studied.
2. To increase reliability many methodical study is needed considering the importance of entrance examination.
3. Difficulitis as some subjects must be controled.
4. To increase discrimination technical study is needed in making multiple choice items.
5. To better option response some more concrete technical study and consideration are needed.