문항반응이론을 활용한 한국어능력시험의 문항 양호도 분석

A study of analyzing the quality of items on the Test of Proficiency in Korean(TOPIK) using item response theory

The purpose of this study is to analyze the quality of items on the 28th through 32nd Test of Proficiency in Korean(TOPIK) using item response theory. Item Response Theory is invariable and insusceptible to the subject and test tool, not of the classical Test Theory that has been generally used in analyzing the difficulty or discrimination of items. The study found that there existed test items that were at an extremely high level or low level of difficulty, and their distribution was not constant amongst different tests and levels. Even in terms of item discrimination, it was found that there existed test items with little or low level of discrimination. In addition, it was shown that the difficulty level in certain types of test items from intermediate-level listening and some free-response test items from intermediate-to-advanced level writing, was either too high or low, and their discrimination level was found to be low. Therefore, measures to raise the quality of those relevant items must be sought out, and a new check on the rubric for making and grading the test questions is deemed necessary especially for the free-response writing questions. TOPIK is scheduled to be reshuffled and implemented with the new system starting from the year 2014. Tests will be implemented six times on a yearly basis. With the increase in the number of tests annually, chances are high that securing the quality of test items will become even more difficult. To tackle this problem, a conversion from the existing method of making test items to the item bank method using item response theory is inevitable.