절대평가 시험으로서의 수능 영어시험 분석: 주요 난이도 예측 변인을 중심으로

Abstract

Given a great variance of the number and percentage of test-takers with high band scores in the CSAT English tests for the 2015-2020 academic years, the present study compared these six tests in terms of facility index (FI), measures of lexical richness and syntactic complexity, distractor feasibility, information density in listening materials, familiarity of speakers’ pronunciation and accent, and levels of background knowledge required to understand reading passages, which are all said be major predictors of question item difficulty. One-way ANOVAs, independent sample Kruskal-Wallis H tests and chi-square tests showed that there was no significant difference concerning FI and any of the predictors not only between the six CSAT English tests but between their reading parts. There was also no significant difference between the listening tests except that the test for the 2017 academic year was significantly easier than for the 2016 and 2018 ones. These findings together mean that the introduction of criterion-referenced testing in the 2018 academic year has brought no substantial changes to CSAT English tests. Interestingly, however, there was a significant difference in FI between the six sets of the 13 most difficult questions drawn from each test, which was a reason for the great variance of high scorers between the six tests. The study concluded with some pedagogical discussions.