Innovation in Language Learning

This study investigated the syntactic complexity of reading passages in the EIKEN Test in Practical English Proficiency using Natural Language Processing tools. The EIKEN is a widely used English proficiency test in Japan that comprises seven grades. In recent years, many high school students have submitted their EIKEN results to university entrance exams, making it important for teachers to understand the specificities of the test. Previous studies on the reading section of the EIKEN have mainly focused on its text coverage [1] and task validity [2]. Hamada [3] investigated the difficulty level of the passages and found that it was mainly manipulated using word-level variables instead of syntactic variables. However, the only metric of syntactic complexity used was syntactic similarity between sentences. In the present study, all 14 metrics of the L2 Syntactic Complexity Analyzer [4] and three other metrics based on dependency trees were used to compare the reading sections from Grades 4 to 1. The results revealed consistent differences across all levels from Grade 4 to Grade Pre-2. For the highest levels, that is, Grades 2, Pre-1, and 1, the length metrics and dependency tree depth differed across each level. In conclusion, the difficulty level of EIKEN reading passages is manipulated by not only lexical variables but also syntactic variables. These findings should be considered by teachers when preparing their students for the EIKEN, for example, by adapting their teaching methods depending on the target level.