한국 초등 학습자 영어 쓰기 코퍼스 구축 및 오류분석

An error analysis of English writing by Korean young learners based on corpora

This study analyzes the types of errors that occur in the English writing of Korean elementary students. Learner corpora were constructed using English writings of 633 elementary students across various regions in Korea. Errors were identified with the assistance of Grammarly and ChatGPT-4o, and tagged according to the Louvain Error Tagging Manual. An analysis using AntConc revealed that the most frequent errors were in the order of form, articles, prepositions, noun number, punctuation, non-finite/finite verb forms, verb number, auxiliary verbs, and verb tense, indicating that articles, verbs, and nouns are the most challenging areas for Korean learners. An analysis of error frequency by grade level showed a decrease in errors related to form, prepositions, punctuation, and verb number as students progressed to higher grades, while errors in articles, noun number, non-finite/finite verb forms, verb tense, and auxiliary verbs increased. This suggests that while upper-grade students become more familiar with basic elements of English, they also attempt to use more complex sentence structures and verb forms. Additionally, there was little difference in the least frequent error types across different grade levels.