영어작문수업에서의 자동작문채점기 FES 활용에 대한 타당성 연구

Validation study of automated essay scoring system applicable to English writing class

This study aims to validate automated essay scoring(AES) named free essay scoring(FES), in order to explore its applicability in English writing classes in Korea. To achieve this aim, a total of 57 essays from 27 high school students were selected and upon the completion of essay writing, a retrospective interview was conducted. Further the grading results of AES were compared with those of human raters. The analysis using Generalizability theory revealed that the grading results through AES were stable by presenting generalizability coefficient relatively robust enough to apply to low stake testing situations, and the two writing tasks provided by the AES proved to be valid in terms of the task difficulty as intended by the system. That is, a variety of evidences investigated in this research showed that the writing tasks the AES provided were classified according to the task difficulty. Lastly, students showed somewhat mixed perceptions on the use of AES and its feedback. Further implications were suggested.