토익 수준별에 따른 영작문의 언어적 특성 연구

A study on linguistic features of writings among different TOEIC scoring groups

The study aims to look into linguistic differences of essays among different scoring groups depending on TOEIC scores. In order to investigate such research questions, the Gachon Learner Corpus collected from over 25,000 essays was used. Each essay was classified into 5 different scoring groups. Statistical results were extracted based on 22 linguistic features using open-source natural language processing tools. Unlike the micro-oriented features of Crossley, Kyle, Allen, Guo and McNamara (2014), macro-oriented features were used. Since micro-feature approach needs more features than macro-feature one does, micro-feature based approach is more efficient in feature extraction process. The current study’s observation revealed that the statistically observed behaviors were varied among different scoring groups. Higher scoring groups had different strategies in language use when showing higher linguistic abilities, compared with those of lower scoring groups. Unlike the observation in Cumming, Kantor, Baba, Eouanzoui, Erdosy, and James (2016), the linguistic complexity feature(s) like syntactic complexity was not clearly distinctive between lower and higher scoring groups.