Utilizing an automated tool analysis to evaluate EFL students’ writing performances

Abstract

Abstract Recently, the integration of linguistics and technology has been promoted and widely used in the field of linguistics and English writing research for several purposes. One of those purposes is to evaluate English as a Foreign Language (EFL) writing ability by using electronic assessment tools. In the current study, an automated writing evaluation tool (Coh-Metrix) was used to indicate English-major students’ writing performances based on the discourse components of the texts. The English texts generated for each writing task on two different topics were collected. The corpus analyses gathered from Coh-Metrix identified linguistic and discourse features that were interpreted to determine the 40 EFL undergraduate students’ English writing abilities. The students wrote and revised their essays in hand-written essays in class and resubmitted their essays in digital forms with corrections made. The results showed that these students demonstrated linguistic flexibility across writing assignments that they produced. The analyses also indicated that the length of the texts and the uses of the word concreteness, and the referential and deep cohesion had impacts on the students’ writing performances across the writing tasks. Besides, the findings suggest practical value in using an automated text analysis to support teachers’ instructional decisions that could help to identify improvement of students’ writing skill.

Keywords

• EFL writing
• Coh-Metrix
• Discourse components
• Corpus linguistics
• Computational linguistics