Abstract—English essay writing is a very essential skill for university students to master. This is particularly true for those non-English major students in China. Those students represent a large population of Chinese English as Second Language (ESL) learners and often do not receive timely written feedback on their writing from teachers since the size of their class is big. Therefore, it would be very useful if we can build an automated feedback tool for supporting writing. In order to understand the relationship between textual features and human teacher feedback, and how well those features were used for predicting feedback rating, we conducted a corpus analysis of 84 Chinese non-English majors’ essays with teachers’ feedback, by using Coh-Metrix, a computational linguistic tool, to extract rich semantic features from their writing. The study results demonstrated that teachers tend to give more local feedback, such as grammar and spelling, to Non-English majors. The local feedback are more predictable since the feedback was moderately correlated to some textual features (e.g. word count and frequency of content words were related to sentence diversity feedback). Consequently, this study is considered as a case study which should shed light on the construction of automated writing feedback tool for Non-English majors.

Index Terms—E-Learning, Chinese ESL writing, writing feedback, natural language processing.

The authors are with Southwest University, Beibei District, Chongqing, China, 400715

Cite: Ming Liu, Weiwei Xu, Qiuxia Ran, " An Empirical Study of Writing Feedback Analysis of Non-English Majors in China with Natural Language Processing Technologies," International Journal of e-Education, e-Business, e-Management and e-Learning vol. 5, no. 2, pp. 85-93, 2015.