A comparison between Automated Writing Evaluation and Teachers' Feedbacks - An Experimental Study based on Juku Correcting Net

摘要

In order to find out whether Automated Writing Evaluation exemplified by Juku Correcting Net is reliable and valid and able to improve students' writing ability, the author compares it with teachers' feedbacks by analyzing students' samples and pre-test as well as post-test writing scores. The results show that the scores given by the teacher and Juku are consistent despite an insufficient degree of differentiation. Compared with feedbacks from teachers, the feedbacks from Juku are more general and just give students some hints about their usage of vocabulary and phrases in spite of offering some knowledge about how to distinguish some synonyms. After practising writing on Juku for a semester, students make some progress in it. It concludes that Juku is expected to be developed to a more intelligent degree and be combined with other means of feedbacks.