Application Of Automated Essay Scoring System In College English Writing Instruction

How to help students improve their writing through feedback has long been one of the major issues for foreign language teachers and researchers. The rapid development and broad application of information technology have injected great impetus into the transformation of teaching and learning methods,which have posed challenge and also provided opportunities for the reform of College English Writing Instruction in Mainland China. On the one hand,although remarkable achievements have been made, "bottleneck"problems of the imbalance between the little educational and social success and the enormous funding, resources and manpower arise. The applicability of education informatization technology is thus questioned by the public. One the other hand, the automated essay scoring system of Juku (www.pigai.org) has been widely used throughout China, while very few empirical studies based on real classroom settings have been made to verify the reliability and validity of its use. Such a situation necessitates studying Juku application in the classroom setting,which may cast new light on the proper use of Juku.This empirical study explores three questions: (1) What kinds of feedback can Juku give to the students? (2) Which types of feedback (among the given ones) do the students actively respond to? (3) Is the number of submission times linked to the students' writing scores?In order to answer the questions, this study takes 65 (with 32 in the experimental group and 33 in the control group) non-English major freshmen of a key university in Beijing as participants, collects data through eight writing exercises and informal interview and analyzes the relationship between scores, revision times and the acceptance of different types of feedback provided by Juku, etc.The results show that Juku can indeed influence the improvement of students'writing skills; the number of revision times does not correlate with the last score of a writing, but correlates with the difference value (difference value = last score - first score); this correlation can be illustrated with the coefficient "k". and the higher the value of "k" is,the higher the "efficacy of revision" will be, and thus the greater the improvement will be; there is "distance" between the feedback provided by Juku and the revisions that can effectively improve the scores; the feedback provided by Juku as a system at its present level is far from satisfactory, and that is where teachers can fit in with extra feedback (especially content-focused feedback). Finally about how to provide effective and efficient feedback that can improve writing instruction, this thesis has made suggestions based on the conclusions reached in the research study: firstly, the system of Juku should be improved, especially its ability in giving content-focused feedback;secondly,more training on Juku use for the users,both teachers and students,should be conducted; and finally, other forms of feedback should be incorporated in English writing teaching, so as to form up an integrated feedback system.