Chinese EFL Learners' Written Production In Controlled Writing Test And Task-based Writing Instruction

This thesis reports on a study about Chinese EFL learners' writing performance in terms of accuracy, complexity and fluency in two different writing tasks, i.e. controlled writing exams and task-based classroom writing.The results show that (1) In task-based classroom writing, the accuracy of learners' written production is higher than that of controlled writing exams and learners tend to produce fewer ambiguous or Chinese-like expressions. The possible reasons are that in classroom writing with the sample article available, learners imitate the original material, from the overall layout to sentence patterns and phrases. Though in writing exams learners may be more aware of the importance of grammatical accuracy, their attentional resource given to accuracy is still limited due to factors such as the meanings and task restrictedness. (2) The measure of syntactic complexity shows no significant difference in the two writing tasks, which proves that syntactic complexity is a more cognitively or linguistically determined measure. Lexical complexity is generally higher in classroom writing. The possible reason is that in classroom writing students are given more freedom to choose familiar topics and contents. Meanwhile, the sample articles may play a stimulating and scaffolding role in the writing process. (3) In terms of fluency, learners perform better in classroom writing. The possible reason is that strict restrictedness in topic and structure in controlled writing exams affect learners' language production. Meanwhile, a contrastive study proves that time also plays an important role in fluency. These findings are discussed in the light of cognitive writing model, information processing theory and Skehan's schema for task analysis.To sum up, this study explores the difference of writing tasks in testing and teaching areas and their possible effect on learners' writing performance. The results give useful implications for the improvement of writing task design in both teaching and testing areas.