'Controlled' Writing Portfolio Assessment

This thesis, through a case study over a semester-long period, makes a tentative study into the feasibility of an alternative writing assessment, namely, the 'controlled' writing portfolio assessment.From the analysis in this thesis, we can find that the experimental writing assessment is both valid and reliable, through which the assessment functions not only as a kind of test to evaluate students' competence in writing, but also a diagnostic tool that provides feedback to learners and teachers about the suitability of the instructional materials. Furthermore, by the inclusion of the self-assessment of their own writing, it helps demonstrate to students that they are making progress in their language acquisition linguistic development, hence a boost in motivation. Besides, this assessing approach provides a broad picture of a student's achievement by showing the unfolding of skills over time. This is something that a one-time performance on a test cannot do. Through the long-term progress evaluation, this assessing approach strengthens collaborative efforts of students, teachers, and books instead of individual efforts. The study also demonstrates that the use of analytic scale is a kind of make-up to the holistic scoring in the final examination, and the samples in the portfolios also provide additional information on students' writing ability.Compared with the traditional practice in China, the 'controlled' writing portfolio assessment is believed to be more reliable and could not only better integrate the assessments with classroom instructions but also present a dynamic rather than static picture of students' linguistic development.