基于课堂测评的英语写作同伴互评量表研发及验证  

Development and validation of a college English writing scale for classroom-based peer assessment

作　　者：马晓梅[1] 史晓婷 芦畅 李荣 Ma Xiaomei;Shi Xiaoting;Lu Chang;Li Rong

机构地区：[1]西安交通大学外国语学院 [2]上海交通大学教育学院 [3]荣耀终端有限公司

出　　处：《西安外国语大学学报》2022年第1期56-62,共7页Journal of Xi’an International Studies University

基　　金：国家社会科学基金一般项目“英语认知诊断测评模式构建及有效性论证”(项目编号:17BYY015)的阶段性研究成果。

摘　　要：本研究通过有声思维、专家判断、非参数维度检验(DIMTEST,DETECT)及多层面Rasch模型等方法构建并验证服务于大学英语写作课堂测评的基于实证法的描述语评分量表(EDD checklist),并通过T检验探讨经过验证的量表是否能缩小教师和学生的评分差异。结果表明,由30条描述语构成的量表内部结构稳定,五个构念相互独立。Rasch模型分析结果表明五个构念难易度不同,量表能有效区分写作者能力,且能降低评阅人的宽严度差异。教师和学生的评分结果经T检验比较进一步表明,描述语量表比整体性量表更能有效提高评阅人之间和内部的评分一致性。研究结果对英语写作同伴互评具有一定实际应用价值。This study develops an English writing EDD checklist by employing raters’ think-aloud protocol and expert judgement. The EDD checklist has been verified by nonparametric dimensional analysis(DIMTEST, DETECT) and Multifaceted Rasch model analysis(MFRM). Then the teachers and students’ rating consistency was tested by using T-test under the conditions of EDD checklist and CET-4 holistic scale(CET-4S). The results indicate that the EDD checklist, with 30 descriptors, categorizing into the 5 independent constructs, keeps a good internal consistency. MFRM analysis reveals that the difficulty of constructs varies significantly and the EDD checklist can effectively distinguish students’ writing abilities and reduce raters’ bias. T-test analysis further demonstrates that the EDD checklist can better improve both inter-rater and intra-rater consistency than CET-4S. Thus, it could be hopefully used for classroom peer-assessment of non-English majors.

关 键 词：描述语评分量表 英语写作 构念 评分员一致性 

分 类 号：H319[语言文字—英语]