摘要
评分量表代表了测试构念的操作化定义,是写作评分的重要组成部分。为了解整体和分项量表对EFL作文评分信度的影响,本研究对9名评分员对同样50份作文使用整体和分项量表所作的评分进行了多层面Rasch模型分析。研究发现,分项量表的使用能够使评分员对考生的英语写作能力作出更加细致和准确的区分,而其严厉度上的差异对考生能力估计的影响也更小。同时,较之整体量表,分项量表各分数间的区别也更加明显。总的来说,分项量表的使用对EFL写作测试的评分信度能产生更为有益的影响。
As realizations of theoretical constructs,rating scales play an important role in the scoring of compositions.In order to know the effect of the use of holistic and analytic scales on the reliability of EFL essay rating,the present study,through the application of M any-facet Rasch M odel,analyzed the scores aw arded by 9 raters on the same set of 50 EFL essays w hile using both a holistic and an analytic scale.Results show ed that the use of the analytic scale led to finer distinctions among examinees in terms of their English w riting ability.M eanw hile,the impact of the variability in raters＇ severity on examinee ability estimates w as also smaller w ith the use of the analytic scale.What＇s more,w hile there w as a lack of clear distinction betw een most adjacent holistic scores,this problem w as not detected for the analytic scale categories.On the w hole,the use of the analytic scale led to more favorable impact on the reliability of EFL essay rating.
出处
《外语与外语教学》
CSSCI
北大核心
2015年第2期45-51,共7页
Foreign Languages and Their Teaching
基金
浙江省教育厅科研项目"大学英语六级考试作文评分的信效度研究"(项目编号:Y201120920)
国家社会科学青年基金项目"中国大学生英语写作能力的认知诊断评估"(项目编号:13CYY031)的阶段性成果
