摘要
本研究运用多层面Rasch测量模型,分析了朗读和口头陈述两种口语考试任务评分维度的设置及评分量表的使用情况。结果表明,口头陈述任务的一个评分维度设置不够合理,两个量表的等级之间都存在非等距性问题,而且口头陈述任务评分量表的某些等级与考生的能力之间出现了负增长现象,即考生的能力越强得分反而越低。这些信息对于修改和完善评分维度和评分量表具有非常重要的意义。
By using many-facet Rasch measurement model, the current study analyzed the fimess of rating dimension setting and scale distribution of two speaking tasks, namely reading aloud and oral presentation. The results revealed that one of the rating dimensions of oral presentation was inappropriate and the points on the two rating scales were both not evenly spaced. What's worse, negative growth on certain scales was found for students with better ability on the oral presentation task, that is, some of the less able were given higher scales. These are of great importance to the improvement of the rating dimensions and scales.
出处
《外语与外语教学》
CSSCI
北大核心
2012年第6期66-70,共5页
Foreign Languages and Their Teaching
基金
中国科学院大学2011年院长基金B(项目编号:Y15102MN00)
中国外语教育基金(项目编号:ZGWYJYJJ2012A26)的阶段性成果