基于Rasch模型的多项式选择完型填空与集库式完型填空信效度对比

Rasch Model Based Comparative Study on Validity Between Multiple-Choice Cloze Questions and Banked Cloze Questions in TEM-4

针对2016年英语专业四级水平测试中完型填空试题类型由多项式选择调整为集库式,研究了完型填空改革前后的试题信效度是否有差异值,以便为英语专业四级水平测试的完型填空命题提供实证依据。以48名英语专业学生为研究对象,改革前后的完型填空试题为测试材料,运用Winsteps和SPSS 23.0构建Rasch模型,以探讨改革前后完型填空试题的信效度差异。研究发现多项式选择完型填空和集库式完型填空测试效果具有显著差异,应增加集库式完型填空题项数量以便提高英语专业四级水平测试的信效度。

In 2016, there is a reform of TEM-4 ＇ s question forms. The cloze test changes from muhi- pie-choice cloze questions to banked cloze questions. This study takes muhiple choice cloze ques- tions in 2015 before reform and bank cloze questions alter reform in 2016 as research subjects, and research data are from authentic test results of 48 test-takers. This research aims at exploring the re- lationships between two different cloze forms and the differentiated performance of test-takers by ap- plying Rasch model analyses, and also testing reliability and validity of two cloze forms. The re- search results show that muhiple choice cloze questions and banked cloze questions have great differ- ence on reflecting test-takers＇ English ability, however, banked cloze questions can better reflect test-takers＇ English ability if more items are added.