命题者:影响阅读理解测试效度的一个因素  

Test Constructor as a Factor Affecting the Validity of Multiple-Choice Reading Comprehension Test

作　　者：李雪[1] 曾用强[2] 

机构地区：[1]东华理工大学外国语学院 [2]广东外语外贸大学英文学院

出　　处：《考试研究》2012年第4期49-60,共12页Examinations Research

基　　金：2011年度东华理工大学校长基金资助项目"测试方法对阅读考试效度的影响"成果;项目编号DHXW1111

摘　　要：本研究应用项目反应理论,从被试的阅读能力值和题目的难度值这两个方面,分析阅读理解测试中多项选择题命题者对考试效度的影响。实验设计中,将两组被试同时施测于一项"阅读水平测试",根据测试结果估计出的两组被试能力值之间无显著性差异。再次将这两组被试分别施测于两位不同命题者所命制的题目,尽管这些题目均产生于相同的阅读材料,且题目的难度值之间并没有显著性差异,被试的表现却显著不同。Rasch模型认为,被试表现由被试能力和试题难度共同决定。因此,可以推测,这是由于不同命题者所命制的题目影响了被试的表现,并进而影响了使用多项选择题进行阅读理解测试的效度。This paper reports on a study of the application of Item Response Theory to explore the impact of the test constructor on the performance of test takers in multiple-choice （MC） reading comprehension test from two perspectives, namely, reading ability estimate of candidates and item difficulty. This study arrives at the conclusion that the difference between students＇ reading ability estimation in reading proficiency test is insignificant, while the difference between students＇ reading ability estimation in the two tests is significant. Based on Rasch model, candidates＇ performance is attributed to candidates＇ ability and item difficulty. As there is no significant difference between the item difficulty of the distinct items in the two tests, we can predict the performance between group 1 and group 2 in the Test A and Test B is significantly different. This suggests the distinct items devised by item writer 1 and item writer 2 affect candidates＇ performance and the validity of multiple-choice reading comprehension test.

关 键 词：命题者 阅读理解 效度 测试方法 

分 类 号：G424.74[文化科学—课程与教学论]