摘要
本研究利用建构图设计一套含有六大部分的30道试题。题型包括拼写题、选择题和简答题。共有175名6到14岁儿童参加了此项考试。Rasch分析结果发现题组内局部题目依赖并不严重。信度为0．85。考题的难度和考生能力的配合度相当良好。我们根据建构图来编写考题,因此有一定程度的内容效度。但有9道题的难度稍微与原先预期略有出入。有5道题不大吻合Rasch模式的预期,没有发现在性别上有明显的项目功能差异。考生能力与学习英语的时间有正相关。最后探讨了基于信息通讯技术的远程计算机自适应测验的技术问题。
Based on the four-building-block theory (Wilson, 2005), we created a 30-item test which contains 6 parts and 5 items in each. A set of 175 Chinese studetnts (aged from 6 to 14) took the test. Rasch analysis showed that local item dependence was negligible. The test reliability was . 85. The persons and the items had a good match. A set of 9 items had difficulties slightly beyond the authors expectation. Only 5 items did not fit the model s expectation very well, and no items had substantial gender differential item functioning. The correlation between the duration of English learning and the person measures was. 40. Finally, issues in implementing online computerized adaptive testing are discussed.
出处
《考试研究》
2006年第3期72-86,共15页
Examinations Research
基金
本研究受国家留学基金委资助。
