首页 | 本学科首页   官方微博 | 高级检索  
     

认知元反应理论--IRT直接应用于多值记分题
引用本文:缪源,李绍珠. 认知元反应理论--IRT直接应用于多值记分题[J]. 心理科学, 2000, 23(2): 196-199
作者姓名:缪源  李绍珠
作者单位:南京大学高等教育研究所,210093
摘    要:0-1记分测验的项目反应理论已经得到广泛的研究和应用.但是,许多测验都含有多值记分题,所以需要将IRT推广到此类情况.从认知理论的观点看,每个0-1记分题(项目)和多值记分题的每个测试点都可同样地看成一个由若干知识点构成的集合,称之为认知元;根据认知元之间存在的关系可以确定各受测者对各试题作出特定答案的概率,从而不需要引用任何其它假设就可将IRT的方法直接应用于含多值记分题的测验.本文应用这一理论分析了某些测验样本,结果表明是可行的.

关 键 词:认知元 认知元反应理论 IRT 多值记分题 极大似然估计

Cognition Element Response Theory-Using IRT in Polychotomously Scored Items Directly
Miao Yuan,Li Shaozhu. Cognition Element Response Theory-Using IRT in Polychotomously Scored Items Directly[J]. Psychological Science, 2000, 23(2): 196-199
Authors:Miao Yuan  Li Shaozhu
Abstract:The item response theory (IRT) for dichotomously scored tests has been studied and used extensively However, many tests employed polychotomously scored items So it is necessary that IRT be extended to such tests From the view point of cognition, any dichotomously scored item is a collection of knowledge, so is any test point of polychotomously scored items, which may be called a Cognition Element (CE). According to the relation among CEs, it is not difficult to find the probability of the whole test sample The IRT can be used in polychotomously scored items directly without any other hypothesis We could estimate the parameters of the CEs and the ability parameters of the examinee by maximum likelihood We have utilized this method to analyze several tests and found it works well
Keywords:cognition element ( CE)    cognition element response theory   IRT   polychotomously scored item   maximum likelihood estimate  
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号