基于固定共同题方法的IRT等值模型比较研究  

A Comparative Study of IRT Equivalence Model Based on Fixed Common Item Parameters Method

作　　者：张泉慧[1] 张颖[1] 何佳[1] 邹杰文 王娇艳[1] 

机构地区：[1]国家医学考试中心,北京100097

出　　处：《中国考试》2017年第6期65-69,共5页journal of China Examinations

摘　　要：等值模型的选择是题库建设的重要环节。本研究在IRT理论框架下,选择最为常用的Logistic模型,针对医学考试某一学科测试比较三种Logistic模型下固定共同题等值方法的差异。结果表明,单参数模型对共同题的参数固定更稳定,等值的精确性优于双参数模型和三参数模型,更适合用于该学科测试。Choosing a certain measurement model is an important part of constructing item bank. This study attempts to analyse the differences in three types of Logistic model by fixed common item parameters method under the IRT theory.The object of study is some kind of medical examination.The result is that the method of 1PLM is more stable, the parameters is smaller than 2PLM and 3PLM, so the 1PLM is suitable in this examination.

关 键 词：题库 LOGISTIC模型 等值 IRT 

分 类 号：G405[文化科学—教育学原理]