摘要
教育考试中专题、短测验等形式是命题的一种主要方式。对这类测验的分析,可以从参数项目反应理论和非参数项目反应理论入手。本研究分别选取Rasch模型和Mokken模型对某高三文科综合地理试卷进行分析比较。使用Winsteps和Xcalibre软件进行Rasch分析,得到难度、信息量、项目功能差异等参数;使用MSP软件进行Mokken分析,得到正答率和同质性系数。比较两种结果,得出以下结论:(1)非参数项目反应理论以正答率对题目排序与参数项目反应理论以难度排序一致;(2)而有个别不符合参数项目反应理论标准的题目对提高测验质量同样有意义,不应被删除;(3)进行维度检验和题目筛选时,非参数项目反应理论标准比参数项目反应理论标准更加严格;(4)两种理论的项目功能差异检验结果一致。
As one of the significant types of tests, the test project and short test are popular in educational testing. Parameter and non-parameter item response theory being the starts, these tests were under analysis. Compared was the geography paper in inaugurated arts taken by some senior three students. During this comparison the Rasch and Mokken model were respectively selected. For analyzing software Winsteps and Xcalibre were utilized to analyze item parameters in Rasch model. Analyzed in detail were the parameters of difficulty, differential item functioning and information curve. Software MSP was for the purpose of analyzing items in Mokken model. Besides, the statistics of accurate rate and coefficients of homogeneity were also analyzed in detail. Finally, four conclusions were arrived at as the following： （ 1 ） The estimate results of difficulty between non-parameter and parameter item response theory were equivalent. （2）Those items, which failed to fit parameter item response theory, succeeded in non-parameter item response theory. （3）Non-parameter item response theory is more rigorous than parameter item response theory in dimension testing and item screening. （4）The result was equivalent in the detection for differential item functioning between non-parameter and parameter item response theory.
出处
《中国考试》
2012年第10期18-24,共7页
journal of China Examinations
基金
贵州省高等学校教学质量与教学改革工程重点项目"基于PBL理论改进心理教育测量教学改革研究"(项目批准号:黔高教发[2011]28-1)
贵州师范大学精品课程"心理测量"建设项目阶段性成果
关键词
教育测量
短测验
非参数项目反应理论
摩根模型
Educational Measure
Short Yest
Non-parameter Item Response Theory
Mokken Model
