作　　者： ; ; ;

机构地区： 华南师范大学教育科学学院心理应用研究中心

出　　处： 《心理与行为研究》 2009年第1期27-31,共5页

摘　　要： 以广东省佛山市"升中"考试为例,分析和探讨如何选用合适的等值设计与方法来解决普教"升中"考试不同地区分数转换的问题。采用非随机组锚测验等值设计对三种经典测验等值方法进行比较。结果发现:Tuck-er线性等值方法最优,Levine线性等值方法次之,等百分位等值方法(频数估计)不适合此类等值。等值方差分析表明题型与等值方法具有交互作用,这说明不同的题型宜选用不同的等值方法来进行等值。 This research exemplified high school entering how to select decent equating design and method to exam of Foshan city in Guangdo solve the problem about how to ng province and explored transfer these scores from different zones. Using un-random anchor design, we compared three methods of classical test and found that the Tucker liner method was the best among them. The Levine liner equating method was relatively good. As for this examination, the unsmoothed frequency estimation equipercentile method was not appropriate. Equivalent analysis of variance （3×3） showed that the equivalent methods and the exam items were interacted （p〈0.001） . Different items were suitable for different equivalent methods.

关 键 词： 测验等值 经典测量理论 锚测验设计 升中 考试

领　　域： [哲学宗教] [哲学宗教] [文化科学] [文化科学]