摘要
标准参照测验采用子分数衡量学生在具体内容、知识或技能上的掌握情况,这有助于发挥考试的学习功能.在少量题目上作答数据估计子分数,其信度较难保证.子分数估计方法可充分利用相关辅助信息获得信度较高的子分数,这对补救教学至关重要.在简要介绍测量模型之后,叙述了子分数的7类估计方法的思想和计算过程,并分析了各方法的应用和表现,需重点关注群体和个体、复杂结构、优化测验设计和其他施测方式下子分数估计.
Criterion-referenced tests focus on mastery statuses of students in different content,knowledge or skill areas by subscore,which can be beneficial for playing the learning function of the examination.Subscore only estimated from a small number of test items makes it difficult to establish high reliability.Subscore estimation methods can make full use of auxiliary information on associated test data to obtain subscore with higher reliability,which is crucial to remedial teaching.After a review of measurement models is briefly introduced,the seven subscore estimation methods and their detailed computational procedures are explained.And the application and performance of each method are analyzed.Finally,subscore estimation of group-levels and individual-levels or under complex structures,optimal test design,and test administration needs to be focused on.
作者
宋丽红
汪文义
丁树良
SONG Lihong;WANG Wenyi;DING Shuliang(Elementary Education College,Jiangxi Normal University,Nanchang Jiangxi 330022,China;College of Computer Information Engineering,Jiangxi Normal University,Nanchang Jiangxi 330022,China)
出处
《江西师范大学学报（自然科学版）》
CAS
北大核心
2020年第3期292-300,共9页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
全国教育科学规划教育部重点课题“基础教育质量监测分数报告方法研究”(DHA150285)资助项目.
