摘要
在Rasch测量模型框架下,根据Rudner分类方法,对第三批高考综合改革8省份使用的5等级赋分方式进行等级分数分类一致性和准确性研究。研究侧重试题信息函数对分数等级分类一致性和准确性的影响,结果表明:构成测验的试题类型、试题数量、提供的信息量大小、信息函数分布等是影响分类一致性和准确性的重要因素,其中,试题信息函数分布和试题类型对分类一致性和准确性的影响更为显著。
Based on Rasch test theory and Rudner decision-making classification method,this paper studies the consistency and accuracy of the score classification of the five-level and equal-proportional score division methods used in the eight provinces and cities in the third batch of comprehensive reform of the college entrance examination.The results show that the type and quantity of test items,the number of items,the size of information provided by the items and the distribution of information functions are the important factors affecting the consistency and accuracy of classification,while the normality of information function curve and the type of items have more significant influence on the consistency and accuracy of classification.
作者
宋吉祥
李付鹏
杜海燕
储林林
赵发忠
SONG Jixiang;LI Fupeng;DU Haiyan;CHU Linlin;ZHAO Fazhong(Anhui Education Examinations Authority,Hefei 230022,China)
出处
《中国考试》
CSSCI
北大核心
2021年第3期22-27,共6页
journal of China Examinations
基金
国家教育考试科研规划2019年度重点课题“新高考学业水平考试等级赋分模拟及相关问题研究”(GJK2019024)。
关键词
高考综合改革
分类一致性
分类准确性
IRT模型
等级分数
试题信息函数
comprehensive reform of college entrance examination
consistency of classification
accuracy of classification
Rasch model
grade scores
item information function