摘要
考虑到在实际应用中学生在做题时的猜测和失误(统称为噪音)会影响探索性因素分析法所使用的四分相关矩阵的质量,该文提出四分相关矩阵的一种噪音修正方法,并将其应用于Q矩阵标定.模拟研究结果表明:猜测和失误这2种噪音会对Q矩阵的标定产生不利的影响;基于修正后的四分相关矩阵的探索性因素分析法,在样本量较大和噪音较大等情况下,均能有效地提高Q矩阵标定的准确率.
Considering that the noises of guessing and slip that students answer problems can influence the quality of the tetrachoric correlation coefficient matrix for the exploratory factor analysis method in practical application,a noise preprocessing method of the tetrachoric correlation coefficient matrix is proposed and is used for Q-matrix specification.Simulation studies show that the noises of guessing and slip have an adverse impact on the calibration of the Q-matrix.The exploratory factor analysis method based on the modified tetrachoric correlation coefficient matrix can effectively improve the accuracy of Q-matrix specification when the sample size is relatively large and the noise is relatively high.
作者
汪文义
高朋
宋丽红
汪腾
WANG Wenyi;GAO Peng;SONG Lihong;WANG Teng(College of Computer and Information Engineering,Jiangxi Normal University,Nanchang Jiangxi 330022,China;Elementary Education Collage,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2期136-141,共6页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
国家自然科学基金(61967009,31500909,31360237,31160203,30860084)
全国教育科学规划教育部重点课题(DHA150285)
江西省社会科学规划(17JY10)
江西师范大学教学改革研究(JXSDJG1848)资助项目。
关键词
认知诊断
Q矩阵
探索性因素分析方法
四分相关系数
数据预处理
cognitive diagnosis
Q-matrix
exploratory factor analysis method
tetrachoric correlation correlation
data preprocessing
