经典测量理论中的信度理论与信度系数——与张敏强教授商榷

Reliability Theory and Reliability Coefficient in Classical Test Theory: A Discussion with Prof.Zhang Minqiang

文章指出了《教育测量学》一书关于信度理论和信度系数值得商榷之处:系数的定义表述不严谨,未区分信度系数的定义式和估计式,对信度系数定义的评注不符合科学现状,定理的证明含有错误,证明的条件不全,符号使用混乱;建议在入门教材中删除有关信度系数的证明过程。

This paper elaborates on the doubtful points with regard to reliability theory and reliability coefficient in the textbook Educational Measurement. These points include loose definition of reliability, unjust remarks on the definition, erroneous proof of theorems and inadequacy of necessary conditions for the proof, inconsistency in notation and the use of symbols. It suggests that in textbooks at the introductory level, mathematical proofs of theorems concerning the estimation of reliability coefficient be cut out.