CAT分层选题策略新探——最大信息量组块分层策略

Maximum Information Stratification Method With Blocking in Computerized Adaptive Testing

选题策略在计算机化自适应测验研究中是一项重要内容,它的好坏与考试的信、效度,测验安全性以及测量准确性直接相关。最大信息量组块分层策略(Maximum Information Stratification Method WithBlocking,MIS-B)针对a分层法在实际应用中的不足,将题目猜测度c和组块(blocking)策略引入其中,提高了测量准确性和测验安全性。此外,尝试性地将改良多项式模型(The Modified Multinomial Model,MMM)融入MIS-B中,提出了MMM-MIS-B(4MIS-B)的雏形。

Item selection method is an important component of a computerized adaptive testing and has significant effects on the measurement efficiency, test security, test reliability and validity. Solving the shortage of a-STRatified method in the actual application, Maximum Information Stratification Method With Blocking （MIS-B） incorporates the c parameter and the blocking in the stratification of the bank and in the item-selection rule, improving accuracy and security by comparison with the a-STRatified method, for item banks with a and b parameters correlated and uncorrelated. In addition, this paper attempted to blend the Modified Multinomial Model into MIS-B, proposed the rudiment of a new item selection method ： MMM-MIS-B （4MIS-B）.