论文部分内容阅读
CD-CAT是CDA同CAT的相结合的产物,适用于课堂教学,是教师补救教学、学生自我学习的重要工具。作为CD-CAT重要组成部分的初始阶段项目选取方法是影响测验判准率的重要因素。本文基于现有研究和CDA的项目区分度提出了四种新的初始阶段项目选取方法:CTTID法、CDI法、CTTIDR*法和CDIR*法。通过模拟研究发现,在定长的CD-CAT下,题库质量是HD-HV下,初始阶段结束时,CTTIDR*法的PCCR比现有的T阵法高了.2999,比PWKL高了.1707,其它题库下趋势相同。整个测验结束时CTTIDR*法的判准率仍然是最高的。在变长的CD-CAT下,最大后验概率大于.7、.8、.9下,CTTIDR*法的被试平均测验长度比T阵法分别缩短了2.6170、2.2347、1.7470道题。
CD-CAT is a product of combining CDA and CAT. It is suitable for classroom teaching and is an important tool for teachers to remedy teaching and self-learning. As an important part of CD-CAT, the initial stage project selection method is an important factor affecting the test resolution rate. Based on the existing research and project differentiation of CDA, this paper proposes four new initial stage project selection methods: CTTID, CDI, CTTIDR * and CDIR *. Through the simulation study, we found that under the fixed-length CD-CAT, the quality of the exam bank is HD-HV. At the end of the initial stage, the PCCR of the CTTIDR * method is higher than that of the existing T-matrix method. , The same trend under the other exams. The CTTIDR * method still has the highest resolution at the end of the test. Under variable CD-CAT, the maximum posterior probability is greater than .7, .8, .9, and the average length of test for CTTIDR * method is 2.6170, 2.347 and 1.7470 respectively less than the T matrix method.