4-6年级多重成就测验的编制II:信度考验

目的:以CTT和GT作4-6年级MATs的信度检验。方法:重测信度161例,复本信度180例,重测复本信度200例。项目有效样本2002例,其中A本1004例,B本998例。结果:A、B本分量表和总量表重测信度0.91～0.95,复本信度0.87～0.94,重测复本信度0.82～0.89,分半信度0.79～0.90,系数0.90～0.96,评分者信度0.94～0.98,真分数变异0.82和0.86;概化研究表明分测验条目15个左右为宜,分量表与总量表条目量50与100即可。结论:各种信度考验结果基本符合测量学的计量标准。

Multiple Achievement Tests for Pupils of Grades 4-6 II:Rialibility

Objective: MATs reliability of the 4-6 Grades was tested.Methods: 161 subjects were given the MATs test-retest;180 subjects,parallel test;and 200 subjects retest and parallel test.Item samples consist of 2002 students.Results: The reliabilities of the subscales and full scales ranged from 0.91 to 0.95,in Spearman test-retest correlations,the parallel reliability was 0.87 to 0.94;the retest and parallel reliability 0.82 to 0.89;the split-half reliability 079 to 0.90 and the Cronbach's coefficient 0.90 to 0.96.The scorers reliabilities were 0.94-0.98.The true scores of MATs-A and MATs-B were 0.82 and 0.86,respectively.Generalization analysis showed that it was proper to have about 15 items in each subtest,and 50 and 100 items in the subscales and the fullscale.Conclusion: A series of reliabilities testing in the MATs conformed to the psychometric standards.