基于概化理论构建自学考试信度分析框架  

Framework of Reliability Analysis in Self Taught Examination Based on GT

作　　者：田霖[1,2] 韦小满[2] 王桥影[1] 赵晓茫[1] 

机构地区：[1]北京教育考试院 [2]北京师范大学教育学部教育统计与测量研究所

出　　处：《考试研究》2013年第5期27-33,共7页Examinations Research

摘　　要：概化理论(GT)提出了新的测量信度观并逐渐应用于大规模考试领域。本文将概化理论思想与方法引入自学考试信度分析研究与实践,提出了自学考试信度分析的概化理论研究框架。本框架对自学考试信度分析的基本任务与流程进行了梳理与总结,对一元概化理论(UGT)模型与多元概化理论(MGT)模型的应用进行了整合,选择了更为合理的自学考试信度指标并探讨了及格线决策信度的考察思路,为自学考试研究者进行概化理论研究提供一定程度的参考。Generalizability Theory （GT）, having put forward a new reliability viewpoint, is gradually applied to large-scale tests. In this article, we brought GT into Self-taught Examination as a guiding ideology of reliability analysis, put forward the researcll paradigm for reliability analysis of Self-taught Examination. We introduced general tasks and processes of reliability analysis, compared Univariate Generalizability Theory （UGT） with Multivariate Generalizability Theory （MGT）, and discussed the most appropriate reliability index and decision reliability of passing grade for Self-taught Examination. All of those above will provide positive reference to researchers who would like to adopt GT method in the area of Self-taught Examination.

关 键 词：概化理论 信度分析 自学考试 标准参照测验 

分 类 号：G424.74[文化科学—课程与教学论]