首页 | 本学科首页   官方微博 | 高级检索  
     检索      

自学考试的题型优化——基于多元概化理论的思想
引用本文:田霖,王桥影.自学考试的题型优化——基于多元概化理论的思想[J].教育与考试,2012(4):29-33.
作者姓名:田霖  王桥影
作者单位:北京教育考试院,北京,100083
基金项目:基金项目:本文为北京市2011年财政项目“高等教育自学考试标准化题库建设”(PXM2011_014256_111831)的研究成果之一
摘    要:概化理论提出了新的测量信度观并逐渐应用于大规模考试领域。本文运用多元概化理论对高等教育自学考试课程《综合英语(四)》听力测验的测量信度、测验总分合成、题型设计优化等问题进行探讨。研究发现:本次听力测验的总体测量信度较高,但"短文理解"分测验的测量信度较低;各分测验对全域总分的方差贡献比例与测验编制时的赋分比例存在一定差距;在删除"短文理解"分测验后将各分测验题量同时增至10题,可有效提高听力测验的测量信度。

关 键 词:多元概化理论  可靠性指数  听力测验  自学考试

Item Format Optimization of the Self-taught Examinations Based on MGT
Tian Lin Wang Qiaoying.Item Format Optimization of the Self-taught Examinations Based on MGT[J].Education and Examinations,2012(4):29-33.
Authors:Tian Lin Wang Qiaoying
Institution:Tian Lin Wang Qiaoying
Abstract:Generalizability Theory (GT for short) puts forward a new reliability viewpoint and is gradually applied to large-scale tests. In this article, we adopt Multivariate Generalizability Theory (MGT for short) to appraise Listening Test's reliability, combination of scores as well as item format optimization based on data of Comprehensive English Four. The result shows that: a) the reliability is acceptably high for the whole Listening-Test while low for Passage-Listening-Comprehension; b) the ratio of each subtest's contribution to total variance is different from test-setters' anticipation to some degree; c) to promote test reliability, it is recommended to increase the sample size of each subtest to 10 simultaneously after abandoning Passage-Listening-Comprehension.
Keywords:Multivariate Generalizability Theory  dependent index  Listening Test  the Self-taught examinations
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号