从写作测验信度研究看开放式测评的发展趋势——概化理论的视角  

The Developmental Trend of Open-Ended Assessment Based on Reliability Research of Writing Tests:A Perspective of Generalizability Theory

作　　者：冯瑞龙[1] 朱宇[2] 

机构地区：[1]北京师范大学香港浸会大学联合国际学院,中国珠海519085 [2]厦门大学海外教育学院,中国厦门361102

出　　处：《海外华文教育》2014年第2期160-167,共8页Overseas Chinese Education

基　　金：UIC Research Grant(2012-13)R201307;教育部留学回国人员科研启动基金资助项目

摘　　要：开放式测评是测评界最新动向之一——真实性测评依据的主要题型,也是能给测评的各种用家提供丰富信息的重要测评方式。但同时,其信度较低这一问题也一直困扰着测评界。本文梳理、比较了当代测评界共存的经典测量理论、项目反应理论和概化理论三大信度流派,发现基于概化理论的研究结果通常能对测评各种来源的误差有较好的估算,从而可以为试卷及阅卷的设计与改进提供有价值的参考,并最终达到改善测评信度的目标。在此基础上,本文以写作测验为例,述评了近20年来概化理论在国内外开放式测评信度研究的应用,探索了其发展动向,力求为今后的相关研究拓宽视野。Subjective test has a long history, a strong vitality, and a prosperous future. Meanwhile, its low reliability, considered as its ＇chronic disease＇ , has been disturbing practitioners and researchers in the field of educational measurement. This paper reviewed the three major schools of reliability theory, and clarified that Generalizability Theory is a forward -looking theory which can offer approaches of im- provements based on estimates of variance components of various effects potentially lowering the reliability of a test. Implications to the development of subjective tests were briefly discussed based on a literature review of writing assessment studies from the perspective of Generalizability Theory.

关 键 词：开放式测评 信度 经典测量理论 项目反应理论 概化理论 

分 类 号：G449[哲学宗教—心理学]