基于英语网络测试系统的自适应测试与常规测试的对比实验研究  

A Comparison between the Conventional Test and the Network-based CAT Test of English

作　　者：韩少杰[1] 李新涛[1] 

机构地区：[1]西南财经大学语言文化学院,四川成都610074

出　　处：《外国语言文学》2006年第4期250-253,共4页Foreign Language and Literature Studies

摘　　要：理论研究表明基于题目反应理论的自适应测试具有准确、高效的特点。目前为数不多的实验研究多集中于测试效率的探讨上。本文以成都某高校非英语专业的41名大学一年级学生为研究对象,设计了自适应测试与常规测试的对照实验,实验结果表明自适应测试的测试信度、预测效度均要优于基于经典测试理论的常规测试。Available studies show that the “item response theory”-inspired CAT(computerized adaptive test) enjoys advantages such as greater accuracy and higher efficiency over other methods of testing.However,most experiments have so far failed to take into serious consideration the guiding theory,focusing on the efficiency of CAT only.This paper analyzes the findings of an experiment involving 41 first-year non-English majors of a university in Chengdu which was designed to compare CAT with the conventional test.The data show that as far as validity and reliability are concerned,CAT is superior to CTT(classical test theory)-based conventional test.

关 键 词：自适应测试 常规测试 题目反应理论 经典测试理论 信度 效度 

分 类 号：H319[语言文字—英语]