TEM4语法词汇项目的构念效度研究--基于Rasch和CFA模型的分析  

Investigating the Construct Validity of the TEM4 Grammar and Vocabulary Test through the Rasch Analysis and Confirmatory Analysis

作　　者：彭康洲[1,2] 邹申[3] 

机构地区：[1]北京师范大学外国语言文学学院,北京100875 [2]重庆师范大学外国语学院,重庆400047 [3]上海外国语大学英文学院,上海200083

出　　处：《外语与外语教学》2012年第6期49-55,共7页Foreign Languages and Their Teaching

摘　　要：本文通过调查TEM4语法词汇项目的潜在结构和影响构念的因素来评价该项目的构念效度。采用专家判断法收集内容效度的证据,Rasch模型分析题目作答反应对构念效度的影响,并借助验证性因子分析了解项目的构念构成。研究结果显示,内容考点判断非常一致;绝大部份题目拟合模型,项目有较好的单维性特征,题目作答基本上相互独立,不过,项目相对于考生的能力来说比较难,从而导致少数难度较大的题目不拟合模型,对项目的构念贡献不足;语法和词汇属于两个非常相关但是又相互独立的构念。The study evaluates the construct validity of the Grammar and Vocabulary Section of TEM4 by investigating the underpinning structure of the test, possible construct under-representation and construct-irrelevant threats. To this end, content-related evidence was gathered by expert judgment; construct representation and relevance were explored by Rasch anal- ysis and confirmatory factor analysis. The findings revealed that independent judgments yield consistent results on content-related evidence;construct-irrelevance is absent thanks to the adherence of data to unidimensionality, reasonably local independence, and good measurement properties of most items, but the subtest is relatively difficulty for the sample, which might attribute to some erratic item responses and construct under-representation; grammar knowledge and vocabulary knowledge are two highly correlated traits and few item responses are likely to be depend- ent due to some correlated error terms.

关 键 词：语法词汇项目 RASCH CFA 构念效度 

分 类 号：H0[语言文字—语言学]