基于校本测试的大学英语阅读群体及个性化诊断研究  

The Group and Individualized Diagnosis on EFL Reading in University-Based Final Test

作　　者：张海云[1] 石运章[1] ZHANG Hai-yun;SHI Yun-zhang(College of Foreign Languages,Shandong Agricultural University,Taian,Shandong 271018,China)

机构地区：[1]山东农业大学外国语学院,山东省泰安市271018

出　　处：《外语电化教学》2020年第5期79-85,12,共8页Technology Enhanced Foreign Language Education

基　　金：山东省社会科学规划研究项目(项目编号:18CWZJ36);山东省高校科研计划项目(人文社科类)(项目编号:J18RA230)的阶段性研究成果。

摘　　要：认知诊断评估在语言测试尤其是英语阅读领域的应用得到了越来越多的关注。该研究基于大学英语校本期末测试开展了认知诊断评估,运用认知诊断模型G-DINA对构建的阅读认知属性Q矩阵和6,242名被试在阅读理解部分的作答数据进行了诊断分析,得到了被试的认知属性掌握概率、潜质类型分布情况、试题答对概率等诊断信息,探索对被试进行阅读技能掌握情况的群体和个性化诊断,以期为教育管理者、教师和学生充分发挥大学英语校本期末测试的积极反拨作用提供启示。This paper aims to make group and individualized diagnosis on EFL Reading in universitybased final test and answer the following questions:(1)What is the attribute mastery probability of the participant groups and individuals?Are there any differences among participants of different levels?(2)What is the latent class distribution of the participants?(3)What is the item parameter estimates of the participants?The conduct of the study includes 3 steps.The first is to define the reading attributes with the method of literature research,expert judgement and think-aloud protocols.Four attributes were confirmed and they are complex vocabulary and language knowledge(A1),extracting explicit information(A2),connecting and synthesizing(A3)and making inferences(A4).The Q-matrix was constructed in the second step.The last step is to use a cognitive diagnostic model named G-DINA to analyze the data of the Q-matrix and the other4 matrixes and the reading comprehension response data of6,242 participants.The analysis under the G-DINA model provided 4 types of outputs:the model fitness to the data under absolute sense;the attribute prevalence of participants as a group,groups of different levels and individuals;the latent class distribution to exhibit the participants’attribute distributions and posterior probabilities;and the item parameters estimates.The model-data fitness result is an indication of appropriate fit.As for the attribute mastery probability of all the participants,the A2 attribute gained the highest prevalence(0.8694).According to the prevalence of the other 3 attributes(from 0.6594 to 0.6872),students in general have a good mastery of the reading skills.The mastery probability of participants with different levels shows difference in that the ones of high level have a good mastery of all the four attributes except A3 while the ones of the middle level excel a lot in the mastery of A3 and a little bit in A1.The participants of low level need to enhance their mastery of all the four attributes.The cognitive diag

关 键 词：校本测试 阅读 认知诊断评估 个性化诊断 

分 类 号：H319.3[语言文字—英语]