机构地区: 海南大学外国语学院
出 处: 《海南大学学报(人文社会科学版)》 2006年第2期274-280,共7页
摘 要: 从统计学角度,运用SPSS和G ITEST软件分析了一套试卷的信度与结构效度,发现试卷的总体信度和客观题部分的信度离理想值差距甚远。各大题之间的相关性大都没有实质意义。5道大题实际只考了2个因素,即(1)书面表达能力;(2)听力与语言知识。用G ITEST软件对客观题部分所做的项目分析表明:32%的题目不合格,在难度和区分度方面存在问题。对词汇语法部分不合格的题目进行了分类研究。 The paper first makes an analysis of an achievement test, using two computer softwares-SPSS10.0 and GITEST, which exposes a lack of reliability in both the test as a whole and the 75 multiple-choice questions. Based on the above analysis, the paper continues to categorize and explore the causes of the 10 vocabulary and structure questions that are not good enough, which are representative of many other universities in China that follow the practice of “CET-oriented teaching”.
领 域: [语言文字]