英汉版和英语版词汇量测试异同及效度比较

Comparison of score difference and validity between English-Chinese and English-English version of the vocabulary size test

我们以228名英语专业学生为被试,对Nation的词汇量测试卷英汉版和英语版进行了比较研究,并考察了两者的信度和效度。结果显示,英汉版成绩显著高于英语版;两版本均具有较高的内部一致性信度、结构效度和效标效度,但英汉版信度高于英语版(.855 vs.816),前者各等级得分与总分的相关性皆高于后者(介于.333~.795之间,5K和8K除外),因子分析结果似乎也支持两版本的单维性(即只测量一个潜在变量——词汇知识),三组不同水平被试两版本词汇测试成绩均在.05的水平上差异显著,且随着语言水平的提高呈递增趋势。随着词频等级的增加,测试成绩大致呈阶梯状下降趋势,英汉版各难度等级成绩均显著高于英语版。两版本(特别是英汉版)词汇量测试卷提供给广大英语教师和研究者一种全新的测量工具,以考量学生书面接受性词汇知识。

The primary purpose of this paper is to provide further evidence to previous findings that the bilingual version of Nation＇s vocabulary size test（ VST） can be more efficient than the monolingual one. It conducts a comparison between EnglishChinese（ E-C） and English-English（ E-E） VST based on data from 228 English majors. Statistic results show that scores of E-C version are significantly higher than those of E-E one,which reflects a higher level of internal consistency reliability,structural and criterion-related validity. The correlation between scores of each frequency level and total scores of the former is also higher than those of the latter. The results of factor analyses reveal that single construct,presumably word knowledge,is underlying the two version tests. A one-way between-subject ANOVA also indicates that the two version tests can effectively distinguish English learners of different proficiency levels（ significantly different at. 05 level）. As word frequency increases,participants＇ scores roughly drop in a stair-step pattern though clusters of 1,000-word frequency levels provide a relatively meaningful difficulty order,which is less susceptible to the idiosyncrasies at each 1,000 level. The two versions,especially EC version,offers English teachers and researchers a new valid instrument to measure written receptive vocabulary size,if with some minor revision of the items in question.