基于多维语言学特征变量的汉语二语者写作水平评价研究  

A Study on Writing Proficiency Evaluation of Chinese Second Language Learners Based on Multidimensional Linguistic Feature Variables

作　　者：张漫莉 Zhang Manli(Institute on Educational Policy and Evaluation of International Students,Beijing Language and Culture University,Beijing 100083,China)

机构地区：[1]北京语言大学国际学生教育政策与评价研究院,北京100083

出　　处：《现代语文》2022年第8期81-87,共7页Modern Chinese

基　　金：国家社会科学基金重点项目“基于HSK大数据挖掘的汉语习得研究”(17AYY011);北京语言大学院级研究生创新基金项目“基于多维语言学特征变量的来华留学生汉语写作水平评价研究”(21YJ160008)。

摘　　要：基于韩国学生的HSK写作测试语料,采用多元线性回归分析法,从词汇和句法两个维度出发,以29个典型指标为基础,对韩国学生汉语写作指标进行探讨。通过考察学生的写作成绩与各指标的相关性,筛选有效参项以构建写作成绩评价模型,并对该模型的预测能力进行效度检验。研究表明,词汇对写作成绩的预测力要高于句法;虚词使用比率、最常用词比率、话题链数量、话题链分句数量和零形成分数量等指标,对写作水平具有较好的解释力度与预测效度。Based on the HSK writing test corpus of Korean students,correlation analysis and multiple linear regression analysis were used to explore the Chinese writing indicators of Korean students based on 29 typical indicators in two dimensions:vocabulary and syntax.By examining the correlation between students writing scores and each index,a valid covariate was selected to construct a writing score evaluation model,and the predictive power of the model was tested for validity.The study showed that vocabulary had higher predictive power than syntax;indicators such as the ratio of dummy words used,the ratio of most frequently used words,the number of topic chains,the number of topic chain clauses and the number of zero-formation scores had better explanatory and predictive validity for writing level.

关 键 词：汉语二语写作 写作评价 二语习得 韩国学生 

分 类 号：H195.3[语言文字—汉语]