汉语二语者词汇丰富性与写作成绩的相关性——兼论测量写作质量的多元线性回归模型及方程
发布时间：2018-09-19 16:42
【摘要】：Read(2000)的词汇丰富性框架被视为测量写作质量的有效工具,本研究以该框架中的词汇多样性、词汇复杂性、词频概貌和词汇错误为观察维度,将原有的7种因素扩展为27种。选取北京语言大学“HSK动态作文语料库”中360篇作文为样本,首先考察汉语二语者写作成绩与各因素的相关性;在此基础上,检验现有框架及扩展因素对汉语二语者写作成绩的预测效度,筛选有效参项,建立写作评价模型和回归方程。研究表明,现有框架中的词汇错误比重及扩展因素中的词种数、常用词数可作为构建模型的重要参项。
[Abstract]:Read (2000)'s lexical richness framework is regarded as an effective tool to measure the quality of writing. The present study uses the lexical diversity, lexical complexity, word frequency profile and lexical errors in the framework as the observation dimensions, and extends the original 7 factors to 27. A sample of 360 compositions from the HSK dynamic composition Corpus of Beijing language and language University was selected to investigate the correlation between the writing achievement of Chinese L2 learners and various factors. This paper examines the validity of the existing framework and extended factors in predicting Chinese L2 writing scores, selects valid parameters, and establishes a writing evaluation model and regression equation. The results show that the proportion of lexical errors in the existing framework and the number of words in the extended factors, and the number of common words can be used as important parameters to construct the model.
，
本文编号：2250719
[Abstract]:Read (2000)'s lexical richness framework is regarded as an effective tool to measure the quality of writing. The present study uses the lexical diversity, lexical complexity, word frequency profile and lexical errors in the framework as the observation dimensions, and extends the original 7 factors to 27. A sample of 360 compositions from the HSK dynamic composition Corpus of Beijing language and language University was selected to investigate the correlation between the writing achievement of Chinese L2 learners and various factors. This paper examines the validity of the existing framework and extended factors in predicting Chinese L2 writing scores, selects valid parameters, and establishes a writing evaluation model and regression equation. The results show that the proportion of lexical errors in the existing framework and the number of words in the extended factors, and the number of common words can be used as important parameters to construct the model.
，
本文编号：2250719
本文链接：https://www.wllwen.com/wenyilunwen/yuyanyishu/2250719.html
