初中级汉语水平二语者口语产出质量评估研究

On the Quality Assessment of Spoken Chinese for Low-intermediate Level L2 Learners

本文以55名初中级水平的汉语学习者为研究对象,采用相关和多元线性回归的统计方法,考查了语言复杂度、准确度、流利度中的8个区别特征,即类符、形符、词汇复杂度、语音准确度、词汇准确度、语法准确度、语速、发音时间比与口语评分的关系。研究结果显示,8个区别特征中,类符、形符、语音准确度、语法准确度、语速和发音时间比6个区别特征均与口语评分显著相关,其中类符、形符与评分高相关,语法准确度、语音准确度、语速及发音时间比与评分中等相关,但词汇复杂度和准确度两个指标与口语评分相关不显著。此外,7个区别特征的组合(类符和形符分开)分别可以解释口语评分总变异的70. 9%和70. 8%,而且类符、形符以及语音准确度能显著预测口语成绩,其中类符最强,形符次之,最后是语音准确度。

Targeting at 55 low-intermediate level L2 learners,the paper conducts a correlation and linear regression analysis to explore the relationship between eight language features( type,token,lexical sophistication,pronunciation accuracy,lexical accuracy,grammar accuracy,speed and time ratio) and the score of spoken Chinese test. The results show that six features( type,token,pronunciation accuracy,grammar accuracy,speed and time ratio) are significantly correlated with the score. Specifically, the correlations between type, token and the scores are high. The correlations between grammar accuracy, pronunciation accuracy, speed, time ratio and the scores are moderate,but the correlations between lexical sophistication,lexical accuracy and the scores are not significant. Besides,the combination between seven features( type and token separate) can explain 70. 9% and 70. 8% total variance of the score. Moreover,type,token and pronunciation accuracy can effectively predict the score,with type making the most accurate prediction,then token and then pronunciation.