汉语分级测试分数线划分研究  

Student Placement in TCSL:Multiple Regression Analysis

作　　者：罗莲[1] 

机构地区：[1]中央民族大学国际教育学院,北京100081

出　　处：《语言文字应用》2012年第3期117-125,共9页Applied Linguistics

基　　金：教育部人文社会科学研究项目基金资助的"对外汉语分级测试研究"课题的研究成果之一。项目批准号:09YJC740074

摘　　要：本文将教师对老生样组汉语学习表现的评价作为因变量,将新学期分级测试的分测验成绩作为自变量,建立多元回归模型。之后,将该模型应用于新学期入学的所有考生,确定考生的初始等级。最后,考察新生初始预测等级与教师评价等级的差异。结果表明,将多元回归方法与简化的以学生为中心的专家判定方法相结合,可以为下一步的口语分级测试提供考生的预测等级信息,有利于最终做出精确的分班决定。This paper sets up a multiple regression model,with the ranks given to the old students by the expert teachers at the end of the past semester as dependent variables,with subtest scores in placement test as independent variables.It then applies it to all the students in the new semester to obtain their predicted ranks.Then the teachers＇ ratings on the new students are collected,and are compared with their predicted ranks.As a result,it is found that the multiple regression method could be combined with student-centered expert judgment to provide meaningful information concerning predicted levels for oral test,so as to facilitate the final placement decision.

关 键 词：汉语作为第二语言 分级测试 分班测试 分数线划分 标准设定 

分 类 号：H195[语言文字—汉语]