学科差异下的大学英语成绩——基于北京市七所高校本科生的分析  

College English's Grade under Different Majors:A Multilevel Analysis of Undergraduates from Seven Universities in Beijing

作　　者：傅强[1] 余昕[1] 田野[1] 唐晓雪[1] 李兰[1] 常国珍[1] 

机构地区：[1]北京大学人口研究所,北京100871

出　　处：《市场与人口分析》2007年第6期72-81,共10页Market & Demographic Analysis

摘　　要：在对北京市海淀区七所"211工程"高校计算机及经管专业大四本科学生问卷调查的基础上,及在将新、旧四六级成绩统一转化为可比的旧四级成绩后,利用一般MLE回归模型纠正、解释了学科差异与大学英语成绩间的生态学谬误。学科差异对于个人的英语成绩并不存在直接影响。还利用国外教育界经常采用的分层线性模型重新估计了学科差异对于我国大学生英语成绩的影响。在理工科学生组中,解释变量对于因变量的影响更为强烈。Based on the survey data of the senior students from both the department of computer science and the department of economics （or management） in seven ＂211 Project＂ universities in Haidian District, Beijing, we used ordinary least square regression model to interpret and justify the ecological fallacy relating to the effect of different majors on college English＇s grades at micro- level, which are converted from four different types of College English Tests （CET）. And we found that majors show no direct influence on the students＇ college English＇s grades at micro -level, which refutes the prevailing misconception that the student from social science usually has got better English grade. Besides, we first adopt hierarchical linear model to estimate the indirect effects of different majors on college English grades, demonstrating that majors have a stronger macro - level influence on the partial effect of micro - level explanatory variables in the engineering students, with comparison to those from social science, which could be served as an empirical basis for future study and teaching practice.

关 键 词：分层线性模型 大学英语四六级考试 调查 抽样 回归 

分 类 号：C921[社会学—人口学]