摘要
相关分析和平均数差异比较是外语教学研究中很常用的两种统计手段.然而对两份国内较有影响的外语教学研究类学术刊物近三年刊登的实证研究文章分析表明,许多研究者在这两种方法的使用上存在严重问题,主要表现为:对相关程度或平均数差异未作显著性检验,凭主观印象或错误地应用某些衡量标准作判断;作了显著性检验,但忽视了其实质,依照研究者预先的主观愿望,对数据作出了错误的分析.作者指出外语教学研究者应该深刻领会假设检验的基本原理,提高外语科研的质量.
Correlation analysis and means comparison are two frequently used statistical procedures in FLT research. However, a survey of empirical studies published in two journals in recent three years in China revealed that quite a few researches were seriously flawed because of errors in the employment of the procedures, which manifested themselves in two aspects: 1) failure to test the significance level of correlation coefficients or means differences while drawing conclusions through researcher's objective judgment or misuse of certain measurements; 2) ignorance of the essence of significance tests, which lead to misinterpretation of quantitative results. The author proposes that FLT researchers should have a deeper understanding of hypothesis testing to improve the quality of empirical studies.
出处
《山东外语教学》
2004年第2期62-64,共3页
Shandong Foreign Language Teaching
关键词
显著性检验
相关系数
平均数差异
significance level test
correlation coefficient
means difference
