摘要
大规模教育质量监测数据有两个基本特征:采用抽样调查方式、使用多个试题版本。本文以PISA 2015数据为例,提出了利用大规模教育质量监测数据进行再分析的五个原则:优先使用官方数据、比较研究须选择具有可比性的群体、确保样本对目标总体的代表性、再分析时需确保有效样本量、警惕相关分析中选择效应的影响。同时还提出了三个解读数据分析结果时的注意点:重视系统性误差对统计分析结果精度的影响、全方位考察相关关系的成因以及避免机械地解读统计结果。
This paper first points out two basic characteristics of large-scale education quality monitoring data:using multiple test booklets and adopting sample survey methods.Then taking PISA 2015 data as examples,five basic principles in the use of large-scale educational quality monitoring data are listed:(1)give priority to official data analysis results,(2)choose comparable comparison objects,(3)ensure that the sample is representative of the target population,(4)ensure a sufficient sample size,(5)note the selection effect in the correlation analysis.Three problems that needed to be paid attention to when interpreting the results of data analysis are suggested:(1)the influence of systematic error on the accuracy of statistical analysis results cannot be ignored,(2)the explanation of the relationship should be placed in a broader context,(3)conclusions cannot be given only by the results of statistical analysis.
作者
张一平
许世红
张丽娜
Zhang Yiping;Xu Shihong;Zhang Lina(School of Psychology,South China Normal University,Guangzhou,Guangdong,510631;Guangdong Academy of Education,Guangzhou,Guangdong,510035)
出处
《考试研究》
2019年第4期47-55,共9页
Examinations Research
基金
国家社会科学基金教育学一般课题“PISA视角下中国经济发达区域的中学生发展基础素养状况及提升路径研究”(课题批准号:BHA170147)
关键词
大规模教育质量监测数据
PISA数据分析
结果解读
Large-scale Education Quality Monitoring Data
PISA Data Reanalysis
Interpretation of Data Analysis Results
