多水平模型在中学生数学考试成绩影响因素分析中的应用

Application of Multi-Level Model in the Analysis of Influencing Factors of Mathematics Examination in Middle School

摘要: 教育测量与教学评价是教学活动的重要组成部分，是教学领域中进行科学管理的重要手段和提高教育质量的有效保证。教学评价中的主要内容之一是考试成绩及其影响因素的分析。目前，用于考试(测验)成绩及其影响因素分析的方法主要有描述性分析、聚类分析、主成分分析、因子分析和多元线性回归分析等。然而，在教学评价中，往往需要分析具有层次结构的数据，如不同学校或同一学校不同专业学生的考试成绩。具有层次结构的资料不适宜用上述方法分析，宜用多水平模型。本文基于某地区60所中学1653名初中二年级学生数学考试成绩的样本数据，建立了数学考试成绩及其影响因素的多水平模型。结果表明：学校、性别和家庭受教育程度对学生的数学考试成绩有显著影响。

Abstract: Education measurement and teaching evaluation is an important part of teaching activities. One of the main parts in teaching evaluation is the analysis of examination results and its influencing factors. At present, the main methods include descriptive analysis, cluster analysis, principal component analysis, factor analysis and multiple linear regression analysis. However, the data with hierarchical structure is not suitable to be analyzed by the above methods. Based on the sample data, a multi-level model is established. The results showed that school, gender and family were significantly affected by education.