多层次线性分析模型在网络协作学习中的应用

胡勇

（北京师范大学 教育技术学院，北京100875）

【摘要】多层次线性分析模型就是针对传统线性回归分析方法的缺点发展起来的、可以同时处理跨个体和群体数据的方法。本文首先介绍了网络协作学习研究中收集的数据存在嵌套结构的原因及使用一般线性回归分析存在的缺点；然后简要概述了多层次分析模型的知识及其操作步骤；最后介绍了多层次分析模型在网络协作学习中的应用、需要注意的问题及常见的分析软件。

【关键词】网络协作学习；一般线性回归分析；多层次分析模型

【中图分类号】G434

【文献标识码】A

【文章编号】1007-2179（2011）01-0080-06

【作者简介】胡勇，北京师范大学教育技术学院远程教育研究所在读博士（lchuyong@126.com）。

The Application of Multilevel Analysis Model

in Online Collaboration Learning

HU Yong

(College of Eduicational Technology, Beijing Normal University, Beijing 100875,China)

Abstract: Even though general linear regression analysis is commonly used in online collaborative learning research, the use of this method must guarantee that data collected are independent of each other. However, in online learning environment, subjects can't be randomly assigned to groups, and learning tasks are generally carried out in groups. In addition, there are design effects and interaction effects, which result in the nested structure and intra-group correlation of data collected. If researchers deduce using general linear regression analysis methods will result in the inflation of type 1 error. Multi-level linear analysis model is developed to solve these problems and can be used to analyze data across individuals and groups. The authors first introduce the reasons why there exists nested structure on online collaborative learning research data, and the shortcomings of general linear regression analysis. We then give an overview of the multilevel model and analysis procedure. Finally, we introduce the application of multilevel analysis model on online collaborative learning, and point out frequently occurring problems and commonly used analysis software. 

Keywords: online collaborative learning；general linear regression analysis；multilevel analysis model