教育领域的自然实验因果推断研究 ——以语言教育研究的应用为例-清华大学教育研究

文章信息/Info

Title:: Causal Research Based on Natural Experiments ——Examples from Language Education

Author(s):: FENG Rui-ling¹² GUO Qian²; 1.Foreign Languages College, Tianjin Normal University;2.Department of Foreign Languages and Literatures, Tsinghua University

Keywords:: causal inference; natural experiment; difference-in-differences; language education

摘要:: 自然实验作为常见准实验设计之一,在国外语言教育研究中已应用二十多年,研究数量呈上升趋势且话题丰富;但在国内语言教育研究中应用极少。本文首先介绍自然实验概念、估算方法(DID)、数据、稳健性检验、DID拓展;然后评述国内外运用自然实验的语言教育研究,涵盖教学方法/项目、教学媒介语、教育帮扶、分流教育、教学时长等九个话题。本文旨在推动语言教育领域的跨学科研究,对其他教育领域的研究也具有参考价值。

Abstract:: Natural experiment, as one of the most common quasi-experimental designs, has been applied in language education research outside of China for over 20 years, increasing in both number and topic richness. However, its application is still rare in language education research in China. This study first introduces the natural experiment, including its core concepts, estimation method, data sources and types, robustness check, and extensions. Then we review language education studies that employ natural experiments for causal inferences. These studies fall into nine themes (e.g., teaching method/program, instructional medium, educational support, migrant language policy, tracking, early childhood education and care, and instruction time). This paper is aimed at promoting interdisciplinary research in language education. It also has implications for education research in general.