A Meta-Analysis of Data-Driven Learning (DDL) in EFL/ESL Settings

• 발행기관: 한국교원대학교 뇌·AI기반교육연구소
• 저자명: Kyung-Ho Yoon Dong Ju Lee
• 간행물 정보: 『Brain, Digital, & Learning』제14권 제2호, 283~304쪽, 전체 22쪽
• 주제분류: 사회과학 > 교육학
• 파일형태: PDF
• 발행일자: 2024.06.30

국문 초록

This research is on the effect size (ES) of corpus-based data-driven learning (DDL) in EFL/ESL, which can show the effectiveness/efficiency of this particular instructional method relative to those of others. This meta-analysis methodology in instructed second language acquisition (ISLA) has been developed and established by Norris and Ortega (2000), Plonsky and Oswald (2014), etc. The initial search results in Education Resources Information Center (ERIC) reveal 5,165 research articles, but the final number is reduced to 46 (54 unique samples) after applying step-by-step inclusion criteria. The weighted mean ES (Hedges’s g) between the comparison and experimental groups is 1.11 (SE: 0.13), which is large. The weighted mean ES between the pre-tests and immediate post-tests is 1.81 (SE: 0.16), which is large, too. The delayed post-test analyses are also conducted. In addition, the present meta-analysis investigates the ESs influenced by the seven moderator variables (MVs). The above-mentioned results as a whole indicate that the ES of corpus-based DDL in EFL/ESL is much larger than that of the overall ISLA, and DDL may have some specific MV subgroups where it is more effective/efficient. These results suggest that more detailed research be conducted on DDL which looks promising as a whole.

목차

Introduction

Literature Review

Method

Results

Discussion

Conclusion

키워드

Meta-analysis effect size data-driven learning (DDL) corpus concordance