摘要
Flipped learning’s widespread adoption and multiple meta-analyses notwithstanding, there exists a scarcity of analyses scrutinizing its comprehensive influence on language learning results. This meta-analysis assesses flipped learning’s impact on language learning outcomes vis-à-vis non-flipped instruction, encompassing distinct moderators encompassing study attributes and instructional facets: target language, level of education, school location, treatment duration, classroom activities, and the pre-instructional comprehension check mechanism. Effect sizes were extracted from 70 eligible studies (involving 4,616 students across 19 countries). Cochran’s Q was used to detect moderator variable effects, and post hoc comparisons were performed if significant differences existed among moderator subsets. Results yielded an overall mean weighted effect size (flipped learning vs. non-flipped instruction) of 0.96 under the random-effects model, indicating a large positive effect on student learning outcomes. Treatment duration, school location, and level of education were influential moderators affecting language learning results, while the other three were not. Flipped learning, particularly when implemented for over 6 weeks, consistently yields superior learning outcomes across different educational contexts. Educators are thus encouraged to adopt flipped learning as a versatile pedagogical approach to maximize student learning potential.