Effects of flipped language classrooms on learning outcomes in higher education: A Bayesian meta-analysis

Despite accumulated evidence demonstrating the effectiveness of flipped language classrooms in higher education, there is no quantitative examination of the extant empirical studies to draw a general conclusion. Based on Bayesian methodologies and 26 effect sizes, this study quantitatively examines empirical studies that investigated flipped language classrooms' effects on learning outcomes in higher education. Our results indicate a large overall effect in favour of the effectiveness of flipped language classrooms. Subgroup analyses indicated that intervention duration, target languages, outcome types, allocation, and school locations were significantly related to the variability in language learning outcomes. A low risk of publication bias was identified. This study concluded that the flipped language classroom was a promising pedagogical approach to promoting language learning. Findings provided insights into an evidence-informed application of flipped language classrooms, for example: (1) sufficient face-to-face time to maximise the effectiveness of flipped language classrooms; (2) making flipped design adjustments based on student responses during long-term intervention; (3) giving students pre-training of flipped language classrooms and showing them the underlying benefits; (4) flipping basic contents of language learning and teaching complex contents face-to-face; and (5) adopting scaffolding strategies like code-switching to scaffold lower achievers.