On the use of flipped classroom across various disciplines: Insights from a second-order meta-analysis

Flipped classroom has become a popular buzzword in the post-secondary education setting, and it is one of the most visible trends in smart learning environments. Alongside this popularisation comes the view that the flipped classroom is something desirable. Yet, many educators remain divided over whether flipped classroom is really an improvement over traditional approaches. This paper is the first to synthesise all relevant meta-analytic information using a second-order meta-analysis approach on the effectiveness of the flipped classroom on student learning outcomes. By synthesising the findings of multiple primary meta-analyses instead of individual empirical studies, a second-order meta-analysis can more accurately account for publication bias and generate a more robust result. The present study synthesised and analysed the quality of 15 primary meta-analyses that involved 156,722 participants in flipped and non-flipped conditions to provide the most exhaustive test of the flipped classroom approach on its effect on student learning outcomes in higher education to date. The mean random effect size, after trim-and-fill adjustment, was 0.37, p < 0.001 in support of flipped classrooms. To check the accuracy of the second-order meta-analysis results, we performed a study-level meta-analytic validation. We discuss possible contextual and methodological moderators.