Meta-analysis reveals the effectiveness evaluation of blended learning models across different academic disciplines

With the rapid development of modern electronic information technology, blended learning has gradually gained attention and acceptance among universities and individuals. Most studies indicate that, due to the combination of teacher guidance and active student participation in blended learning models, their effectiveness often surpasses that of traditional, single-learning methods. However, different disciplines possess unique characteristics, and there is currently a paucity of research exploring the effectiveness of blended learning across various academic fields. This study aims to evaluate the effectiveness of blended learning models in different academic disciplines using meta-analysis methods, through a comprehensive quantitative synthesis of 18 relevant experimental and quasi-experimental studies. The results reveal that blended learning demonstrates a moderately high level of effectiveness in the humanities (g = 0.795, P < 0.01) and an exceptionally high level of effectiveness in science and engineering disciplines (g = 1.017, P < 0.01). Lastly, this study offers suggestions on effectively implementing blended learning, considering three aspects: discipline characteristics, learning resources, and evaluation criteria.