Abstract
INTRODUCTION: With the educational reform for skills in the 21st century, a large number of scholars have explored project-based learning. However, whether project-based learning can effectively improve the learning effect of students has not yet reached a unified conclusion. METHOD: This study uses a meta-analysis method to transform 66 experimental or quasi-experimental research papers based on project-based learning over the past 20 years into 190 effect values from the sample size, mean, and standard deviation of experimental data during their experiments, and to conduct in-depth quantitative analysis. RESULTS: The results of the study showed that compared with the traditional teaching model, project-based learning significantly improved students' learning outcomes and positively contributed to academic achievement, affective attitudes, and thinking skills, especially academic achievement. DISCUSSION: The results of the moderating effects test indicated that the effectiveness of project-based learning and teaching was influenced by different moderating variables, including country region, subject area, type of course, academic period, group size, class size, and experimental period : (1) from the perspective of country geography, the effects of project-based learning in Asia, especially in Southeast Asia, were significantly better than those in Western Europe and North America; (2) in terms of curriculum, project-based learning promotes student learning effects more significantly in engineering and technology subjects, and is better applied in laboratory classes than in theory classes; (3) from a pedagogical point of view, project-based learning is more suitable for small group teaching, in which the group size is 4-5 people teaching the best results; (4) in view of the experimental period, 9-18 weeks is more appropriate and has more obvious advantages for application at the high school level.