Effectiveness of Computer-Based Scaffolding in the Context of Problem-Based Learning for Stem Education: Bayesian Meta-analysis

Computer-based scaffolding plays a pivotal role in improving students' higher-order skills in the context of problem-based learning for Science, Technology, Engineering and Mathematics (STEM) education. The effectiveness of computer-based scaffolding has been demonstrated through traditional meta-analyses. However, traditional meta-analyses suffer from small-study effects and a lack of studies covering certain characteristics. This research investigates the effectiveness of computer-based scaffolding in the context of problem-based learning for STEM education through Bayesian meta-analysis (BMA). Specifically, several types of prior distribution information inform Bayesian simulations of studies, and this generates accurate effect size estimates of six moderators (total 24 subcategories) related to the characteristics of computer-based scaffolding and the context of scaffolding utilization. The results of BMA indicated that computer-based scaffolding significantly impacted (g = 0.385) cognitive outcomes in problem-based learning in STEM education. Moreover, according to the characteristics and the context of use of scaffolding, the effects of computer-based scaffolding varied with a range of small to moderate values. The result of the BMA contributes to an enhanced understanding of the effect of computer-based scaffolding within problem-based learning.