This research aimed to evaluate and synthesize the effectiveness of problem-based learning (PBL) compared to traditional instruction (TI) in various learning outcomes. The study, using a sequential explanatory mixed research approach, reviewed 74 postgraduate dissertations from Türkiye, spanning from 2009 to 2021, that matched specific inclusion criteria. Qualitative data analysis revealed that the majority of these dissertations were master’s level and completed in 2010. These studies frequently implemented experimental interventions in middle school science classes, typically over five weeks with groups of 21-30 students. The quantitative analysis showed that PBL had a significant overall impact, with a General Effect Size (GES) of 0.734 across all variables. This effect was divided into specific clusters: knowledge (GES of 0.992), skills (GES of 0.696), and emotions (GES of 0.406). Further, PBL particularly affected academic achievement (GES of 0.842), scientific process skills (GES of 0.351), and course attitude (GES of 0.402), which were the most commonly assessed variables in the respective dimensions of knowledge, skills, and emotions. Additionally, PBL showed a substantial impact (GES of 0.799) on variables covering 21st-century competency framework skills. In summary, the study found that PBL is significantly more effective than traditional instructional methods across various learning outcomes, particularly in enhancing knowledge, skills, and emotional aspects of learning. This study suggests that integrating PBL activities can significantly enhance the attainment of curriculum objectives.