A mixed-meta analysis of research on problem-based learning in science education was the goal of this study. In this instance, document analysis utilizing the mixed-meta method—which combines meta-analysis and meta-thematic analysis procedures-was used to investigate both quantitative and qualitative data. Using CMA and MetaWin programs, studies from certain databases that fit predetermined criteria were analyzed for a meta-analysis. Of the 78 studies that were included in the study, the effect size was found to be large, with g = 0.909. It has been suggested that science education benefits from the problem-based learning approach. Within the framework of the established criteria, 15 researches based on document analysis were subjected to a meta-thematic analysis using the Maxqda program in accordance with content analysis. Codes were acquired based on specific topics. By calculating the compliance values of the relevant codes under the relevant themes, a good degree of agreement was found regarding the effects of problem-based learning in science education on academic achievement, the affective-social dimension, 21st century skills, and the negative aspects of problem-based learning in science education. The research results indicate that there is a consensus among the researchers regarding the benefits of using problem-based learning techniques in science education.