Conducting Problem-Based Learning Meta-Analysis: Complexities, Implications, and Best Practices

Abstract

Almost fifty years ago Barrows claimed that problem-based learning (PBL) was broad enough that a single methodological description was not possible (1986). It has only grown more complex since. In addition to meaningful variations of problem-based learning there are several related problem-centered pedagogies, such as case-based learning, project-based learning, and inquiry-based learning among others. Even within PBL primary research is conducted using a wide variation in measurement approaches, with diverse audiences, in a myriad of disciplines. Some of these challenges are unique and still pair with common meta-analysis challenges such as multiple definitions of “control groups,” pre-experimental designs, and multiple treatment designs. This intersection of complexity and common challenges makes PBL meta-analysis research particularly difficult. This article will explore best practices for conducting meta-analysis including search strategies, inclusion/exclusion criteria, asking clear research questions of the primary source literature, coding schemes, analysis techniques including key decisions of random vs. fixed effects models, analyzing publication bias and interpreting, presenting, and discussing results. Beginning with comprehensive coverage of the PBL meta-analysis and meta-synthesis contributions to date and suggestions for new meta-analysis work in PBL, we will mostly be devoted to the best meta-analysis practices alongside their implications unique to a PBL context.