Using forest plots to introduce meta-analysis, including simple moderator analysis, early in statistics education

Abstract

Meta-analysis is the quantitative integration of empirical studies that address the same or similar issues. It is usually the best way to draw research-based conclusions that can guide evidence-based practice by professionals, and evidence-based decision making by public policy makers. Meta-analysis is so important that students should learn about it very early in their statistics education. The close links between meta-analysis and practical conclusions drawn from bodies of research mean that meta-analysis is a vital element in outreach from statistics education. I describe software that uses forest plots to make the basic ideas of meta-analysis accessible, and my experience using it with beginning students. I use the software to illustrate two major models for meta-analysis, and introduce graphical extensions to forest plots that illustrate how the crucial topic of moderator analysis can be explained and, in simple cases, interpreted visually.