How to compare multiple intervention effects: Educational research using network meta-analysis

Abstract

ABSTRACT This study aims to introduce network meta-analysis (NMA) to provide educational researchers with an extended view of the reviewing educational research. Meta-analytic methods have been widely used in educational research reviews. However, weaknesses have emerged in the multi-group comparison analysis of educational studies where different interventions are applied because general meta-analysis is structured for a two-group comparison (i.e. treatment versus control group) of the same structure. Alternatively, when synthesizing experimental results of previous studies, NMA aims to verify related effect sizes by considering various experimental environments and research variables. Using NMA with simulation data, along with five types of online learning designs, this study describes how educational researchers could compare complex intervention effects between ten networks derived from five groups. The use of NMA in review research can provide rich interpretations of multi-group comparison results. Additionally, it can reflect on the complexity of real-world experimental design conditions.