Using Bayesian Meta-Analysis to Explore the Components of Early Literacy Interventions

Improving literacy instruction so that all students achieve grade-level proficiency in reading and writing 
remains a critical challenge in the United States. In this report, which is primarily aimed at researchers 
rather than education practitioners, the What Works Clearinghouse (WWC) applies two methodological 
approaches new to the WWC that together aim to improve our understanding of how early literacy 
interventions may work to improve student outcomes. First, this report pilots a new taxonomy developed 
by early literacy experts and intervention developers as part of a larger IES effort to develop standard 
nomenclature for the components of literacy interventions. Most, if not all, education interventions include 
multiple components that, when implemented together, are meant to improve student outcomes. These 
components might also be thought of as “active ingredients” or “key features.” Then, the WWC uses 
Bayesian meta-analysis—a statistical method to systematically summarize evidence across multiple studies
that is new to the WWC—to estimate the associations between intervention components and intervention 
impacts. If some components are associated with positive intervention effects, researchers and innovators
might be able to use this information to refine interventions, or develop new ones, and then design studies 
to test whether the interventions provide even greater benefits to students. The purpose of this 
methodological report is to explore to what extent observed components of interventions can explain 
which interventions have positive effects and develop take-aways for future applications of similar 
methods. The reportuses studies of early literacy interventions to conduct this trial of the new method