Evidence-based Education: The development of a model to use protocols and small-scale aggregated trials to create a prospective cumulative meta-analysis as an evidence base for interventions.

Abstract

In education, there has been a worldwide increase in the use of evidence in education to inform policy and practice. In the USA, bodies such as the Institute of Educational Sciences’ (IES) What Works Clearinghouse (WWC), the Best Evidence Encyclopaedia (BEE) and the Department General Administrative Regulations (EDGAR) and in the UK the Education Endowment Foundation (EEF) have been established to inform education decision making. The global increase in the use of evidence in education is based on the premise that if programmes are selected on the basis of more robust evidence, using these tested interventions should increase the chance of positive outcomes if they are deployed in other schools and contexts.
The aim of this thesis is to explore evidence-based education in the literature review before proposing a new theoretical model for evidence generation, introducing the use of protocols and small-scale aggregated trials linked to a prospective cumulative meta-analysis (PCM). The application of a PCM allows replication as a way to test how likely the intervention effect sizes will translate when they are tested using large scale randomised controlled trials and also as a method to test stability and improve dissemination after the initial research. The thesis includes two primary research studies for online cross-age peer tutoring across the transition boundary between primary and secondary schools and online small group teaching. The purpose of the trials were to test the implementation of the methodology for the use of protocols and small-scale aggregated trials linked to a prospective cumulative meta-analysis (PCM). Study One focuses on online peer tutoring across the transition for primary and secondary pupils and Study Two investigates the effectiveness of online small group teaching for mathematics.
The thesis demonstrates how the use of the model can be used to increase replication in the testing phase of an intervention, using empirical evidence from the online peer tutoring trials involving data collected across three cohorts of schools in an academic year. The impact of this research will provide an alternative testing framework for deciding in future trials if the evidence is robust for the commissioning of large scale randomised controlled trials in education.