Randomised Controlled Trials in Educational Research

It has been more than twenty years since we lamented the lack of large, pragmatic, randomised controlled trials to evaluate novel educational policies. Whilst there was an evaluation landscape in education this was dominated by quasi-experimental designs that could not, by their very nature, enable the minimisation of selection bias. Consequently, even well conducted quasi-experiments have the potential to be misleading as their results may be influenced by selection effects. However, over the last 20 years there has been a sea change, with large pragmatic randomised controlled trials dominating the evaluation landscape especially in the United Kingdom (UK) and United States (US). Key funding agencies for educational evaluations in both countries encourage the adoption and prioritisation of ‘true’ experimental designs. Furthermore, the methodological quality of recent experiments is high with trial registration, large sample sizes, independent randomisation and blinded pre and post testing featuring in many if not most of the published examples. In this paper we summarise the massive progress in the quality and quantity of educational trial design over the last 20 years. We also describe innovative trial designs that have and are being used to control for various confounders that can still be present in a trial design despite adequate randomisation.