Introduction

Impact evaluation has a choice of approaches for assessing effect and confronting how much of that change could be due to the intervention being evaluated (attribution). This returns us to the territory of the counterfactual (Chapter 5) where, among the many ways for tackling this, fully experimental methods – that is, randomised control trials (RCTs) – continue to be seen as the purest expression.

Those seeking evidence may refer to hierarchies of evidence quality for counterfactual designs, and these consistently place RCTs at the top of the list as the ‘gold standard’ (Cook and Campbell, 1979). But, RCTs do not lend themselves to evaluating all situations, and there is a persisting debate and controversy about where they best fit, as well as how much design flexibility is possible without compromising validity. Nonetheless, they have a major role to play in purposeful evaluation as a prospective preferred choice for evaluators.

The appetite for RCTs among many commissioners and users is undiminished and will remain so. However, their enthusiasm is not always grounded in a recognition of their resource and other demands or limitations on effective practice. From time to time, most evaluators will likely face situations where they need to educate and guide funders and users on more realistic aspirations of RCTs alongside alternative options (covered in Chapters 9 and 10).

RCTs and the gold standard

Without taking any credit away from Dr Lund and his maritime experiment for preventing scurvy (Chapter 1), RCTs have rather less deep historic foundations. Their roots are set in the combination of treatment and non- treatment (control) groups, robust management, and randomisation and appropriate ‘blinding’ which can be traced to the start of the 20th century in the US.