Randomised controlled trials

Randomised controlled trials (RCTs) are the gold standard for evaluating the causal impact of interventions, yet their successful implementation requires careful planning from the outset. This chapter provides a non-technical introduction to RCTs, guiding researchers through the key considerations for
designing rigorous impact evaluations. It begins by explaining the fundamental logic of random assignment and why alternative approaches — such as before and-after designs and comparisons of non-random samples — fall short in establishing causal inference. The chapter then describes four types of randomisation commonly used in social science research: simple, clustered, phase-in, and stratified designs, highlighting the conditions under which each
is most appropriate. It further covers data collection strategies, the core components of power calculations, and practical sample size examples using Stata. Finally, the chapter addresses two critical implementation challenges —
Hawthorne effects and spillovers — and strategies to mitigate their threat to internal validity.