Randomized Control Trials
Randomized Control Trials (RCTs) are experiments that randomly allocate participants between treatment and control groups. They are considered the 'gold standard' for impact evaluation.
Individual-level RCTs are impact evaluation design where the outcomes are measured on an individual basis. Randomization for individual-level RCTs are also done on an individual (per participant) level.
Clustered RCTs are a type of RCT in which randomization is done on the basis of a group i.e. cohort, villages, etc. This is the preferred type of RCT when the intervention is by definition applied at the cluster, rather than the individual level (for example an intervention that is targeted towards schools or health facilities in a given setting, rather than the students or patients who might attend these schools or clinics).
A key consideration in the design and analysis of cluster RCTs is that statistical power in cluster RCTs is typically lower than that for individually randomized trials, since outcomes within clusters are typically somewhat similar to each other. This means that the number of clusters in a cluster RCT, rather than the number of individuals who participate, is most relevant to the statistical power of the study. This also means that cluster RCTs are often more expensive than individually-randomized RCTs. However, an advantage is that cluster RCTs can enable measurement of spillover effects.
Roll-out of the intervention is randomized. This is typically done at the cluster-level. For example, an intervention is intended to treat 100 villages. 50 villages are randomly selected to receive interventions in year 1, and 50 villages are selected to receive interventions in year 2 (and therefore serve as a control group in year 1). A primary advantage of the randomized phase-in is that it is easily applied to project implementation schedules (as roll-outs typically happen over multiple years). A primary disadvantage is that once the intervention is fully rolled-out, there is no remaining control group, and thus no way to measure long-run effects (although long-run analyses can still examine differences between groups with degrees of exposure).
Back to Parent
This article is part of the topic Experimental Methods
- Impact Evaluation in Practice. Paul J. Gertler, Sebastian Martinez, Patrick Premand, Christel Vermeersch, Laura B. Rawlings.World Bank Publications, 2016. 
- JPAL's Introduction to Evaluations 
- Running Randomized Evaluations: A Practical Guide. Rachel Glennerster, Kudzai Takavarasha. Princeton University Press, 2013. 
- Evidence in Governance and Politics 
- Field Experiments: Design, Analysis, and Interpretation. Alan S. Gerber, Donald P. Green. W. W. Norton, 2012.