Randomisation
Make assignment independent of prognosis
Randomisation assigns eligible units to conditions using a chance process with known probabilities. Simple, blocked and stratified schemes balance different practical considerations. Allocation concealment prevents foreknowledge of the next assignment and protects recruitment decisions. Random assignment differs from random sampling, which concerns how units enter the study rather than which condition they receive.
Use when an intervention can be assigned ethically and feasibly, and a causal comparison is the objective. Choose the assignment unit, sequence method and concealment process before recruitment begins.
Strengths
- Breaks systematic links between treatment assignment and baseline prognosis in expectation
- Supports design-based causal comparisons
Limitations
- Chance imbalance remains possible in finite samples
- Non-adherence, interference and attrition can complicate causal interpretation
Know the boundary
Alternation, birth date and researcher judgement are not random allocation; randomisation is not blinding.