Study design

Randomisation

Assigning participants to groups by chance to reduce systematic differences between groups.

What it means

Definition

Randomisation. Randomisation is a major strength of controlled trials because it helps balance known and unknown confounders. It does not guarantee perfect balance, especially in small studies, but it improves causal inference.

What it can tell you

It can show how close the evidence is to a real human health outcome.

What it cannot tell you

It cannot prove that a treatment works in people unless human outcome data are available.

Headline trap

Why this matters in headlines

A randomised trial is usually stronger evidence for treatment effect than a comparable observational association.