Explore · Making a comparison
Same numbers. Different groups.
Rearrange the same 24 markers. Keep the group sizes unchanged and see how a starting difference can still affect the comparison.
Invented arrangements, not trial data. The 24 markers stand for people with or without one abstract starting feature. No age, diagnosis, severity, treatment or outcome is assigned. Filled and outlined markers do not mean good or bad health.
24 markers / two groups of 12
Starting features only
Group A / 12 people
12 with the marked feature; 0 without
Group B / 12 people
0 with the marked feature; 12 without
Group size and group composition are different.
Allocation can affect a comparison
If people with an important starting characteristic are systematically placed in one group, later differences may reflect that characteristic as well as treatment. Equal group sizes do not remove that problem.
In a randomised trial, a planned chance-based process assigns treatment groups rather than the clinician or participant choosing the allocation.
Chance does not make people identical
Chance allocation helps reduce selection bias. It does not guarantee identical starting characteristics, particularly in a small study. Trial design and analysis can address important differences.
Not every clinical trial is randomised, and a control group may receive active usual treatment. The information for the actual study explains what each group receives.
What this illustration leaves out
The button cycles through three fixed, deliberately selected arrangements that could arise with equal-size chance allocation. It does not generate genuine random allocations or represent their probabilities. Real studies may use different ratios, more groups and more complex methods.
No response, survival, side effect or treatment benefit is shown. There is no better-performing group to find and no score to achieve. The aim is to see why the way groups are formed matters before outcomes are compared.
Continue exploring
Sources & clinical review
Clinically reviewed and approved by Dr AlHilali on 4 November 2026.