Statistical Experiments

Full topic overview

An experiment (AS91265) is how statistics tests cause and effect. Instead of just observing, you actively apply a treatment and use random allocation to split units into a treatment and a control group — which is exactly what lets you claim the treatment caused any difference, not just that they are linked. After running it you compare the two groups, then use re-randomisation to decide whether the difference you see is a real effect or something the random split could easily have produced by chance. The marks come from a fair design, a clear comparison, and a causal conclusion that is honest about chance.

Experiments & Observational Studies

What makes it an experiment→

Designing a Fair Experiment

Designing a fair experiment→

Analysing an Experiment

Covers AS91265

Statistical Experiments

Full topic overview

An experiment (AS91265) is how statistics tests cause and effect. Instead of just observing, you actively apply a treatment and use random allocation to split units into a treatment and a control group — which is exactly what lets you claim the treatment caused any difference, not just that they are linked. After running it you compare the two groups, then use re-randomisation to decide whether the difference you see is a real effect or something the random split could easily have produced by chance. The marks come from a fair design, a clear comparison, and a causal conclusion that is honest about chance.