Replication
Replication means assigning each treatment to more than one experimental unit, so treatment effects can be told apart from unit-to-unit variation.
Replication is about the number of units per treatment inside one study, not about repeating the whole study later. For example, giving a drug to 30 patients and a placebo to 30 patients replicates each treatment 30 times, while giving each to a single patient does not replicate at all. Without replication, one unusual unit looks exactly like a treatment effect. More units per treatment also makes the comparison between the groups more precise.
Where this comes up
More collecting data and study design terms, or browse the full statistics glossary.