A duplicate or repeat performance of an experiment or experimental trial/run. It involves performing the same trial from scratch using new experimental material. This results in additional measurements of the response variable for the same treatment conditions and allows greater precision of the estimates obtaiend from them. In a designed experiment, a replicate may involve repetition of either a single trial, a set of trials, or the entire experiment.
In a cake-baking experiment, making a new batch of dough using the same recipe (trial/run) and baking a new cake constitutes a replicate of that recipe (trial/run).
Design of Experiments (DOE)