In an experiment, we deliberately change one or more factors in order to observe the effect of these changes on one or more response variables. The design of experiments is an efficient procedure for planning experiments, so that the data obtained can be used to explain the changes on the response variables.
During the design of an experiment, the hypothesis to test needs to be considered first. When trying to test the effect of a certain factor (e.g. concentration of a chemical compound) on a certain response variable (e.g. growth rate of a macroalgae) it is important to have both "treatment" (addition of the chemical compound) and "control" (no addition of the chemical compound) groups. By having both groups it is ensured that the observed effect on the variable will be caused by the factor that was modified in the treatment group.
Once the necessary treatments are decided, it is also important to decide the number of replicates that are going to be used in the experiment. A replicate is a repetition of the same experimental conditions (treatment). Replicates are necessary to ensure that the changes that occurred in response to the variable, are indeed caused by the modified factor, and not due to changes in other factors or differences between individuals. Typically, the minimum amount of replicates (of the same treatment) that are necessary to consider a result reliable is three. With less than three replicates most statistical tests cannot be applied to the data. On the other hand, the more replicates, the more reliable the results will be. However, compromises in the number of replicates and/or treatments are usually required to reduce the complexity of the experiments.