completely randomized design An experimental design in which the treatments are assigned completely at random so that each experimental unit has the same chance of receiving any one treatment. Any difference among experimental units receiving the same treatment is considered as experimental error. Hence, it is appropriate only for experiments with homogeneous experimental units, such as laboratory experiments, where environmental effects are relatively easy to control. CRD agronomic experimental design