ABSTRACT
The experimental design is the complete sequence of steps involving randomization, assignment and application of treatments, collection of data, etc., for the experimental material, such that valid inferences may be drawn. In the design stage the consulting statistician should share the responsibility for making sure that objective inferences can be drawn, that valid estimates of experimental error can be computed, and that differences between treatments are in fact unbiased. The experimental unit is the smallest unit of experimental material to which a treatment is assigned by a single act of randomization. The unit of experimental material on which the observation(s) is recorded is referred to as the sampling unit. The sampling unit may be the same as the experimental unit; the entire experimental unit may be divided into sampling units; or one or a few sampling units may be randomly sampled from the experimental unit. The sampling unit is sometimes also referred to as the observational unit.
