Completely Random Design

When the treatments are tested under homogeneous conditions of the experimental material, that is, the experimental units are homogeneous, and when the treatments are assigned to these experimental units randomly, then the candidate model is the Completely Randomized Design (CRD). In this design, it is assumed that if we have n=tr, homogeneous UEs, then the experimental units can be partitioned into t treatments of r units each. The CRDs are the simplest designs in which treatments are assigned to experimental units completely at random. This allows each experimental unit to have the same probability of receiving any treatment. The main objective in this design is to compare the means of the response variable between treatments that is achieved with analysis of variance (ANOVA) and multiple mean pair comparisons.