ABSTRACT
The randomized complete block design is different from a completely randomized design. Instead of randomly selecting each sample unit from all the N observations, randomization is conducted within the specific test group samples. This design is very useful when experimental units can be grouped meaningfully. Such a grouping is designated a block. The objective of blocking is to have the within-block units as uniform as possible. This is not a new concept, for we encountered it in pairing in Chap. 4 for the matched paired t-tests. Variability among the sample units in different blocks will be greater, on the average, than variability among the units in the same block. Ideally, the variability among experimental units is controlled so that the variation among blocks is maximized and the variability within minimized.
