ABSTRACT
By definition, a block consists of a group of experimental units, i.e., subjects, such that the variability within each block is less than the variability among all the blocks. There must be at least two blocks for inclusion in a statistical analysis and three conditions must be met in order for blocking to be valid
1. There must be at least one treatment with at least two levels and one nuisance variable that can be classified into at least two
levels. For pretest-posttest designs, the nuisance variable is the pretest score.
