Latin square

n. A within-subjects research design in which a square array of letters representing treatments occurs once in every row and every column so that each subject receives each treatment, but different groups receive them in different orders in an attempt to neutralize any order effects.