ABSTRACT
In row-column designs the experimental units are grouped in two directions, i.e. two blocking factors are used with one factor representing the rows of the design and the other factor representing columns. For instance, in an agricultural experiment the rows and columns of a design might be used to control field gradient and soil type respectively. In a preference testing experiment, the columns might represent different subjects and the rows the order in which different products are presented to the subjects. In an insulin-response experiment, the rows, columns and treatments may represent respectively individual animals, date of injection and levels of insulin. In all these experiments it is hoped that there will be a gain in the accuracy of estimating treatment comparisons resulting from eliminating the effects of the row and column factors.
