Particular Questions About Treatments
A simple example of treatment structure is an experiment in which a number of new strains of grass are being compared with a standard strain. This chapter considers an experiment on spring cabbage in which there were eight treatments arranged in four randomized blocks of eight plots each. The experiment was to compare three sources of nitrogen, each at two levels, with a control treatment of no nitrogen, with two replicates of the control in each block. The chapter provides information on several useful components of the treatment sum of squares (SS). Many statistical packages will estimate linear contrasts between treatments, which are specified in this way, either by partitioning the treatment SS in the analysis of variance into components to test for the significance of the contrast or by calculating the estimate of the comparison and the standard error of that estimate. The contrasts are specified by the appropriate coefficients of the treatment means.