chapter  4
18 Pages

A Simple Experiment

ByRoger Mead, Robert N. Curnow, Anne M. Hasted

Randomization must be used in the allocation of treatments to units. This chapter discusses the method of allocating treatments to these units. Random number generators are available in spreadsheets and most statistical software packages will generate simple experimental designs that automatically randomize the allocation of treatments to units. The chapter assumes that the variance is the same for the different treatments, which is equivalent to assuming that the effect of each treatment is to increase or decrease the yield of the unit to which it is applied by a fixed amount. In view of the potential difficulty of examining each pair of treatments, it seems desirable to test the overall variation between treatments to see whether the treatments, viewed as a group, are homogeneous. The completely randomized design with treatments is about the simplest experiment that can be analyzed statistically and there are obviously a number of extensions that can be made.