Experimental Designs for Off-station Agronomy Trials

Abstract

The aim of off-station agronomy trials is to determine, under a wide variety of conditions, which of the factors of production have a significant effect on yield and how the factors interact with each other as they affect yield. In designing these trials, several considerations come into play. They should be conducted at several locations so that the effects may be examined under a wide variety of conditions. They should be relatively small because of the limited resources of any research program, yet they should provide reasonable precision for estimating the effects of interest. For the practical reason, they should allow the use of large plots for some treatments and small plots for others. They should permit relatively uncomplicated statistical analysis of the results. The experimental designs presented here are an attempt to meet these restrictions. Each design requires only thirty-two plots. They permit the simultaneous examination of five or six factors of production. They provide at least nine degrees of freedom for estimating error. Some of the plans allow the trial to be split into one or more ways of convenience in applying treatments. In all of these designs each factor is tested at only two levels, or rates, a high level, and a low level. Usually, but not necessarily, the low level would be the absence of the factor. In some cases, it might be known that a minimum amount of the factor, such as nitrogen, must be added to produce any yield at all. In this case, the low level would consist of a minimum application of the factor. In some instances, a factor might have qualitative rather than quantitative levels. For example, drilling and broadcasting might represent a “method of seeding” factor. In this case on method, drilling would arbitrarily be called the high levels while the other method, broadcasting, would be called the low.