THE USE OF NEAREST NEIGHBOUR (NN) MODELS IN FIELD TRIALS

Wilkinson's recent method of NN analysis for field experiments and the topic of NN balanced designs are described and discussed, indicating the types of situations where this method may be useful. An example is presented that shows the improvements that may be obtained by this method. It is concluded that having the opportunity to apply an NN analysis provides an "insurance" against some of the factors that may lead to the standard statistical analyses being inefficient.