ABSTRACT
The analysis of variance (ANOVA) is used by agronomists to compare the means of treatment groups in field trials. To perform an ANOVA, the treatment groups are assumed to be randomly sampled, normally distributed, and have equal variances. This chapter proposes such a procedure involving a spatial statistical technique known as autoregressive response (AR) modeling. It advocates the method of analysis of field experiments for ANOVA of agronomic field trials because, through its specification of the geographic configuration of areal units, the AR model accommodates plots that are located on the nodes of a regular lattice. The chapter provides guidance for the use of AR-based ANOVA and accompanying SAS computer code for the analysis of agronomic field data. It shows clearly the basic difference between AR-based ANOVA and conventional ANOVA with regard to the results from the 102-plot variety trial.
