The main aim of the analysis of groups of experiments is to produce information that can be applied to a population either in space or in time and in both. The aim of this paper is to apply the specific error technique in order to solve the heterogeneity problem of the interactions of experiments with treatments, and to obtain the expectations of the mean squares for the considered effects in the analysis of variance. The data that were used refer to a group of 11 experiments, in a 3³ factorial design with sugar cane, conducted at different places of Sao Paulo State, Brazil. The results showed that the specific error technique is adequate when the heterogeneity is present in the interactions.
interaction; heterogeneity of variances; orthogonal contrast; groups of experiments
