ABSTRACT
In an orthogonal design all the information on the treatments is obtained from the intra-block analysis. That is, contrasts amongst the treatment parameters are estimated entirely from comparisons made within blocks. In a non-orthogonal design some of the information is obtained from these within-block comparisons, but the remaining information has to be recovered, using an inter-block analysis, from comparison between blocks. Different designs of the same size can be assessed, therefore, on the basis of the amount of information available from both within and between blocks. For this purpose efficiency factors, obtained from the intra-block ana lysis, will be defined and used to provide criteria for comparing different non-orthogonal block designs.
