An Application of the Augmented Randomized Complete Block Design to Poultry Research ,

Avian health researchers are often faced with the need to evaluate large numbers of potential disease agents, but limited facilities and the desire to minimize animal usage and costs prohibit large, all-inclusive experiments. Plant breeders encounter a similar problem with variety screening experiments in which a large number of varieties must be evaluated, but the supply of seed may be limited. In that situation, the use of augmented experimental designs has proven effective for obtaining the replication needed for valid statistical analysis. Those designs rely on the use of multiple trials consisting of different sets of plant varieties; standard varieties are included as controls or checks in every trial to provide an estimate of experimental error. Although the comparisons among varieties are not all equally precise and the estimate of experimental error may be biased, use of augmented designs can reduce the time and number of experimental units needed to isolate the most promising new varieties for further in-depth study. Augmented designs can similarly be applied to screening experiments in poultry research. This paper describes the augmented randomized complete block design and its application in an avian health setting. Data collected in an exploratory study of infectious enteritis in chickens illustrate details of the statistical analysis. Suggestions for implementing this analysis using the SAS statistical analysis package are also given.