The use of generalized linear mixed models (GLMM) is widespread and growing. Perhaps the most utilized GLMM is the binomial-normal, which involves the measurement of an outcome on the binary scale and incorporation of random effects, assumed to be normally distributed, on the link function scale. In this talk, we investigate the performance of GLMM using the SAS® GLIMMIX procedure for small, multi-location studies consistent with the design used in veterinary biologics and medicine studies. Traditional methods, such as log-linear analyses, are included for comparision.