Numerous genes that might contribute to the development of diabetes mellitus and/or its complications have been isolated and characterized. One approach to determining whether these "candidate" genes influence susceptibility to diabetes is to compare the frequency of a DNA marker(s) (restriction-fragment-length polymorphism) for each gene is appropriately matched groups of patients and control subjects. The identification of a DNA-marker association would suggest that genetic variation at this gene may increase or reduce the risk of developing diabetes. However, the absence of an association does not necessarily imply that this gene does not contribute to the development of diabetes. We discuss the genetic rationale of disease association studies and the importance of sample size and disease-marker allele frequencies in these studies.