The present report describes the basic epidemiological measures and concepts used in association studies, i.e. measures of disease frequency and the concept of risk, together with their internal relationships. Further, the different types of designs for HLA and association studies are reviewed with particular reference to estimation of association strength. It is shown that the odds ratio from a 2 x 2 table estimates the incidence density ratio which again under certain conditions estimates the relative risk, if the study is based on incident (new) cases. However, most HLA and association studies are prevalence-based, which implies that the relative risk and incidence density ratio cannot be estimated unless a series of conditions is fulfilled. Finally, it is stressed that a detailed description of study design and the age classes studied as well as the procedure of case ascertainment should be mandatory before comparisons between and pooling of estimates from different studies are performed.