The underlying cohort, or study base, representing the group of subjects that gives rise to cases. Controls should be selected to represent the proportions of exposed and non-exposed persons in the source population to help lessen the introduction of bias.