Human longevity is a complex phenotype with a strong genetic contribution. The search for allelic variation that impacts on longevity often involves population-based association studies of long-lived individuals (centenarians and nonagenarians) and younger controls. A major methodological problem with this approach is that the cases and controls in these experiments are often born several generations apart. The inherent absence of "age matching" can lead to serious misinterpretations. For example, allele frequency differences between long-lived individuals and younger controls may not reflect an effect on longevity but rather a change in population structure over time or different gene-environment interactions across generations.