There are countless reports—some in this ARCHIVES—about the association between candidate genes (genes encoding products with neurobiological function such as dopamine receptors, tyrosine hydroxylase, and tryptophan hydroxylase) and a host of behavioral traits ranging from normal personality variants to psychopathologic conditions. Although conflicting results and other inconsistencies have spawned wary commentaries,1-7 the tide of ‘positive' results continues unabated. In what follows, I highlight pivotal issues in the design and interpretation of association studies, with an eye toward potential pitfalls.