Research on the genetic factors conferring risk for schizophrenia has not provided definitive answers. In the present review, we will discuss potential clinical and genetic limitations intrinsic to the strategies using a diagnostic phenotype. Among clinical factors, uncertainty of the phenotype is certainly a major limitation. Genetic problems include locus heterogeneity and the complex genetic architecture of the phenotype. Given these limiting factors, we will also discuss another hypothesis-driven strategy to uncover genetic risk: the use of quantitative measures (intermediate phenotypes) within more specific neurobiological mechanisms. As a clear example of all these issues and because of the longstanding involvement in the pathophysiology of this disorder, we will review the association of the gene for dopamine D2 receptors (DRD2) with diagnosis of schizophrenia and with specific working memory behavioral and brain activity phenotypes. We conclude by suggesting that hypothesis-free and hypothesis-driven are not mutually exclusive strategies and may provide information at different levels that are both useful and equally valid about genetic risk for a complex diagnostic entity like schizophrenia and for a complex phenotype like psychosis.