To address the controversy surrounding DRD2 and alcoholism, we performed linkage and association studies utilizing alcoholic men from high-density families largely uncontaminated by other psychopathology and female alcoholics for whom secondary drug dependence (averaging 10 years later onset) was a prominent feature. The males and females were combined for a total of 52 alcoholics, and compared to 30 controls screened for the absence of alcoholism and other psychopathology, revealing a significant association between the frequency of the TaqI A1 allele and alcoholism. However, linkage and family-based association studies conducted on 20 families of male alcoholics found no evidence for association or linkage between Taq A and alcoholism. The results of our population-based association study, placed in the context of the literature, suggest that minimizing psychopathology in control groups is probably a more important explanation for divergent results than either sampling error or population stratification. When combined with the complete lack of within-family evidence, we concluded that the association, while not appearing to be artifactual, is not specific to the alcoholism phenotype, per se.