Previous studies testing the dopamine D2 receptor (DRD2) locus for association with alcoholism have produced conflicting results. Failure to screen controls for substance abuse and failure to assess alcoholics for severity have been proposed as causes for the inability of some studies to detect an association. We have reevaluated the involvement of DRD2 mutations in susceptibility to alcoholism with a cladistics-based association analysis after restricting an alcoholic sample to more severe cases. For the present study we tested 55 alcoholic probands and 80 normal controls for differences in the frequency of six haplotypes at the DRD2 locus. The haplotypes were derived from five di-allelic polymorphisms spanning all but the first exon of the DRD2 gene. A cladogram constructed from the haplotypes provided the evolutionary context for a nested statistical analysis. We found no significant evidence for association of the DRD2 haplotypes analyzed with the more severe alcoholic phenotype.