There is strong evidence for a genetic contribution to schizophrenia, but efforts to identify susceptibility genes have been largely unsuccessful because of the low power of individual studies. The authors’ goal was to evaluate the collective evidence for an association between the Val158/108Met polymorphism of the catechol O-methyltransferase (COMT) gene and schizophrenia. They performed separate meta-analyses of existing case-control and family-based association studies. Overall, case-control studies showed no indication of an association between either allele and schizophrenia, and family-based studies found modest evidence implicating the Val allele in schizophrenia risk. The pooled analyses of studies from diverse geographical regions may have obscured ethnic differences in patterns of genetic risk for schizophrenia. Stratification of the studies by ethnicity of the subjects yielded evidence for an association with the Val allele in case-control studies of European samples and, especially, in family-based studies of European samples. Case-control and family-based studies of Asian samples produced mixed results and, overall, little evidence for association. The results of the two types of association studies diverged somewhat, but the evidence from the family-based studies, although based on fewer reports, may be more accurate. The Val allele may be a small but reliable risk factor for schizophrenia for people of European ancestry, but the influence of this polymorphism on risk in Asian populations remains unclear. These results call for more family-based studies to confirm the association between COMT and schizophrenia in European samples and to clarify its contribution to risk in Asian samples. They also suggest that case-control studies should use methods of genomic control to avoid being confounded by population stratification.