Results of earlier studies have shown that rating of prior stress exposure in preadolescent boys influenced the association between DRD2 genotypes and alcoholism risk factors, suggesting that variability in stress exposure, either in patient or control samples, could readily account for at least part of the confusion in DRD2 study outcomes. In order to test the hypothesis that the DRD2 A1 allele is only associated with alcoholism in subjects with elevated stress exposure, we examined the gene-stress interactional model in a sample of males of Mayan descent in the Olancho district of Honduras. Ascertainment was based on an epidemiologic, observational cross-sectional design, and the study was approved by the Institutional Review Board. A total of 309 adult males (age range 18-87 years) were interviewed by a physician or a public health nurse, blood samples were obtained for genetic studies, and participants were administered the short version of the Michigan Alcoholism Screening Test (S-MAST) and the Hispanic Stress Inventory (HSI). Three explanatory models were evaluated. The first model tested the effect of the demographic variables alone as predictors of MAST scores, the second tested the effects of stress and DRD2 genotypes separately, and the third tested the effect of the interaction between stress and the DRD2 genotypes. Neither model 1 nor model 2 yielded significant results; neither MAST scores nor HSI scores were found to be associated with DRD2 genotypes. However, Model 3 was confirmed reflecting a significant (P<.05) interaction between DRD2 genotype and stress score as a predictor of MAST score. Additionally, this difference was found to be largely accounted by the HSI occupational/economic stress score, which had a highly significant (P=.003) interaction with DRD2 genotype as a predictor of MAST score. This stress score was the only one of four that showed levels of stress as high as HSI scores in a US population. The MAST scores of A2A2 genotype participants were found to be nearly identical in low stress and high stress participants, whereas the MAST scores of A1A2 participants increased modestly with stress (P=.01) and that of A1A1 participants increased markedly with stress (P=.001). These findings support the hypothesis that DRD2 genotype-phenotype associations depend on the magnitude of stress exposure, and they lend support to the view that variability in DRD2 study outcomes may in part be explained by this gene-environment interaction.