Data from an early study of alcoholism among half‐siblings of alcoholic probands (17) was reanalyzed using log‐linear analysis. The original study, which used chi‐square analysis, identified genetic factors but found no environmental effects that contributed to the prediction of alcoholism. Log‐linear analysis, however, reveals that being reared with a proband is associated with a reduced incidence of alcoholism in half‐siblings. The results do not disconfirm the presence of genetic factors, but they do indicate that family dynamics make an independent and statistically significant contribution to the development of alcoholism not heretofore demonstrated. Similarities between these results and the results from studies of identical twins are