The nature of the relationship between social contact and attitude similarity between twins was investigated using longitudinal data from a sample of Australian twins. Earlier research has suggested that social attitudes are not explained solely by shared environment; rather there are both genetic and environmental components that explain variance in social attitudes. Using three types of analyses we investigated the magnitude of the relationship and the direction of causation between attitude similarity and social contact. Longitudinal analysis of within-pair variance by level of contact suggests that attitude similarity leads to contact among the females and that similarity is both genetically and environmentally based. Analyses using a crosslag regression model suggest that similarity causes contact among MZ females. Biometrical analyses indicate differences in direction of causation for males and females. Among females, both genetic and shared environmental parameter estimates could be equated across contact groups, suggesting little relationship between contact and similarity. Among males, findings of smaller estimated heritability in the high-contact group suggest that similarity causes contact. However, an increased estimate of the contribution of shared environmental variance in the highcontact males could additionally suggest that contact leads to similarity.