We examined the utility of a broad framework that separated positive, negative, and ambivalent social network members. One hundred thirty-three young and older participants completed the social relationships index, measures of mental health, and a cardiovascular reactivity protocol. Results replicated prior research on the beneficial influence of positive (supportive) ties on psychological outcomes. More important, analyses also revealed that the number of ambivalent network ties predicted age-related differences in depression and sympathetic control of heart rate reactivity during stress. The statistical interactions between age and ambivalent ties on cardiovascular responses during stress were not changed when statistically controlling for other social network categories, demographic variables, and various personality factors. These data suggest that social network ambivalence was a relatively unique predictor of cardiovascular reactivity and highlight the utility of separating the variance due to positive, negative, and ambivalent network ties. Implications for the study of social relationships, physiological processes, and health outcomes are also discussed.