Giving social support is lauded in society and regarded as contributing to the development and maintenance of social relationships, but research has typically focused on the benefits of receiving such support and the costs of giving. An emerging literature has begun to show the potential for giving to yield health benefits, but these studies have examined diverse possible benefits (e.g., relationship quality, mortality) and documented highly variable effect sizes. The literature could benefit from a theory-driven meta-analysis to organize and evaluate the robustness of these effects. However, the only existing theory addressing the benefits of giving support, the Caregiver Systems model (Konrath & Brown, 2013), suggests that giving support buffers stress during discrete stressful experiences. However, it does not address giving that occurs outside of stressful experiences, and does not account for the potential direct salutary effects of giving on stable health factors that help individuals prevent and manage future stress, including social relationships, social resources (e.g., trust and reciprocity), and personal resources (e.g., self-esteem). The present meta-analysis tested relationships posited via a new integrative framework – the Roles and Resources model (LaPorte, 2016) – to explain the variable levels of associations among giving and relationships, resources, and stress within and outside of stressful contexts. Results showed giving is (a) not significantly correlated with social relationships; (b) correlated with social and personal resources; and (c) unrelated to stress, even in high stress contexts. These findings provide incomplete but essential support for the Roles and Resources model and do not support the Caregiver Systems model’s stress-buffering framework. Implications of these findings for future research and application are discussed.