The Open Science paradigm has brought the dissemination of experimental artifacts on the agenda of funding agencies, research institutions, and academic publishers. Managing research data is a crucial part of guaranteeing the reusability and reproducibility of published results. In this research, we suggest a conceptualization of reproducibility based on threats, risks, and vulnerabilities identified in current research data management (RDM) practices. By doing so, we can describe a range of threats to reproducibility and pinpoint areas where current RDM practices and the scholarly communication infrastructure insufficiently address these threats. Further, we elaborate on a socio-technical approach to reproducibility in RDM by collecting evidence from researchers and scientific publications. We show that the STS approach complements current IS research on RDM by offering a holistic view of reproducibility challenges in RDM.