DataStaR (http://datastar.mannlib.cornell.edu/), a Data Staging Repository developed and maintained by Cornell University's Mann Library, was designed as a platform and related services to support data sharing among collaborators, as well as the eventual publication of data to permanent, domain-specific repositories and institutional repositories. As a staging repository, providing temporary storage for data (whether preliminary or final), DataStaR assumes no long-term responsibility for preservation of content. However, because one of the goals of DataStaR is to facilitate the publication of data to permanent, external repositories, it is critical that DataStaR's operations are consistent with digital preservation best practices. Toward that end, we consider how DataStaR's design and function map to the OAIS reference model, and use the Trustworthy Repositories Audit & Certification (TRAC): Criteria and Checklist as a framework for specifying system, policy, and documentation requirements to ensure that DataStaR is a responsible partner in the entire chain of preservation activities. We present a description of this process and a summary of the types of elements that are most important in a staging environment.