Digital libraries (DLs) are adapting to accommodate research data and related services. The complexities of this new content spans the elements of DL development, and there are questions concerning data selection, service development, and how best to align these with local, institutional initiatives for cyberinfrastructure, data-intensive research, and data stewardship. Small science disciplines are of particular relevance due to the prevalence of this mode of research in the academy, and the anticipated magnitude of data production. To support data acquisition into DLs - and subsequent data reuse - there is a need for new knowledge on the range and complexities inherent in practice-data-curation arrangements for small science research. We present a flexible methodological approach crafted to generate data units to analyze these relationships and facilitate cross-disciplinary comparisons.