Paper short abstract:

We study Federal Statistical Research Data Centers (RDCs), exploring how scientific work is shaped by efforts to repair the RDC system. We find that efforts to democratize RDCs ignore key dynamics of data access (e.g. relationships), undermining the expansion of high-quality knowledge production.

Paper long abstract:

Recent legislation in the U.S. (e.g. the 2018 Evidence-Based Policymaking Act) aims to democratize access to public-sector data, with the aims of advancing equity and transparency and facilitating policy-relevant research. But in the face of legal, fiscal, and reputational uncertainty, efforts to streamline statistical infrastructures may undermine the scientific work they support.

In this project, we consider Federal Statistical Research Data Centers (RDCs), which are secure enclaves that offer access to sensitive data from multiple agencies. A network of RDCs operates across the country, enabling researchers to apply for access and conduct multi-year analyses. The application process is notoriously long and mysterious (“like a speakeasy,” according to one informant), requiring guidance from administrators who provide informal, iterative feedback.

Through interviews with civil servants, administrators, and researchers, as well as analysis of public meetings, documents, and (un)successful research proposals, we explore how knowledge production is affected by efforts to reconfigure the RDC system. We find that: (1) relationships and status networks shape not just access to data, but data work itself, (2) painful bureaucratic processes help manage risk and visibility, and (3) RDCs engage in deep boundary-work between science and policy, which shapes understandings of scientific merit and expertise. Efforts to democratize access (such as the web-based Standard Application Process portal, which limits the capacity of administrators to guide proposals) hinders all three of these dynamics. In attempting to repair the seemingly-broken RDC system, open-data initiatives may accelerate institutional collapse, and in turn, threaten the very goal of expanding knowledge production.