Data integration is undertaken for the public good, yet institutions rarely address structural bias in their history, or the ways data are biased due to systemic inequities in the administration of policies and programs. Meanwhile, the public are rarely consulted in data use. Though data infrastructure can be a powerful tool to support equity-oriented reforms, equity is rarely a stated goal for data integration. This raises fundamental concerns, as integrated data increasingly provide the raw materials for evaluation, research, and risk modeling that inform policy, practice, and resource allocation.
Actionable Intelligence for Social Policy (AISP) is an initiative of the University of Pennsylvania that focuses on the development, use, and innovation of integrated data systems (IDS). We convene a network of IDS across the United States while supporting developing sites, and as such are uniquely situated to convene experts to develop guidance for centering equity within integrated data infrastructure.
This project aims to generate guidance for agencies supporting data sharing infrastructure to ensure an emphasis on equity and public engagement for ethical use.
A variety of data collection methods are being used, including expert panel convenings and interviews with sites piloting or exemplifying strategies for public engagement and equitable data access and use. An extensive literature review is also in progress and will inform a suite of forthcoming products, including a white paper, communications and training materials.
The results will provide strategies for centering equity across the spectrum of data integration activities, including inclusive governance, staffing considerations, decisions about data quality, and the ethical use of data models and algorithms. Initial findings indicate there are few exemplar sites that routinely center equity within data integration efforts, yet there are promising incremental steps that sites can take to ensure ethical use.
While centering equity within data integration is an emerging focus, initial findings indicate the importance of such efforts, particularly in acknowledging and mitigating the risks of unacknowledged bias across use of administrative data for research and evaluation purposes.