Department of Statistics Unitmark
Dietrich College of Humanities and Social Sciences


Census-taking around the world is under assault, due to concerns about privacy, costs, and diminishing response rates. Our node's proposal responds to these issues in the following ways:
• Privacy: While we have done and are doing extensive research on confidentiality and privacy in principle, this node focuses on the practical problems of insuring confidentiality and privacy in the Census process as it is likely to develop, while not losing sight of the purpose of the Census to produce useable data both for public and for private purposes.
• Costs: We are pursuing two possible routes to minimize costs of future censuses. We think that administrative records could be used effectively to create an initial census frame. There is no reason why the federal government needs to collect the same information repeatedly. A second idea is the possibility of an online census form as a full or partial replacement for the traditional mailing method of census-taking. This idea has to be studied for accuracy, for preservation of privacy and confidentiality, as well as for cost.
• Diminishing Response Rates: A more efficient census would allow resources to be concentrated on the most difficult-to-reach subpopulations. To some extent, the response rate problem is an intractable issue of social change, but we expect to learn new ways to encourage participation through experiments that study participatory response.

Our node consists of a multi-faceted methodological research program to assist the Census Bureau in increasing census accuracy for 2020, while at the same time decreasing total census cost. In addition, we plan to use our program to engage colleagues and students, both at Carnegie Mellon University and more broadly, in thinking about research issues of direct relevance to the decennial census and other Census Bureau programs.