CAREER: Accountable Democracy: Mathematical Reasoning and Democratic Processes in America

In recent years, it has become increasingly apparent that computational techniques are deeply embedded in every stage of the American democratic process. Prominent examples include computational redistricting and the increased use of statistical analysis in polling and forecasts. This project will historicize the role of computers, as well as algorithmic thinking and mathematical rationales, in the constitution of American representative democracy in the twentieth century. Even before digital computers and automated decision making became prevalent, controversies about mathematical definitions of 'fairness,' 'bias,' and 'discrimination' abounded in American life, and nowhere were these ideas debated more forcefully than in the realm of democratic decision making. Focusing on congressional apportionment, the census, proportional voting, congressional redistricting, polling, and forecasts, this project investigates how changing computational practices, from statistical modeling to computational geometry, have insinuated themselves into the most basic definitions of 'fairness' in the American electorate in the twentieth century.

Applying theoretical insights from Science and Technology Studies, this project will integrate archival research and oral histories in order to investigate the extent to which ideas about what constitutes 'bias,' 'fairness,' and 'discrimination' are determined by mathematical and computational metrics as well as by legal, legislative, and social ones. This research will therefore elucidate how technical and social definitions of fairness are entangled in computational and mathematical ways of thinking. Furthermore, it will reckon with the public understanding of science and, in particular, the public understanding of mathematics and statistics. Finally, it will extend the study of how scientific expertise is received and adjudicated when the expert's skill is a method (e.g., statistics), rather than a subject matter (e.g., climate change, DNA fingerprinting). The project will increase public engagement with science and technology by unpacking and clearly outlining the ways in which mathematical and computational claims to knowledge have at times obscured political and social agendas. One of the primary goals of the project is to understand how mathematical claims to truth are adjudicated not only in the courts, but also by the general public. Some of the debates that will be examined, like the one regarding census adjustment, are highly technical in nature. Examining the successful (as well as unsuccessful) strategies that U.S. journalists and the public have developed to communicate and disseminate these complex issues is crucial to contemporary society, when so many aspects of daily life are governed by technically intricate, if not entirely opaque, computational systems. As data analytics become increasingly foundational to how Americans engage with the most basic constitution of democracy, it is of the utmost importance to publicly disseminate information and toolkits to help citizens evaluate such claims and their assumptions.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.