What does it mean for an algorithm or decision-making process to be fair? When making decisions using data, how do we account for historical biases or systemic inequalities? Starting with affirmative action in the United States, this course explores the methodological, ethical, and legal ramifications of quantitative decision-making.  In the context of affirmative action, this course will review the empirical literature and evolving legal doctrine that reflect how courts and policymakers in the United States have struggled with these questions. This course then reviews the literature on algorithmic fairness, examining definitions of fairness and efforts to create fair machine learning models. In particular, this course will explore what causal inference frameworks can offer in addressing these issues. This is an interdisciplinary course, with readings that span the fields of statistics, computer science, economics, and law. This is also an applied course that will connect abstract concepts of fairness to issues in American criminal justice, higher education, and the labor market.

No textbook required. A syllabus with the readings will be distributed.