Causal inference and machine learning

This course presents the theory and practice of contemporary causal inference. It is strongly interdisciplinary and focused on the practical applications of theory.

Significant advances in causal inference have come from a wide variety of disciplines, including philosophy, economics, statistics, computer science and epidemiology. The course will introduce students to methods from all of these disciplines and discuss whether and how they can illuminate questions of public policy. In keeping with its contemporary focus, the course will also emphasise the connections, tensions and opportunities that arise from addressing causal questions with machine learning tools.

While the course will not shy away from theoretical questions, it will pay particular attention to how theoretical tools help with practical issues in experimental, quasi-experimental and observational data analysis in a public policy context. We will consider what theory and tools for causal inference can add to a wide range of policy debates, from traditional policy impact assessment, through research design criticism and correction to recent conceptual frameworks of fairness as they illuminate normative policy dilemmas. Because nothing is more practical than a good theory.
 

Instructor