Workshop
Supercharge your causal inference with Machine Learning
Wednesday 23 October 10.30
Organizer: Alexander O.K. Marin, University of Southern Denmark
This session explores how machine learning is revolutionizing policy recommendations by advancing causal inference methods. Traditionally, scholars rely on causal evidence over correlation for policy debates. However, high costs or ethical concerns of randomized controlled trials often limit their use.
This has spurred the development of causal inference techniques using observational or quasi-experimental data. Machine learning has transformed these methods, enhancing our ability to detect heterogeneous causal effects with precision. Targeting subpopulations who benefit from interventions while avoiding potential harm is now more feasible, opening new avenues for policy design.
This session blends presentations with brief hands-on activities, guiding participants to apply causal inference in their work. After introducing classical methods, we will focus on estimating heterogeneous causal effects using machine learning and present their application to real-world academic research.
Introduction to the session and overview of the topic (5 min – Alexander)
Estimating Causal Effects (50 min – Phillip & Torben)
Heterogeneous Causal Effects in Academic Research – The Causal Effect of Community Nurseries on Child Health
Causal Inference Resources and Closing Remarks (5 min – Alexander)
Level: Introductory