49th Annual Summer Institute of Applied Statistics 2025
Making Causal Inferences Robust to Unobserved Confounding
June 17 - 18, 2025

The featured speaker will be Dr. Sam Pimentel, Assistant Professor for the Department of Statistics at the UC Berkeley.
Measuring the causal impact of an exposure is an important scientific goal in social science, public policy, and medicine. Since randomized experiments are often impractical, most attempts to interrogate causal effects use observational data in which the researcher did not assign the exposure to subjects. In this setting groups with different exposures may also differ systematically in baseline variables. While a rich literature addresses corrections for such confounding between groups due to observed variables, confounding due to unobserved variables remains a more difficult challenge.
This course provides an overview of methodology for learning about causal effects from observational data in the presence of unobserved confounding. After a review of the most common tools used for causal analyses in observational data — including matching, weighting, and doubly-robust methods — we will explore leading methods for sensitivity analysis, which quantify the degree of unobserved confounding needed to explain an apparent effect if none is truly present. In addition, we will discuss design sensitivity, which permits researchers to plan observational studies in a way that maximizes robustness to potential confounders, and devices to detect or combat the impact of unobserved confounders, including negative control outcomes and multiple control groups. We will also conduct illustrative analyses from the health and social sciences using R.
To view past presenters, click here.
To register, click here
For questions regarding SIAS, contact Kimri Mansfield at (801) 422-4506 or kmansfield@stat.byu.edu.