Position title: Associate Professor, Statistics
1220 Medical Sciences Center
1300 University Avenue
- Department of Statistics
My training and expertise is in the area of causal inference, instrumental variables (IV)/Mendelian randomization (MR), and econometrics. Throughout the last six years, I have established the theoretical limit of inferring causal effects when the IV/MR assumptions known as the exclusion restriction (i.e. no direct effect assumption) is violated and proposed two statistical methods that address the problem. I have also developed a new, model-free, statistical method to infer causal effects using instrumental variables based on matching methods common in causal inference. These methods have been utilized in my collaborations with epidemiologists and clinicians using matching methods to study the causal effect of malaria on stunted growth in children in Ghana, Africa (using the sickle cell gene as the instrument). I continue to develop new statistical methods for causal inference, with a focus on building robust methods that allow relaxation of assumptions, such as modeling assumptions or causal assumptions required for identification of causal effects.
CDE research theme area affiliations
Biodemography; Health and the Life Course
Kang, Hyunseung, Laura Peck, and Luke Keele. “Inference for Instrumental Variables: A Randomization Inference Approach.” Journal of the Royal Statistical Society: Series A (Statistics in Society) (2018): online first.