Seminar by Jared Huling, University of Minnesota

Abstract

Studying causal effects of continuous treatments is important for gaining a deeper understanding of many interventions, policies, or medications, yet researchers are often left with observational studies for doing so. In the observational setting, confounding is a barrier to estimation of causal effects. Weighting approaches seek to control for confounding by reweighting samples so that confounders are comparable across different values of the treatment. Yet, for continuous treatments, weighting methods are highly sensitive to model misspecification. In this work we elucidate the key property that makes weights effective in estimating causal quantities involving continuous treatments. We show that to eliminate confounding, weights should make treatment and confounders independent on the weighted scale. We develop a measure that characterizes the degree to which a set of weights induces such independence and propose a new model-free method for weight estimation by optimizing our measure. The empirical effectiveness of our approach is demonstrated in a suite of challenging numerical experiments, where we find that our weights are quite robust and work well under a broad range of settings. We briefly touch on ongoing work and open challenges in the area.

Bio

Jared Huling develops statistical methods for the analysis of complex and high-dimensional observational studies. His work focuses on the development of statistical approaches to better understand causal relationships from observational data, to optimally match individual patients with the right treatments, and statistical learning approaches for accurate and interpretable risk models using high-dimensional, heterogeneous data.