Title: Causality and covariate adjustment
is commonly used in practice to adjust for confounders. In this talk, we argue
that the covariate set should be chosen with care. We will discuss graphical
criteria that identify covariate sets that give consistent and efficient
estimates of total effects in causal linear models. We will illustrate the
concepts in examples and discuss various generalizations.
Marloes Maathuis is Professor of Statistics at the Seminar for Statistics at ETH Zurich. She studied Applied Mathematics at Delft University of Technology and obtained a PhD in Statistics from the University of Washington, Seattle. Her research interests include causality, graphical models, high-dimensional models, and applications of statistics. For more information, please see https://stat.ethz.ch/~maathuis/