Day 1 | Tutorial: Causal Inference | Virtual Option

Speaker:

• Jose Zubizarreta (Associate Professor, Department of Health Care Policy, Harvard Medical School; Associate Professor, Department of Biostatistics, Harvard School of Public Health; Faculty Affiliate, Department of Statistics, Harvard University)

Abstract:

In this tutorial, we will provide an introduction to causal inference. We will describe ideal study design principles to establish cause and effect relationships. We will explain the potential outcomes framework for causal inference, the central role of randomization for identification and inference in randomized experiments, and then focus on observational studies distinguishing between study design strategies that aim to control for observed and for unobserved confounding. With both strategies, we will consider machine learning prediction approaches and emphasize the use of modern matching and weighting methods for transparent, sample-bounded estimation of causal effects. As time permits, we will also discuss sensitivity analyses to hidden biases and provide some practical considerations. Attendees should have familiarity with regression approaches.