Classical approaches to causal inference largely rely on the assumption of “lack of interference”, according to which the outcome of an individual does not depend on the treatment assigned to others. In many applications, however, such as designing and evaluating the effectiveness of healthcare interventions that leverage social structure, assuming lack of interference is untenable. In fact, the effect of interference itself is often an inferential target of interest. In this talk, we will discuss technical issues that arise in estimating causal effects when interference can be attributed to a network among the units of analysis, and develop a strategy for optimal experimental design in this context that involves a piecewise constant approximation of a certain graphon.

Bio: Edo Airoldi is an associate professor of statistics at Harvard University. His research interests include statistical methodology and theory for network data analysis, stochastic optimization strategies for inference with large data sets, and modeling and inference in molecular biology. He is the recipient of a Sloan Research Fellowship and a Shutzer Fellowship, has received several best paper awards, and was recently a speaker at the Sackler Colloquium on “Drawing Causal Inference from Big Data” at the National Academy of Sciences.