Causal Inference through a Witness Protection Program

Part of Advances in Neural Information Processing Systems 27 (NIPS 2014)

Bibtex »Metadata »Paper »Reviews »Supplemental »

Authors

Ricardo Silva, Robin Evans

Abstract

<p>One of the most fundamental problems in causal inference is the estimation of a causal effect when variables are confounded. This is difficult in an observational study because one has no direct evidence that all confounders have been adjusted for. We introduce a novel approach for estimating causal effects that exploits observational conditional independencies to suggest <code>weak'' paths in a unknown causal graph. The widely used faithfulness condition of Spirtes et al. is relaxed to allow for varying degrees of</code>path cancellations'' that will imply conditional independencies but do not rule out the existence of confounding causal paths. The outcome is a posterior distribution over bounds on the average causal effect via a linear programming approach and Bayesian inference. We claim this approach should be used in regular practice to complement other default tools in observational studies.</p>