Partial Identification with Noisy Covariates: A Robust Optimization Approach
This addresses the issue of unreliable causal inference for researchers and practitioners when covariate measurements are imperfect, offering a method to extend various adjustment techniques, though it is incremental in building on existing partial identification frameworks.
The paper tackles the problem of biased causal estimates from noisy covariates in observational studies by proposing a robust optimization approach for partial identification of average treatment effects, showing it provides bounds with higher coverage probability than existing methods across synthetic and real datasets.
Causal inference from observational datasets often relies on measuring and adjusting for covariates. In practice, measurements of the covariates can often be noisy and/or biased, or only measurements of their proxies may be available. Directly adjusting for these imperfect measurements of the covariates can lead to biased causal estimates. Moreover, without additional assumptions, the causal effects are not point-identifiable due to the noise in these measurements. To this end, we study the partial identification of causal effects given noisy covariates, under a user-specified assumption on the noise level. The key observation is that we can formulate the identification of the average treatment effects (ATE) as a robust optimization problem. This formulation leads to an efficient robust optimization algorithm that bounds the ATE with noisy covariates. We show that this robust optimization approach can extend a wide range of causal adjustment methods to perform partial identification, including backdoor adjustment, inverse propensity score weighting, double machine learning, and front door adjustment. Across synthetic and real datasets, we find that this approach provides ATE bounds with a higher coverage probability than existing methods.