Detecting and Measuring Confounding Using Causal Mechanism Shifts
This work addresses a key problem in causal inference for researchers and practitioners by providing methods to handle unobserved confounding, though it appears incremental as it builds on recent advancements in causal discovery.
The paper tackles the challenge of detecting and measuring confounding effects in causal inference by relaxing unrealistic assumptions like causal sufficiency and strong parametric models, proposing a comprehensive approach that includes tailored methodologies for various definitions of confounding, with empirical results supporting the theoretical analysis.
Detecting and measuring confounding effects from data is a key challenge in causal inference. Existing methods frequently assume causal sufficiency, disregarding the presence of unobserved confounding variables. Causal sufficiency is both unrealistic and empirically untestable. Additionally, existing methods make strong parametric assumptions about the underlying causal generative process to guarantee the identifiability of confounding variables. Relaxing the causal sufficiency and parametric assumptions and leveraging recent advancements in causal discovery and confounding analysis with non-i.i.d. data, we propose a comprehensive approach for detecting and measuring confounding. We consider various definitions of confounding and introduce tailored methodologies to achieve three objectives: (i) detecting and measuring confounding among a set of variables, (ii) separating observed and unobserved confounding effects, and (iii) understanding the relative strengths of confounding bias between different sets of variables. We present useful properties of a confounding measure and present measures that satisfy those properties. Empirical results support the theoretical analysis.