A general condition for bias attenuation by a nondifferentially mismeasured confounder

TL;DR

This paper establishes a general condition under which adjusting for a nondifferentially mismeasured confounder reduces bias in causal effect estimation, even when measurement error is present.

Contribution

It provides a new theoretical framework showing that bias attenuation can occur under broader conditions than previously known, especially with nondifferential measurement error.

Findings

Adjusting for mismeasured confounders often reduces bias.

The paper offers a general condition for bias attenuation.

Empirical scenarios demonstrate the practical relevance.

Abstract

In real-world studies, the collected confounders may suffer from measurement error. Although mismeasurement of confounders is typically unintentional -- originating from sources such as human oversight or imprecise machinery -- deliberate mismeasurement also occurs and is becoming increasingly more common. For example, in the 2020 U.S. Census, noise was added to measurements to assuage privacy concerns. Sensitive variables such as income or age are oftentimes partially censored and are only known up to a range of values. In such settings, obtaining valid estimates of the causal effect of a binary treatment can be impossible, as mismeasurement of confounders constitutes a violation of the no unmeasured confounding assumption. A natural question is whether the common practice of simply adjusting for the mismeasured confounder is justifiable. In this article, we answer this question in the affirmative and demonstrate that in many realistic scenarios not covered by previous literature, adjusting for the mismeasured confounders reduces bias compared to not adjusting.