Identifying Causal Effects With Proxy Variables of an Unmeasured Confounder

By Wang Miao, Zhi Geng, E. T. Tchetgen Tchetgen. 2016

385 citations recorded by the source corpus.

We consider a causal effect that is confounded by an unobserved variable, but with observed proxy variables of the confounder. We show that, with at least two independent proxy variables satisfying a certain rank condition, the causal effect is nonparametrically identified, even if the measurement error mechanism, i.e., the conditional distribution of the proxies given the confounder, may not be identified. Our result generalizes the identification strategy of Kuroki & Pearl (2014) that rests on identification of the measurement error mechanism. When only one proxy for the confounder is available, or the required rank condition is not met, we develop a strategy to test the null hypothesis of no causal effect.

Subfields: Artificial Intelligence, Statistics and Probability, Causal Inference & Discovery

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

In Causal Inference & Discovery

Citations in the corpus

Deconfounding Reinforcement Learning in Observational Settings (2018, methodology): “However, using proxy variables to correctly recover causal effects requires strict mathematical assumptions (Louizos et al., 2017; Edwards et al., 2015; Kuroki and Pearl, 2014; Miao et al., 2018; Pearl, 2012; Wooldridge, 2009) and, in practice, we do not know whether or not the available data meets ”
Multiple Causes: A Causal Graphical View (2019, methodology): “Completeness is a common assumption posited in nonparametric causal identiﬁcation (Miao et al., 2018; Yang et al., 2017; D’Haultfoeuille, 2011); it is often used to guarantee the existence and the uniqueness of solutions to integral equations.”
Multi-cause causal inference with unmeasured confounding and binary outcome. (2019, background): “, 2016), negative controls (e.g., Kuroki and Pearl, 2014; Miao et al., 2018) and sensitivity analysis (e.”
Policy Evaluation with Latent Confounders via Optimal Balance (2019, methodology): “In particular, there is a variety of work that studies sufficient conditions for the identifiability of latent confounder models [6, 34, 37].”
Ethical Machine Learning in Healthcare (2020, methodology): “In practice, when potential hidden confounding is suspected, either mediating features or proxies can be leveraged (103, 99) or sensitivity analysis methods can be used to determine potential sources of errors in effect estimates (104).”
Long-term Causal Inference Under Persistent Confounding via Data Combination (2022, background): “The seminal work of Miao et al. [2016] first prove the identification of treatment effects with two different types of proxy variables for unmeasured confounders: one type of proxy variables (called negative control outcomes) are not affected by the treatment, and other type of proxy variables (call”
De-Confounded Variational Encoder-Decoder for Logical Table-to-Text Generation (2021, background): “Proxy variables are the proxies of hidden or unmeasured variables (Miao et al., 2018).”
A confounding bridge approach for double negative control inference on causal effects (2018, positive): “Confounder proxies used by Miao, Geng, et al. (2018) and Kuroki and Pearl (2014) can be viewed as special negative controls in our framework.”