Doubly Robust Estimation in Missing Data and Causal Inference Models

By H. Bang, James M. Robins. 2005

2,329 citations recorded by the source corpus.

Subfields: 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 Statistics and Probability

In Causal Inference & Discovery

Citations in the corpus

Counterfactual reasoning and learning systems: the example of computational advertising (2012, background): “Doubly robust variance reduction has also been extensively used for causal inference applied to biostatistics (see Robins et al., 2000; Bang and Robins, 2005).”
More Robust Doubly Robust Off-policy Evaluation (2018, methodology): “The DR estimator was first developed in statistics (e.g., Cassel et al. 1976; Robins et al. 1994; Robins & Rotnitzky 1995; Bang & Robins 2005) to estimate from incomplete data with the ar X iv :1 80 2.”
Optimal and Adaptive Off-policy Evaluation in Contextual Bandits (2016, methodology): “In statistics, this is studied as the problem of mean estimation for treatment regimes in the causal inference literatre (see e.g. [Horvitz and Thompson, 1952, Holland, 1986, Bang and Robins, 2005, Rotnitzky et al., 2012] as well as the references in [Dudík et al., 2014]).”
P3O: Policy-on Policy-off Policy Optimization (2019, methodology): “These ideas have also been employed in reinforcement learning (Kang et al., 2007; Bang and Robins, 2005; Dudı́k et al., 2011).”
Learning Representations for Counterfactual Inference (2016, methodology): “Doubly robust methods combine aspects of parametric and non-parametric methods, typically by using a propensity score weighted regression (Bang & Robins, 2005; Dud´ık et al., 2011).”
Comment: Demystifying Double Robustness: A Comparison of Alternative Strategies for Estimating a Population Mean from Incomplete Data. (2008, methodology): “Further explanation and evaluation of DR estimators has been given by Lunceford and Davidian (2004), Carpenter, Kenward and Vansteelandt (2006), Davidian, Tsiatis and Leon (2005) and Bang and Robins (2005).”
Demystifying Double Robustness: A Comparison of Alternative Strategies for Estimating a Population Mean from Incomplete Data (2007, background): “However, in the process of arguing that we misunderstood the message of Bang and Robins (2005), they have apparently misunderstood ours.”
The Selective Labels Problem: Evaluating Algorithmic Predictions in the Presence of Unobservables (2017, methodology): “(3) doubly robust methods which combine the two aforementioned classes of techniques typically via a propensity score weighted regression [5, 10].”
Automated versus Do-It-Yourself Methods for Causal Inference: Lessons Learned from a Data Analysis Competition (2017, background): “A strong focus of the causal inference literature over the past few decades has been on preprocessing data to reduce reliance on parametric assumptions (Scharfstein, Rotnitzky and Robins, 1999; Bang and Robins, 2005; Sekhon, 2007).”
Asymptotic inference of causal effects with observational studies trimmed by the estimated propensity scores (2018, neutral): “…X ) = E(Y | A = a, X ) by μ̂(a, X ) and obtain the residual R̂i = Yi − μ̂(Ai, Xi), then the augmented weighting estimator is N −1 ∑N i=1 τ̂ aug(Xi) (Lunceford & Davidian, 2004; Bang & Robins, 2005), where τ̂ aug(Xi) = { AiR̂i e(X Ti θ̂ ) + μ̂(1, Xi) } − { (1 − Ai)R̂i 1 − e(X Ti θ̂ ) + μ̂(0, Xi) } .”
When the money runs out: Do cash transfers have sustained effects on human capital accumulation? (2016, neutral): “We can also use OLS to control for the same set of maternal baseline characteristics iX to provide estimates that are “doubly robust” to the extensive margin selection controls (Robins and Rotnitzky 2001; Van der Laan and Robins 2003; Bang and Robins 2005).”
Does more balanced survey response imply less non‐response bias? (2016, methodology): “Bang and Robins (2005)). For an excellent discussion on the various estimators see Schafer and Kang (2008). We restrict ourselves to the estimation of the location of distributions of interest.”
Evaluating online ad campaigns in a pipeline: causal models at scale (2010, methodology): “Doubly robust estimation is used in applications ranging from medicine [2] to criminal justice [15].”