Adjusting for Nonignorable Drop-Out Using Semiparametric Nonresponse Models

By D. Scharfstein, Andrea Rotnitzky, James M. Robins. 1999

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

Citations in the corpus

Estimation Considerations in Contextual Bandits (2017, background): “The term “doubly-robust” comes from the extensive literature on offline policy evaluation [54]; it means that when comparing two policies using historical data, we get consistent estimates of the average difference in outcomes”
Balanced Policy Evaluation and Learning (2017, methodology): “[11] propose a weighting-based approach that combines the direct and IPW approaches by adapting the doubly robust (DR) estimator [9, 27, 28, 31]:”
Approximate residual balancing: debiased inference of average treatment effects in high dimensions (2016, methodology): “Xi/} as sample weights when estimating β̂c, whereas Scharfstein et al. (1999) and Van Der Laan and Rubin (2006) considered adding these weights as features in the outcome model; see also Robins et al. (2007) and Tan (2010) for further discussion.”
Efficiently Breaking the Curse of Horizon in Off-Policy Evaluation with Double Reinforcement Learning (2019, methodology): “The third approach is the doubly robust (DR) method, which combines DM and IS and is given by adding the estimated q -function as a control variate (Dudik et al., 2014; Jiang and Li, 2016; Scharfstein et al., 1999).”
Learning When-to-Treat Policies (2019, methodology): “We propose an advantage doubly robust estimator that draws upon the semiparametrics and orthogonal moments literature (Belloni et al., 2014; Chernozhukov et al., 2016a; Newey, 1994; Robins and Rotnitzky, 1995; Scharfstein et al., 1999).”
Quasi-oracle estimation of heterogeneous treatment effects (2017, methodology): “…inference and constructions of orthogonal moments (Athey and Wager, 2017; Belloni et al., 2017; Bickel et al., 1998; Luedtke and van der Laan, 2016; Newey, 1994; Robins, 2004; Robins and Rotnitzky, 1995; Robins et al., 2017; Scharfstein et al., 1999; Tsiatis, 2007; van der Laan and Rose, 2011).”
The Elephant in the Corner: A Cautionary Tale About Measurement Error in Treatment Effects Models (2010, methodology): “The DR estimator in Scharfstein et al. (1999) and Bang and Robins (2005) entails OLS estimation of the following speci cation Yi = 0 + ( 1 0)Ti +Xi + 0(1 Ti)f1 bP (Xi)g 1 + 1Tif bP (Xi)g 1 + i: The DR estimate of the ATE is given by ̂OLS = \ 1 0+(1=N) P i[ b 1f bP (Xi)g 1 b 0f1 bP (Xi)g 1].”
A Semiparametric Estimation of Mean Functionals With Nonignorable Missing Data (2011, methodology): “The exponential tilting model for nonignorable nonresponse is similar in spirit to the stratified Cox proportional hazards model considered in Scharfstein, Rotnitzky, and Robins (1999), which was used to model nonignorable drop out in the analysis of longitudinal data.”
An Introduction to the Augmented Inverse Propensity Weighted Estimator (2010, background): “Most interestingly, ÂTEAIPW is doubly robust in that it will be consistent for ATE whenever (1) the propensity score model is correctly specified or (2) the two outcome regression models are correctly specified (Scharfstein et al., 1999).”
Cross-fitting and fast remainder rates for semiparametric estimation (2017, methodology): “The double robustness of the inﬂuence function for the functionals we consider is shown in Chernozhukov et al.(2016), where the doubly robust estimators of Scharfstein, Rotnitzky, and Robins (1999), Robins, Rotnitzky, and van der Laan (2000), Robins et. al.”