The central role of the propensity score in observational studies for causal effects

By Paul R. Rosenbaum, Donald B. Rubin. 1983

Subfields: Probabilistic & Bayesian Methods

How researchers see this paper

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

In Probabilistic & Bayesian Methods

Citations in the corpus

Improving a Neural Semantic Parser by Counterfactual Learning from Human Bandit Feedback (2018, background): “…(DPM) objec-tive (Lawrence et al., 2017b), without the possibility to correct the sampling bias of the logging policy by inverse propensity scoring (Rosenbaum and Rubin, 1983): This objective can show degenerate behavior in that it overﬁts to the choices of the logging policy (Swaminathan and…”
Avoiding Discrimination through Causal Reasoning (2017, methodology): “From the perspective of causality, the idea of a similarity measure is akin to the method of matching in counterfactual reasoning [13, 14].”
A survey on measuring indirect discrimination in machine learning (2015, methodology): “s also use clustering on X to take into account more than one explanatory variable at the same time. Then z denotes a cluster, one strata is one cluster. 4.3.2. Propensity measure. Propensity models [Rosenbaum and Rubin 1983] are typically used in clinical trials or marketing for estimating the prob”
Deep-Treat: Learning Optimal Personalized Treatments From Observational Data Using Neural Networks (2018, background): “need to correct the bias of the representations by weighting the instances by their inverse propensities (Rosenbaum and Rubin 1983) since not all bias will be removed during the auto-encoder stage.”
Detecting and Correcting for Label Shift with Black Box Predictors (2018, background): “Also related, Rosenbaum & Rubin (1983) introduce propensity scoring to design unbiased experiments.”
FLAME: A Fast Large-scale Almost Matching Exactly Approach to Causal Inference (2017, methodology): “In particular, a randomization scheme leads to the strong ignorability of treatment (Rosenbaum and Rubin, 1983): each individual has a positive probability of being assigned to treatment, and there is conditional independence of potential outcomes ( Y ( 0 ) , Y ( 1 )) from T given relevant…”
Metalearners for estimating heterogeneous treatment effects using machine learning (2017, background): “To aid our ability to estimate τ , we need to assume that there are no hidden confounders (13): Condition 1: (ε(0), ε(1))⊥W |X .”
Double/Debiased/Neyman Machine Learning of Treatment Effects (2017, methodology): “We consider estimation of ATE and ATTE under the unconfoundedness assumption of Rosenbaum and Rubin (1983).”
Generalized random forests (2016, positive): “If Wi is a binary treatment, this condition is equivalent to the unconfoundedness assumption used to motivate propensity score methods (Rosenbaum and Rubin, 1983).”
Approximate residual balancing: debiased inference of average treatment effects in high dimensions (2016, background): “…matching, propensity score weighting and subclassification, as well as doubly robust combinations thereof (e.g. Abadie and Imbens (2006), Heckman et al. (1998), Hirano et al. (2003), Robins et al. (1994, 1995, 2017), Rosenbaum (2002), Tan (2010), Tsiatis (2007) and Van Der Laan and Rubin (2006)).”
Estimation and Inference of Heterogeneous Treatment Effects using Random Forests (2015, methodology): “In the widely used potential outcomes framework we use to formalize our results [Holland, 1986, Imbens and Rubin, 2015, Rosenbaum and Rubin, 1983, Rubin, 1974], a treatment eﬀect is understood as a diﬀerence between two potential outcomes: e.g., would the patient have died is they received the drug…”
Deep Counterfactual Networks with Propensity-Dropout (2017, methodology): “Throughout this paper, we adopt Rubin’s potential outcomes model (Rubin, 2011; 1973; Rosenbaum & Rubin, 1983).”
Causal inference in statistics: An overview (2009, methodology): “The selection criterion can be applied systematically to diagrams of any size and shape, thus freeing analysts from judging whether “X is conditionally ignorable given S,” a formidable mental task required in the potential-response framework (Rosenbaum and Rubin, 1983).”
Comment: Demystifying Double Robustness: A Comparison of Alternative Strategies for Estimating a Population Mean from Incomplete Data. (2008, background): “This probability is called the propensity score (Rosenbaum and Rubin, 1983).”