Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction

By Guido W. Imbens, Donald B. Rubin. 2015

793 citations recorded by the source corpus.

Subfields: Management Science and Operations Research, 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 Management Science and Operations Research

In Statistics and Probability

In Causal Inference & Discovery

Citations in the corpus

Recommendations as Treatments: Debiasing Learning and Evaluation (2016, neutral): “This process is typically called the Assignment Mechanism in causal inference (Imbens & Rubin, 2015) or the Missing Data Mechanism in missing data analysis (Little & Rubin, 2002).”
Counterfactuals uncover the modular structure of deep generative models (2018, background): “Causality frameworks (using structural equations or potential outcomes ) allow evaluating with counterfactuals how the outcome of an observed system would have changed, had some variables taken different values [Pearl, 2009, Imbens and Rubin, 2015].”
Learning stable and predictive structures in kinetic systems (2018, background): “Causal models (22, 23) explicitly model such changes by the concept of interventions.”
Evaluating Reinforcement Learning Algorithms in Observational Health Settings (2018, methodology): “While formal statistical methods have been developed to assess the quality of new policies based on observational data alone [Thomas and Brunskill, 2016, Precup et al., 2000, Pearl, 2009, Imbens and Rubin, 2015], these methods rely on strong assumptions and are limited by statistical properties.”
Invariant Causal Prediction for Sequential Data (2017, neutral): “ a core problem in many scientiﬁc ﬁelds. Performing controlled randomized intervention experiments can be considered the gold standard for inferring causal relations (e.g., Peirce, 1883; Pearl, 2009; Imbens and Rubin, 2015; Peters et al., 2017). In many situations however, randomization and interven”
Local Linear Forests (2018, positive): “As in Athey, Tibshirani, and Wager (2019), we frame our discussion in terms of the Neyman–Rubin causal model (Imbens and Rubin 2015).”
Double/De-Biased Machine Learning of Global and Local Parameters Using Regularized Riesz Representers (2018, background): “We refer to Imbens and Rubin (2015), Hernan and Robins (2019), and Peters et al. (2017) for the relevant formalizations that enable causal interpretation.”
The Econometrics of Randomized Experiments (2016, neutral): “In an intention-to-treat analysis the receipt of treatment is ignored, and outcomes are compared by the assignment to treatment (Imbens and Rubin, 2015; Fisher et al, 2000).”
Learning When-to-Treat Policies (2019, methodology): “In contrast, we work in a nonparametric setting, and adopt a potential outcomes model in which we only get to observe outcomes corresponding to the sequence of actions we choose to take (Robins 1986; Imbens and Rubin 2015).”
Double/Debiased Machine Learning for Treatment and Causal Parameters (2016, positive): “robust" combinations of these methods; see, for example, Robins and Rotnitzky (1995), Hahn (1998), Hirano et al. (2003), and Abadie and Imbens (2006) as well as the textbook overview provided in Imbens and Rubin (2015). In this section, we present results that complement this important classic ”
Quasi-oracle estimation of heterogeneous treatment effects (2017, neutral): “Ding, Feller, and Miratrix (2019) consider heterogeneous treatment effect estimation in a strict randomization inference setting, where we the features and potential outcomes {Xi, Yi(0), Yi(1)}ni=1 are taken as fixed and only the treatment Wi is random (Imbens and Rubin, 2015); they then show how to”
Interpretable Almost Matching Exactly With Instrumental Variables (2019, methodology): “…a variable to be a valid instrument it must satisfy the following standard assumptions (see, e.g., Imbens and Angrist, 1994; Angrist et al., 1996; Imbens and Rubin, 2015): (A1) Relevance: 1n ∑n i=1 ti(1) − ti(0) 6= 0, that is, the variable does indeed have a non-zero causal effect on treatment…”
Causal inference with confounders missing not at random (2017, neutral): “If all the confounders of the treatment-outcome relationship are observed, one can use standard techniques, such as propensity score matching, subclassiﬁcation and weighting, to adjust for confounding (e.g., Rosenbaum & Rubin, 1983; Imbens & Rubin, 2015).”