Controlling Selection Bias in Causal Inference

By Elias Bareinboim, Judea Pearl. 2011

218 citations recorded by the source corpus.

Selection bias, caused by preferential exclusion of units (or samples) from the data, is a major obstacle to valid causal inferences, for it cannot be removed or even detected by randomized experiments. This paper highlights several graphical and algebraic methods capable of mitigating and sometimes eliminating this bias. These nonparametric methods generalize and improve previously reported results, and identify the type of knowledge that need to be available for reasoning in the presence of selection bias

Subfields: Artificial Intelligence, Statistics and Probability

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

Citations in the corpus

Counterfactual Normalization: Proactively Addressing Dataset Shift and Improving Reliability Using Causal Mechanisms (2018, methodology): “Others have studied methods and conditions for identiﬁcation of causal effects under simultaneous selection and confounding bias (e.g., Bareinboim and Pearl (2012); Bareinboim and Tian (2015); Correa et al. (2018)).”
Recovering Causal Effects from Selection Bias (2015, methodology): “…expresses the requirement that effects are computable from the available (biased) data and assumption embodied in an augmented causal graph Gs (Bareinboim and Pearl 2012b): Definition 2 (Recoverability from Selection Bias) Given a causal graph Gs augmented with a node S encoding the selection…”
Recovering from Selection Bias in Causal and Statistical Inference (2014, background): “Using this language, (Bareinboim and Pearl 2012) provided a complete treatment for selection relative to the OR 4 .”
Causal Effect Identification by Adjustment under Confounding and Selection Biases (2017, background): “Formally, the task of estimating a probabilistic quantity from a selection-biased distribution is known as recovering from selection bias (Bareinboim and Pearl 2012).”
Causal inference and the data-fusion problem (2016, background): “We also should note that the odds ratio can be recovered from selection bias even in settings where the risk difference cannot (36, 37, 40).”
Causal Intervention for Weakly-Supervised Semantic Segmentation (2020, background): “The purpose of causal inference [39, 43] is to empower models the ability to pursue the causal effect: we can remove the spurious bias [5], disentangle the desired model effects [6], and modularize reusable features that generalize well [35].”
Transportability across studies: A formal approach (2011, methodology): “Elsewhere, we analyze S variables representing selection of units into the study pool [BP11]; there, the arrows will be pointing towards S.”
Linear Models: A Useful “Microscope” for Causal Analysis (2013, unknown): “[12] and [14] we obtain βyx 1⁄4 αþ βγ βyx z 1⁄4 α Thus, if α has a different sign from βγ, it is quite possible to have the regression of Y on X, βyx, change sign upon conditioning on Z 1⁄4 z; for every z.”
A General Algorithm for Deciding Transportability of Experimental Results (2013, background): “Bareinboim and Pearl [50] extended these conditions and provided a complete characterization, together with an algorithm, for deciding when a bias-free estimate of the odds ratio (OR) can be recovered from selection-biased data.”
Multiple Causes: A Causal Graphical View (2019, methodology): “Following Bareinboim & Pearl (2012), we introduce a selection operator S ∈ { 0 , 1 } into the causal graph.”
TRYGVE HAAVELMO AND THE EMERGENCE OF CAUSAL CALCULUS (2014, unknown): “A complete theory can be found in Bareinboim and Pearl (2012) and Bareinboim et al. (2014). 16. According to White and Lu (2010) “A common exercise in empirical studies is a ‘robustness check,’ where the researcher examines how certain ‘core’ regression coefficient estimates behave when the regressi”
Estimating individual treatment effect: generalization bounds and algorithms (2016, background): “…the causal inference community relates to bounding the estimate of the average treatment effect given an instrumental variable (Balke & Pearl, 1997; Bareinboim & Pearl, 2012), or under hidden confounding , for example when the ignorability assumption does not hold (Pearl, 2009; Cai et al., 2008).”
Adjustment Criteria for Generalizing Experimental Findings (2019, background): “S can be seen as an indicator where S=1 if a unit is included in the sample, and S=0 otherwise (Bareinboim and Pearl, 2012).”
Generalized Adjustment Under Confounding and Selection Biases (2018, background): “1(c), for example, S represents a binary indicator of entry into the data pool, such that S=1 if unit is included in the sample and S=0 otherwise (Bareinboim and Pearl 2012b).”