Causal diagrams for empirical research

By Judea Pearl. 1995

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

Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias (2020, neutral): “, variable selection methods (Pearl, 1995).”
Learning Neural Causal Models from Unknown Interventions (2019, background): “The recovery of the underlying structural causal graph from observational and interventional data is a fundamental problem (Pearl, 1995; 2009).”
A Meta-Transfer Objective for Learning to Disentangle Causal Mechanisms (2019, methodology): “Pearl’s seminal work on do-calculus Pearl (1995, 2009); Bareinboim and Pearl (2016) lays a foundation for expressing the impact of interventions on probabilistic graphical models – we use it in our work.”
Causal inference in statistics: An overview (2009, methodology): “Pearl (1995a) has presented examples in which there exists no set of variables that is sufficient for adjustment and where the causal effect can nevertheless be estimated consistently.”
Characterization and greedy learning of interventional Markov equivalence classes of directed acyclic graphs (2011, background): “Extending the do() operator (Pearl, 1995) to stochastic interventions, we denote the density of X under such an intervention by f ( x | do D ( X I = U I )).”
Theoretical Impediments to Machine Learning With Seven Sparks from the Causal Revolution (2018, background): “ is available, called do-calculus, which predicts the eﬀect of policy interventions whenever feasible, and exits with failure whenever predictions cannot be ascertained with the speciﬁed assumptions (Pearl, 1995; Tian and Pearl, 2002; Shpitser and Pearl, 2008). Pillar 3: The Algorithmization of Coun”
Interventional Few-Shot Learning (2020, methodology): “2 Derivation of Backdoor Adjustment for the Proposed Causal Graph We will show the derivation of the backdoor adjustment for the causal graph in Figure 3(a) using the three rules of do-calculus [47].”
Invariant Risk Minimization Games (2020, background): “(2019) is the most related work, and is motivated from the theory of causality and causal Bayesian networks (CBNs) (Pearl, 1995).”
Informativeness and Invariance: Two Perspectives on Spurious Correlations in Natural Language (2022, methodology): “…data generating process, we can compute an interventional distribution P ( Y | do ( X 1 := x ) , X 2 , X 3 ) , which corresponds to the distribution over Y in a data generating process in which the variable X 1 is surgically set to the value x (Pearl, 1995; Peters et al., 2017; Feder et al., 2021).”
Recovering Causal Effects from Selection Bias (2015, methodology): “The problem of confounding has been broadly studied in the literature and a number of conditions for non-parametric identification had emerged, for instance (Spirtes, Glymour, and Scheines 1993; Galles and Pearl 1995; Pearl and Robins 1995; Halpern 1998; Kuroki and Miyakawa 1999).”
Pathways between education and health: a causal modelling approach (2006, methodology): “SEMs can be used to specify ‘causal models’ through which causal hypotheses about the relationships between variables can be constructed and tested (Pearl, 1995; Greenland, 2000; Sobel, 2000; Greenland and Brumback, 2002).”
Signed directed acyclic graphs for causal inference (2010, methodology): “A backdoor path from some node A to another node Y is a path which begins with a directed edge into A. Pearl (1995) showed that, for intervention variable A and outcome Y , if a set of variables Z is such that no variable in Z is a descendant of A and such that Z blocks all backdoor paths from A to…”
An Extended Class of Instrumental Variables for the Estimation of Causal Effects (2011, methodology): “Developments along these lines include “selection on observables” (Barnow, Cain, and Goldberger, 1980; Heckman and Robb, 1985), ignorability and the “propensity score” (Rubin, 1974; Rosenbaum and Rubin, 1983), the “back-door” method (Pearl, 1995), and “predictive proxies” (White, 2006; WC).”
Direct and Indirect Causal Effects via Potential Outcomes * (2004, background): “Nevertheless, I still feel as I did in my discussion with Imbens (Imbens & Rubin, 1995) of Pearl (1995) , that the framework is inherently less revealing than the potential outcomes framework because it tends to bury essential scientific and design issues.”
Causal Network Learning from Multiple Interventions of Unknown Manipulated Targets (2016, methodology): “If a causal network model (G,P ) satisfies the causal Markov condition, then the joint distribution of p variables X can be factored as follows (Pearl, 1995; Spirtes et al., 2001) P (X) = p∏ i=1 P (xi|pai), (1) where P (xi|pai) is the conditional probability of xi given its parent set pai.”
Identification of Joint Interventional Distributions in Recursive Semi-Markovian Causal Models (2006, methodology): “(cid:2) | Results in [Pearl, 1995], [Halpern, 2000] take a different approach, and provide sound rules which are used to manipulate the expression corresponding to the effect alge-braically.”