Direct and Indirect Effects
2013/01/10 by Judea Pearl, Pearl, Judea · 107 citations
Computer Science · Decision Sciences · Mathematics · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Methodology (stat.ME) #cs.AI #stat.ME
paper · pdf · doi:10.48550/arxiv.1301.2300
Appears in Proceedings of the Seventeenth Conference on Uncertainty in Artificial Intelligence (UAI2001)
arxiv created 2013/01/10 · arxiv updated 2013/01/14
Abstract
The direct effect of one eventon another can be defined and measured byholding constant all intermediate variables between the two.Indirect effects present conceptual andpractical difficulties (in nonlinear models), because they cannot be isolated by holding certain variablesconstant. This paper shows a way of defining any path-specific effectthat does not invoke blocking the remainingpaths.This permits the assessment of a more naturaltype of direct and indirect effects, one thatis applicable in both linear and nonlinear models. The paper establishesconditions under which such assessments can be estimated consistentlyfrom experimental and nonexperimental data,and thus extends path-analytic techniques tononlinear and nonparametric models.
Citations
Cited by
- Is Chain-of-Thought Really Not Explainability? Chain-of-Thought Can Be Faithful without Hint Verbalization
- The positivity assumption in causal mediation analyses? Checked!
- Measuring Mechanistic Independence: Can Bias Be Removed Without Erasing Demographics?
- In-Context Algebra
- Estimation of Heterogeneous Causal Mediation Effects in a Hypertension Treatment Trial
- Bounds for causal mediation effects
- Causal Inference: A Tale of Three Frameworks
- Pseudo-strata learning via maximizing misclassification reward
- Representation-Level Counterfactual Calibration for Debiased Zero-Shot Recognition
- Linear Causal Discovery with Interventional Constraints
- On the magnitude, sign and ranking of recanting-twin path-specific effects
- Testing Full Mediation of Treatment Effects and the Identifiability of Causal Mechanisms
- A Distributional Perspective on Pearl's Causal Hierarchy: From Marginal to Joint and Individualized Potential Outcomes
- Learning density ratios in causal inference using Bregman-Riesz regression
- Reframing cross-world independence for identifying path-specific effects
- Sparsely gated tiny linear experts
- Counterfactual Identifiability via Dynamic Optimal Transport
- What Drives Compositional Generalization? The Importance of Continuous Training Objectives in Visual Generative Models
- Mechanistic Interpretability as Statistical Estimation: A Variance Analysis
- Partial Identification Approach to Counterfactual Fairness Assessment
- Calibrated Counterfactual Conformal Fairness (C3F): Post-hoc, Shift-Aware Coverage Parity via Conformal Prediction and Counterfactual Regularization
- RAPSEM: Identifying Latent Mediators Without Sequential Ignorability via a Rank-Preserving Structural Equation Model
- Towards Understanding Subliminal Learning: When and How Hidden Biases Transfer
- Causally-Enhanced Reinforcement Policy Optimization
- Mediation Analysis with Multiple Mediators Subject to Missing Not at Random
- Position, Not Provenance: Separating Reasoning Mediation from Sycophancy in Medical Vision-Language Models
- Using non-targeted HPV infections in studies with risk compensation
- DISCO: Disentangled Communication Steering for Large Language Models
- A Tutorial on Conducting Mediation Analysis with Exposure Mixtures
- Statistical Methods in Generative AI
- Crosscoding Through Time: Tracking Emergence & Consolidation Of Linguistic Representations Throughout LLM Pretraining
- Causal mechanism and mediation analysis for macroeconomics dynamics: a bridge of Granger and Sims causality
- Nonparametric Bounds in Causal Mediation Analysis in the Presence of Unmeasured Confounding and Imperfect Compliance
- Mediation Analysis in the Presence of Sample Selection Bias with an Application to Disparities in Liver Transplantation Listing
- From Indirect Object Identification to Syllogisms: Exploring Binary Mechanisms in Transformer Circuits
- Orientability of Causal Relations in Time Series using Summary Causal Graphs and Faithful Distributions
- Detecting and Characterizing Planning in Language Models
- How Causal Abstraction Underpins Computational Explanation
- Identifying Unmeasured Confounders in Panel Causal Models: A Two-Stage LM-Wald Approach
- A Framework for FAIR and CLEAR Ecological Data and Knowledge: Semantic Units for Synthesis and Causal Modelling
- Who cuts emissions, who turns up the heat? causal machine learning estimates of energy efficiency interventions
- Dissecting Persona-Driven Reasoning in Language Models via Activation Patching
- Causal Mediation Analysis for Zero-inflated Mixture Mediators
- Riesz representers for the rest of us
- Ethnic and social class discrimination in education: Experimental evidence from Germany
- Explaining Causal Findings Without Bias: Detecting and Assessing Direct Effects
- Eight Myths About Causality and Structural Equation Models
- Causal mediation analysis with double machine learning
- Combining Statistical and Causal Mediation Analysis
- On causal inference in the presence of interference
- Endogenous Selection Bias: The Problem of Conditioning on a Collider Variable
- A causal framework for classical statistical estimands in failure‐time settings with competing events
- Learning Individually Fair Classifier with Path-Specific Causal-Effect Constraint
- Contrasting Identifying Assumptions of Average Causal Effects: Robustness and Semiparametric Efficiency
- Decomposition of the Total Effect for Two Mediators: A Natural Counterfactual Interaction Effect Framework
- Nested Counterfactual Identification from Arbitrary Surrogate Experiments
- Causal Abstractions of Neural Networks
- Testing Mediation Effects Using Logic of Boolean Matrices
- Unbiased Scene Graph Generation from Biased Training
- Identification of causal intervention effects under contagion
- Some Doubly and Multiply Robust Estimators of Controlled Direct Effects
- Counterfactual fairness: removing direct effects through regularization
- Path-Free Decomposition for Direct, Indirect and Interaction Effects in Mediation Analysis
- Visual Relationship Detection using Scene Graphs: A Survey
- Same Task, Different Circuits: Disentangling Modality-Specific Mechanisms in VLMs
- Counterfactual Variable Control for Robust and Interpretable Question Answering
- Bayesian Mediation Analysis for Individualized Treatment Rules
- Causal Models and Learning from Data
- Causal indirect effect of an HIV curative treatment: mediators subject to an assay limit and measurement error
- Unequal Democracies
- Mediation Analysis: A Practitioner's Guide
- On relation between separable indirect effect, natural indirect effect, and interventional indirect effect
- Transformers Don't Need LayerNorm at Inference Time: Scaling LayerNorm Removal to GPT-2 XL and the Implications for Mechanistic Interpretability
- Efficient nonparametric estimation of causal mediation effects
- Diffusion Counterfactual Generation with Semantic Abduction
- Mechanistic Interpretability Needs Philosophy
- Latent Concept Disentanglement in Transformer-based Language Models
- Visual symbolic mechanisms: Emergent symbol processing in vision language models
- Causal Mediation Analysis with Multiple Mediators: A Simulation Approach
- DISCO: Mitigating Bias in Deep Learning with Conditional Distance Correlation
- Optimal Adjustment Sets for Nonparametric Estimation of Weighted Controlled Direct Effect
- When Measurement Mediates the Effect of Interest
- Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety
- Residual Balancing: A Method of Constructing Weights for Marginal Structural Models
- Inference for natural mediation effects under case-cohort sampling with applications in identifying COVID-19 vaccine correlates of protection
- Ratio of Mediator Probability Weighting for Estimating Natural Direct and Indirect Effects
- Local Markov Equivalence for PC-style Local Causal Discovery and Identification of Controlled Direct Effects
- M-learner:A Flexible And Powerful Framework To Study Heterogeneous Treatment Effect In Mediation Model
- Regression-based approach for natural direct and indirect relative risk in case of multiple mediators
- Walking the Tightrope: Disentangling Beneficial and Detrimental Drifts in Non-Stationary Custom-Tuning
- Causal Mediation Analysis with Multiple Treatments and Latent Confounders
- Causal mediation analysis with one or multiple mediators: a comparative study
- Post-treatment problems: What can we say about the effect of a treatment among sub-groups who (would) respond in some way?
- Decomposition of Probabilities of Causation with Two Mediators
- Identification of Causal Diffusion Effects Under Structural Stationarity
- The Do-Calculus Revisited
- Personalizing Path-Specific Effects
- Causal Etiology of the Research of James M. Robins
- Organic direct and indirect effects with post-treatment common causes of mediator and outcome
- Semiparametric theory for causal mediation analysis: Efficiency bounds, multiple robustness and sensitivity analysis
- Sunlight and vitamin D for bone health and prevention of autoimmune diseases, cancers, and cardiovascular disease
- Do Language Models Track Entities Across State Changes?
- Reverse-engineering censorship in China: Randomized experimentation and participant observation
- An Introduction to Causal Inference
- GoT-CD: Graph-of-Thoughts Causal Discovery and the Fragility of Post-hoc Path-Specific Fairness Audits
- Definition, Identification, and Estimation of the Direct and Indirect Number Needed to Treat
- Improving Fair Predictions Using Variational Inference In Causal Models
- 3. The Foundations of Causal Inference
- Multimodal Neuroimaging Data Integration and Pathway Analysis
- Causal machine learning for high-dimensional mediation analysis using interventional effects mapped to a target trial
- Scaling sparse feature circuit finding for in-context learning
- Gender differences in wage expectations
- Path-specific Effects Based on Information Accounts of Causality
- Decomposition of Total Effect with the Notion of Natural Counterfactual Interaction Effect
- MIB: A Mechanistic Interpretability Benchmark
- How do Large Language Models Understand Relevance? A Mechanistic Interpretability Perspective
Related