VanderWeele T.J. Explanation in Causal Inference: Methods for Mediation and Interaction

Oxford: Oxford University Press, 2015. – 706 p. – ISBN: 978-0-19-932587-0The book provides an accessible but comprehensive overview of methods for mediation and interaction. There has been considerable and rapid methodological development on mediation and moderation/interaction analysis within the causal-inference literature over the last ten years. Much of this material appears in a variety of specialized journals, and some of the papers are quite technical. There has also been considerable interest in these developments from empirical researchers in the social and biomedical sciences. However, much of the material is not currently in a format that is accessible to them. The book closes these gaps by providing an accessible, comprehensive, book-length coverage of mediation.
The book begins with a comprehensive introduction to mediation analysis, including chapters on concepts for mediation, regression-based methods, sensitivity analysis, time-to-event outcomes, methods for multiple mediators, methods for time-varying mediation and longitudinal data, and relations between mediation and other concepts involving intermediates such as surrogates, principal stratification, instrumental variables, and Mendelian randomization. The second part of the book concerns interaction or "moderation," including concepts for interaction, statistical interaction, confounding and interaction, mechanistic interaction, bias analysis for interaction, interaction in genetic studies, and power and sample-size calculation for interaction. The final part of the book provides comprehensive discussion about the relationships between mediation and interaction and unites these concepts within a single framework. This final part also provides an introduction to spillover effects or social interaction, concluding with a discussion of social-network analyses.
The book is written to be accessible to anyone with a basic knowledge of statistics. Comprehensive appendices provide more technical details for the interested reader. Applied empirical examples from a variety of fields are given throughout. Software implementation in SAS, Stata, SPSS, and R is provided. The book should be accessible to students and researchers who have completed a first-year graduate sequence in quantitative methods in one of the social- or biomedical-sciences disciplines. The book will only presuppose familiarity with linear and logistic regression, and could potentially be used as an advanced undergraduate book as well.Mediation Analysis
Explanation and Mechanism
Causal Inference and Explanation
Forms of Explanation and Types of Mechanisms
Motivations for Assessing Mediation, Interaction, and Interference
Organization of this Book
Mediation: Introduction and Regression-Based Approaches
Classic Regression Approach to Mediation Analysis
Counterfactual Approach to Mediation Analysis: Continuous Outcomes
Assumptions about Confounding
Binary and Count Outcomes
Binary Mediators
Comparison of Approaches: Product-of-Coefficient and Difference Methods
Description of the SAS Macro
Description of the SPSS Macro
Description of the Stata Macro
Hypothetical Example with Output
Empirical Example in Genetic Epidemiology
When to Include an Exposure–Mediator Interaction
Proportion Mediated
Proportion Eliminated
Study Design and Mediation Analysis
Counterfactual Notation for Natural Direct and Indirect Effects
An Alternative Regression-Based Estimation Approach Using Simulations
Code for the Simulation-Based Approach in R
Discussion
Sensitivity Analysis for Mediation
Sensitivity Analysis for Unmeasured Confounding for Total Effects
Sensitivity Analysis for Unmeasured Confounding for Controlled Direct Effects
Sensitivity Analysis for Unmeasured Confounding for Natural Direct and Indirect Effects
Sensitivity Analysis Using Two Trials
Sensitivity Analysis for Direct and Indirect Effects in the Presence of Measurement Error
Discussion
Mediation Analysis with Survival Data
Earlier Literature on Mediation Analysis with Survival Models
Mediation Analysis with an Accelerated Failure Time Model
Mediation Analysis with a Proportional Hazards Model
Mediation with an Additive Hazard Model
A Weighting Approach to Direct and Indirect Effects with Survival Outcomes
Sensitivity Analysis with Survival Data
Discussion
Multiple Mediators
Regression-Based Approaches to Multiple Mediators
A Weighting Approach to Multiple Mediators
Controlled Direct Effects and Exposure-Induced Confounding
Effect Decomposition with Exposure-Induced Confounding
Path-Specific Effects
Sensitivity Analysis for Exposure-Induced Confounding
Discussion
Mediation Analysis with Time-Varying Exposures and Mediators
Notation and Definitions
Controlled Direct Effects with Time-Varying Exposures and Mediators
Natural Direct and Indirect Effects and their Randomized Interventional Analogues with Time-Varying Exposures and Mediators
Counterfactual Analysis of MacKinnon’s Three-Wave Mediation Model
Discussion
Selected Topics in Mediation Analysis
Other Estimation Approaches
Ill-Defined Mediators and Multiple Versions of the Mediator
Controversies Over Assumptions and Alternative Interpretations of Effects
Direct and Indirect Effects in Health Disparities Research
Rubin’s Seemingly Problematic Examples
A Three-Way Decomposition into Direct, Indirect, and Interactive Effects
Alternative Identification Strategies Using Confounding Control
Identification Using Baseline Covariates that Interact with Exposure
Power and Sample Size Calculations for Mediation Analysis
Discussion
Other Topics Related to Intermediates
Principal Stratification
Surrogate Outcomes
Instrumental Variables
Mendelian Randomization
Discussion
Interaction Analysis
An Introduction to Interaction Analysis
Measures of Interaction and Scale of Interaction
Statistical Interactions and Statistical Inference
Inference for Additive Interaction
SAS and Stata Code for Additive Interaction from Logistic Regression
Additive Versus Multiplicative Interaction
Confounding and the Interpretation of Interaction: Interaction Versus Effect Heterogeneity
Presenting Interaction Analyses
Synergism and Mechanistic Interaction
Interactions for Continuous Outcomes and Time-to-Event Outcomes
Identifying Subgroups to Target Treatment
Qualitative Interaction
Attributing Effects to Interactions
Discussion
Mechanistic Interaction
Sufficient Causes and Synergism
Statistical Interaction with No Mechanistic Interaction
Empirical Tests for Sufficient Cause Synergism
Sufficient Cause Interaction and Statistical Interactions
“Epistatic” or Singular Interactions
Extensions to Ordinal Exposures
Extensions to Three or More Exposures
Other Extensions
Antagonism
Limits of Inference Concerning Biology
Discussion
Bias Analysis for Interactions
Sensitivity Analysis and Robustness for Additive Interaction
Sensitivity Analysis and Robustness for Multiplicative Interaction
Sensitivity Analysis for the Relative Excess Risk Due to Interaction
Measurement Error and Additive Interaction
Measurement Error and Multiplicative Interaction
Discussion
Interaction in Genetics: Independence and Boosting Power
Case-Only Estimators of Interaction
Joint Tests for Interactions and Main Effects
Multiple Testing
Discussion
Power and Sample-Size Calculations for Interaction Analysis
Power and Sample-Size Calculations for Interaction for Continuous Outcomes
Power and Sample-Size Calculations for Binary Outcomes: Multiplicative Interaction
Power and Sample Size Calculations for Binary Outcomes: Additive Interaction
Power and Sample Size Calculations for Binary Outcomes: Mechanistic Interaction
Excel Spreadsheets for Sample-Size and Power Calculations for Additive and Multiplicative Interaction for a Binary Outcome
Discussion
Synthesis and Spillover Effects
A Unification of Mediation and Interaction
Notation and Definitions
Fourfold Decomposition: The Unification of Mediation and Interaction
Identification of the Effects
Relation to Statistical Models
Binary Outcomes and the Ratio Scale
Illustration in Genetic Epidemiology
Relation to Mediation Decompositions
Relation to Interaction Decompositions
SAS Code for the Four-Way Decomposition
Discussion
Social Interactions and Spillover Effects
Notation and Definitions for Spillover Effects
Basic Spillover and Individual/Direct Effects
Assessing “Infectiousness” Effects
Contagion versus Infectiousness Effects
Tests for Specific Forms of Interference Using Causal Interactions
Inferential Challenges with Many Individuals per Cluster
Spillover Effects and Observational Data
Spillover Effects and Social Networks
Discussion
Mediation and Interaction: Future and Context
The Present State of Methods and Future Methodological Development
Philosophical Questions
Appendix. Technical Details and Proofs