Causal Inference

• Main Author: Rosenbaum, Paul R
• Format: eBook Book
• Language: English
• Published: Cambridge MIT Press 2023; The MIT Press
• Edition: 1
• Series: The MIT Press Essential Knowledge Series
• Subjects: Causation; Inference; Mathematical statistics; Observation (Scientific method); Probabilities; Science; Science > Experiments; randomized experiment; sensitivity analysis; Causation; causality; observational study; epidemiology; health outcomes; randomized trials; treatment effects; econometrics; propensity; data science; randomization; program evaluation; medical statistics; policy analysis; clinical trials; public health; statistics

A nontechnical guide to the basic ideas of modern causal inference, with illustrations from health, the economy, and public policy. Which of two antiviral drugs does the most to save people infected with Ebola virus? Does a daily glass of wine prolong or shorten life? Does winning the lottery make y...

Table of Contents:

• Intro -- Contents -- Series Foreword -- List of Examples -- List of Methodological Topics -- 1 The Effects Caused by Treatments -- 2 Randomized Experiments -- 3 Observational Studies: The Problem -- 4 Adjustments for Measured Covariates -- 5 Sensitivity to Unmeasured Covariates -- 6 Quasi-Experimental Devices in the Design of Observational Studies -- 7 Natural Experiments, Discontinuities, and Instruments -- 8 Replication, Resolution, and Evidence Factors -- 9 Uncertainty and Complexity in Causal Inference -- Postscript: Key Ideas, Chapter by Chapter -- Glossary -- Notes -- Bibliography -- Further Reading -- Index
• 7: Natural Experiments, Discontinuities, and Instruments -- Bits and Pieces of Random Assignment in an Otherwise Biased World -- Natural Experiments from Lotteries -- Nature's Natural Experiment, I: The Genes of Siblings -- Nature's Natural Experiment, II: Hypothetical Siblings -- Discontinuity Designs as Natural Experiments -- Encouragement Experiments: Can You Learn about One Treatment When You Randomize Another? -- Instrumental Variables and the Complier Average Causal Effect -- The Effect of Being Offered a Housing Voucher or the Effect of Accepting It -- Choosing Situations in Which Biases in Treatment Assignment Are Smaller -- 8: Replication, Resolution, and Evidence Factors -- Replication Is Not Repetition -- Repetition without Resolution -- Varied Views of a Single Object -- Evidence Factors -- 9: Uncertainty and Complexity in Causal Inference -- Are Small Daily Doses of Alcohol Beneficial? -- Oncologists versus Cardiologists -- A Dissenting Voice from a New Tactic: Mendelian Randomization -- The Answer Might Be Complex -- A Traditional Toxin -- Total Mortality -- Is Part or All of the Supposed Heart Benefit Simply a Mistake? -- So Are Small Daily Doses of Alcohol Beneficial? -- Postscript -- Glossary -- Notes -- Chapter 1 -- Chapter 2 -- Chapter 3 -- Chapter 4 -- Chapter 5 -- Chapter 6 -- Chapter 7 -- Chapter 8 -- Chapter 9 -- Bibliography -- Further Reading -- Index
• Intro -- Contents -- Series Foreword -- List of Examples -- List of Methodological Topics -- 1: The Effects Caused by Treatments -- What Is a Causal Effect? Why Is Causal Inference Difficult? -- Notation for a Precise Question -- Would a Control Group Solve the Problem? -- It Is Just the Same with More Than Two People -- Flipping a Fair Coin to Assign Treatments to People -- The Average Treatment Effect (ATE) -- 2: Randomized Experiments -- The Effects of Treatments for Ebola Virus Disease -- A Randomized Controlled Trial -- Why Assign Treatments at Random? -- Randomized Treatment Assignment and Causal Inference -- Estimating the Average Treatment Effect -- Testing the Hypothesis of No Effect -- What Is Special about Coin Flips? -- Bleeding George Washington and the Theory of Humors -- 3: Observational Studies -- What Are Observational Studies? -- Smoking and Periodontal Disease -- Interlude: Tukey's Boxplot -- Do Treated and Control Groups Look Comparable in Boxplots? -- Comparing Outcomes in Groups That Are Not Comparable -- 4: Adjustments for Measured Covariates -- Matching for Covariates as a Method of Adjustment -- Imbalances in the Propensity Score -- How Matching for the Propensity Score Balances Measured Covariates -- Comparing Outcomes While Controlling Measured Covariates -- Other Matching Techniques -- 5: Sensitivity to Unmeasured Covariates -- Objections, Counterclaims, and Rival Hypotheses -- Smoking and Lung Cancer -- The First Sensitivity Analysis in an Observational Study -- Modern Sensitivity Analysis: Smoking and Periodontal Disease -- The Role of Sensitivity Analysis -- 6: Quasi-Experimental Devices in the Design of Observational Studies -- Anticipated Counterclaims -- Two Control Groups -- The Logic of Two Control Groups -- Untreated Counterparts in Addition to Untreated Controls -- Resolving Anticipated Counterclaims