Causal Analysis

Impact Evaluation and Causal Machine Learning with Applications in R

by Huber

ISBN: 9780262374910 | Copyright 2023

Click here to preview

Instructor Requests

Digital Exam/Desk Copy Print Desk Copy Ancillaries
Tabs

A comprehensive and cutting-edge introduction to quantitative methods of causal analysis, including new trends in machine learning.

Reasoning about cause and effect—the consequence of doing one thing versus another—is an integral part of our lives as human beings. In an increasingly digital and data-driven economy, the importance of sophisticated causal analysis only deepens. Presenting the most important quantitative methods for evaluating causal effects, this textbook provides graduate students and researchers with a clear and comprehensive introduction to the causal analysis of empirical data. Martin Huber's accessible approach highlights the intuition and motivation behind various methods while also providing formal discussions of key concepts using statistical notation. Causal Analysis covers several methodological developments not covered in other texts, including new trends in machine learning, the evaluation of interaction or interference effects, and recent research designs such as bunching or kink designs.

•Most complete and cutting-edge introduction to causal analysis, including causal machine learning
•Clean presentation of rigorous material avoids extraneous detail and emphasizes conceptual analogies over statistical notation
•Supplies a range of applications and practical examples using R

Expand/Collapse All
Contents (pg. vii)
1. Introduction (pg. 1)
1.1 About Causality and This Book (pg. 1)
1.2 Overview of Topics (pg. 3)
2. Causality and No Causality (pg. 11)
2.1 Potential Outcomes, Causal Effects, and the Stable Unit Treatment Value Assumption (pg. 11)
2.2 Treatment Selection Bias (pg. 14)
3. Social Experiments and Linear Regression (pg. 19)
3.1 Social Experiments (pg. 19)
3.2 Effect Identification by Linear Regression (pg. 22)
3.3 Estimation by Linear Regression and Its Properties (pg. 26)
3.4 Variance Estimation, Inference, and Goodness of Fit (pg. 35)
3.5 Extensions to Multiple or Continuous Treatments (pg. 49)
3.6 Including Covariates (pg. 58)
4. Selection on Observables (pg. 65)
4.1 Identification under Selection on Observables (pg. 65)
4.2 Linear, Series, and Kernel Regression (pg. 70)
4.3 Covariate Matching (pg. 80)
4.4 Propensity Score Matching (pg. 88)
4.5 Inverse Probability Weighting, Empirical Likelihood, and Entropy Balancing (pg. 94)
4.6 Doubly Robust Methods (pg. 99)
4.7 Practical Issues: Common Support and Match Quality (pg. 102)
4.8 Multivalued or Continuous Treatments and Distributional Effects (pg. 113)
4.9 Dynamic Treatment Effects (pg. 120)
4.10 Causal Mechanisms (Mediation Analysis) (pg. 126)
4.11 Outcome Attrition and Posttreatment Sample Selection (pg. 133)
5. Causal Machine Learning (pg. 137)
5.1 Motivation for Machine Learning and Fields of Application (pg. 137)
5.2 Double Machine Learning and Partialling out with Lasso Regression (pg. 138)
5.3 A Survey of Further Machine Learning Algorithms (pg. 145)
5.4 Effect Heterogeneity (pg. 151)
5.5 Optimal Policy Learning (pg. 159)
5.6 Reinforcement Learning (pg. 163)
6. Instrumental Variables (pg. 169)
6.1 Evaluation of the Local Average Treatment Effect (pg. 169)
6.2 Instrumental Variable Methods with Covariates (pg. 177)
6.3 Nonbinary Instruments and Treatments (pg. 184)
6.4 Sample Selection, Dynamic and Multiple Treatments, and Causal Mechanisms (pg. 189)
7. Difference-in-Differences (pg. 195)
7.1 Difference-in-Differences without Covariates (pg. 195)
7.2 Difference-in-Differences with Covariates (pg. 203)
7.3 Multiple Periods of Treatment Introduction (pg. 207)
7.4 Changes-in-Changes (pg. 213)
8. Synthetic Controls (pg. 219)
8.1 Estimation and Inference with a Single Treated Unit (pg. 219)
8.2 Alternative Estimators and Multiple Treated Units (pg. 225)
9. Regression Discontinuity, Kink, and Bunching Designs (pg. 231)
9.1 Sharp and Fuzzy Regression Discontinuity Designs (pg. 231)
9.2 Sharp and Fuzzy Regression Kink Designs (pg. 243)
9.3 Bunching Designs (pg. 248)
10. Partial Identification and Sensitivity Analysis (pg. 255)
10.1 Partial Identification (pg. 255)
10.2 Sensitivity Analysis (pg. 264)
11. Treatment Evaluation under Interference Effects (pg. 271)
11.1 Failure of the Stable Unit Treatment Value Assumption (pg. 271)
11.2 Partial Interference (pg. 272)
11.3 Interference Based on Exposure Mappings (pg. 281)
12. Conclusion (pg. 285)
References (pg. 287)
Index (pg. 311)

Martin Huber

Martin Huber is Professor of Applied Econometrics at the University of Fribourg, Switzerland, where his research comprises both methodological and applied contributions in the fields of causal analysis and policy evaluation, machine learning, statistics, econometrics, and empirical economics.

eTextbook
Go paperless today! Available online anytime, nothing to download or install.

Features

  • Bookmarking
  • Note taking
  • Highlighting