>> Home >> Bookstore >> Econometrics >> Propensity Score Analysis: Statistical Methods and Applications

Propensity Score Analysis: Statistical Methods and Applications

Shenyang Guo and Mark W. Fraser
Publisher: Sage
Copyright: 2009
ISBN-13: 978-1-4129-5356-6
Pages: 370; hardcover
Price: $82.50

Comment from the Stata technical group

Propensity Score Analysis by Shenyang Guo and Mark W. Fraser is an excellent book on estimating treatment effects from observational data. Researchers and graduate students interested in the analysis of observation data will find this book invaluable. This book is the first to provide step-by-step instructions for using Stata to estimate treatments by propensity-score analysis. In addition, this book also covers Heckman’s sample-selection estimator, nearest-neighbor matching estimators, propensity-score matching, and propensity-score nonparametric regression estimators.

After providing a useful history of these techniques, the authors discuss the issues involved in using observational data. The authors have consolidated the notation and theory found in the propensity-score literature and have carried this notation thr oughout the book. This book introduces the philosophy and methods of matching estimation in an approachable manner, even for those unfamiliar with the subject.

Readers embark on an interesting journey from the early formulation of the Neyman–Rubin counterfactual framework to a discussion of current methods and research issues. These ideas are discussed at theoretical and applied levels. The authors formally describe the assumptions underlying the analyses in a way that provides insight into the issues related to applying these techniques to practical problems.

Propensity-score analysis has it roots in both economics and statistics. Economists will find the discussion of Heckman’s sample-selection estimator familiar, and statisticians will find the Rosenbaum and Rubin counterfactual framework familiar. The authors do a great job is bringing these two viewpoints together in a unified, balanced manner.

While this book makes an excellent text for a graduate-level course on the analysis of observational data, it is readily accessible to researchers using observational data to estimate treatment effects. This book should be a required read for students in economics and statistics.

Table of contents

List of Tables
List of Figures
1. Introduction
1.1 Observational Studies
1.2 History and Development
1.3 Randomized Experiments
1.3.1 Fisher’s Randomized Experiment
1.3.2 Types of Randomized Experiments and Statistical Tests
1.3.3 Critiques of Social Experimentation
1.4 Why and When a Propensity Score Analysis Is Needed
1.5 Computing Software Packages
1.6 Plan of the Book
2. Counterfactual Framework and Assumptions
2.1 Causality, Internal Validity, and Threats
2.2 Counterfactuals and the Neyman–Rubin Counterfactual Framework
2.3 The Ignorable Treatment Assignment Assumption
2.4 The Stable Unit Treatment Value Assumption
2.5 Methods to Estimate Treatment Effects
2.5.1 The Four Models
2.5.2 Other Balancing Methods
2.6 The Underlying Logic of Statistical Inference
2.7 Types of Treatment Effects
2.8 Heckman’s Econometric Model of Causality
2.9 Conclusions
3. Conventional Methods for Data Balancing
3.1 Why is Data Balancing Necessary? A Heuristic Example
3.2 Three Methods of Data Balancing
3.2.1 The Ordinary Least Squares Regression
3.2.2 Matching
3.2.3 Stratification
3.3 Design of the Data Simulation
3.4 Results of the Data Simulation
3.5 Implications of the Data Simulation
3.6 Key Issues Regarding the Application of OLS Regression
3.7 Conclusions
4. Sample Selection and Related Models
4.1 The Sample Selection Model
4.1.1 Truncation, Censoring, and Incidental Truncation
4.1.2 Why Is It Important to Model Sample Selection?
4.1.3 Moments of an Incidentally Truncated Bivariate Normal Distribution
4.1.4 The Heckman Model and Its Two-Step Estimator
4.2 Treatment Effect Model
4.3 Instrumental Variables Estimator
4.4 Overview of the Stata Programs and Main Features of treatreg
4.5 Examples
4.5.1 Application of the Treatment Effect Model to Analysis of Observational Data
4.5.2 Evaluation of Treatment Effects From a Program With a Group of Randomization Design
4.5.3 Running the Treatment Effect Model After Multiple Imputations of Missing Data
4.6 Conclusions
5. Propensity Score Matching and Related Models
5.1 Overview
5.2 The Problem of Dimensionality and the Properties of Propensity Scores
5.3 Estimating Propensity Scores
5.3.1 Binary Logistic Regression
5.3.2 Strategies to Specify a Correct Model Predicting Propensity Scores
5.3.3 Hirano and Imbens’s Method for Specifying Predictors Relying on Predetermined Critical t Values
5.3.4 Generalized Boosted Modeling
5.4 Matching
5.4.1 Greedy Matching
5.4.2 Optimal Matching
5.4.3 Fine Balance
5.5 Postmatching Analysis
5.5.1 Multivariate Analysis After Greedy Matching
5.5.2 Stratification After Greedy Matching
5.5.3 Computing Indices of Covariate Imbalance
5.5.4 Outcome Analysis Using the Hodges–Lehmann Aligned Rank Test After Optimal Matching
5.5.5 Regression Adjustment Based on Sample Created by Optimal Pair Matching
5.5.6 Regression Adjustment Using Hodges–Lehmann Aligned Rank Scores After Optimal Matching
5.6 Propensity Score Weighting
5.7 Modeling Doses of Treatment
5.8 Overview of the Stata and R Programs
5.9 Examples
5.9.1 Greedy Matching and Subsequent Analysis of Hazard Rates
5.9.2 Optimal Matching
5.9.3 Post–Full Matching Analysis Using the Hodges–Lehmann Aligned Rank Test
5.9.4 Post–Pair Matching Analysis Using Regression of Difference Scores
5.9.5 Propensity Score Weighting
5.9.6 Modeling Doses of Treatment
5.9.7 Comparison of Models and Conclusions of the Study of the Impact of Poverty on Child Academic Achievement
5.9.8 Comparison of Rand-gbm and Stata’s Boost Algorithms
5.10 Conclusions
6. Matching Estimators
6.1 Overview
6.2 Methods of Matching Estimators
6.2.1 Simple Matching Estimator
6.2.2 Bias-Corrected Matching Estimator
6.2.3 Variance Estimator Assuming Homoscedasticity
6.2.4 Variance Estimator Allowing for Heteroscedasticity
6.2.5 Large Sample Properties and Correction
6.3 Overview of the Stata Program nnmatch
6.4 Examples
6.4.1 Matching With Bias-Corrected and Robust Variance Estimators
6.4.2 Efficacy Subset Analysis With Matching Estimators
6.5 Conclusions
7. Propensity Score Analysis With Nonparametric Regression
7.1 Overview
7.2 Methods of Propensity Score Analysis With Nonparametric Regression
7.2.1 The Kernel-Based Matching Estimators
7.2.2 Review of the Basic Concepts of Local Linear Regression (lowess)
7.2.3 Asymptotic and Finite-Sample Properties of Kernel and Local Linear Matching
7.3 Overview of the Stata Programs psmatch2 and bootstrap
7.4 Examples
7.4.1 Analysis of Difference-in-Differences
7.4.2 Application of Kernel-Based Matching to One-Point Data
7.5 Conclusions
8. Selection Bias and Sensitivity Analysis
8.1 Selection Bias: An Overview
8.1.1 Sources of Selection Bias
8.1.2 Overt Bias Versus Hidden Bias
8.1.3 Consequences of Selection Bias
8.1.4 Strategies to Correct for Selection Bias
8.2 A Monte Carlo Strategy Comparing Corrective Models
8.2.1 Design of the Monte Carlo Study
8.2.2 Results of the Monte Carlo Study
8.2.3 Implications
8.3 Rosenbaum’s Sensitivity Analysis
8.3.1 The Basic Idea
8.3.2 Illustration of the Wilcoxon’s Signed-Rank Test for Sensitivity Analysis of Matched Pair Study
8.4 Overview of the Stata Program rbounds
8.5 Examples
8.5.1 Sensitivity Analysis of the Effects of Lead Exposure
8.5.2 Sensitivity Analysis for the Study Using Pair Matching
8.6 Conclusions
9. Concluding Remarks
9.1 Common Pitfalls in Observational Studies: A Checklist for Critical Review
9.2 Approximating Experiments With Propensity Score Approaches
9.2.1 Criticism of Propensity Score Methods
9.2.2 Criticism of Sensitivity Analysis (Γ)
9.2.3 Group Randomized Trials
9.3 Other Advances in Modeling Causality
9.4 Directions for Future Development
Author Index
Subject Index
About the Authors
The Stata Blog: Not Elsewhere Classified Find us on Facebook Follow us on Twitter LinkedIn Google+ Watch us on YouTube