Demystifying Double Robustness
A Comparison of Alternative Strategies for Estimating a Population Mean from Incomplete Data
Bibliographic Data
| ID | 23319307 |
|---|---|
| Authors | Joseph D Y Kang, Joseph L Schafer |
| Year | 2007 |
| Volume | 22 |
| Issue | 4 |
| Pages | 569-573 |
| Publication date | 2007-11-01 |
| Peer Reviewed | Yes |
| Open Access | No |
| Type | ARTICLE |
| Venue | Statistical Science (JOURNAL) |
| Journal identifiers | ISSN: 0883-4237 • E-ISSN: 2168-8745 |
| Publisher | Institute of Mathematical Statistics (PUBLISHER • US) |
| DOI | 10.1214/07-sts227 |
| PMID | 18516239 |
| PMCID | PMC2397555 |
| OpenAlex | W2761665523 |
| Language | EN |
| Citations received | 69 |
| References cited | 42 |
When outcomes are missing for reasons beyond an investigator’s control, there are two different ways to adjust a parameter estimate for covariates that may be related both to the outcome and to missingness. One approach is to model the relationships between the covariates and the outcome and use those relationships to predict the missing values. Another is to model the probabilities of missingness given the covariates and incorporate them into a weighted or stratified estimate. Doubly robust (DR) procedures apply both types of model simultaneously and produce a consistent estimate of the parameter if either of the two models has been correctly specified. In this article, we show that DR estimates can be constructed in many ways. We compare the performance of various DR and non-DR estimates of a population mean in a simulated example where both models are incorrect but neither is grossly misspecified. Methods that use inverse-probabilities as weights, whether they are DR or not, are sensitive to misspecification of the propensity model when some estimated propensities are small. Many DR methods perform better than simple inverse-probability weighting. None of the DR methods we tried, however, improved upon the performance of simple regression-based prediction of the missing values. This study does not represent every missing-data problem that will arise in practice. But it does demonstrate that, in at least some settings, two wrong models are not better than one.
Covariate · Missing data · Outcome (game theory) · Population · Propensity score matching · Simple (philosophy) · Advanced Causal Inference Techniques · Meta-analysis and systematic reviews · Statistical Methods and Bayesian Inference
Sweet are the fruit of adversity? The impact of fathers’ death on child non-cognitive outcomes in Ethiopia
Supporting treatment adherence for resilience and thriving (START)
The causal effect and impact of reproductive factors on breast cancer using super learner and targeted maximum likelihood estimation
Semiparametric Estimation of Treatment Effects in Observational Studies with Heterogeneous Partial Interference
Estimating the effects of Head Start on parenting and child maltreatment
A Data Analysis Method for Using Longitudinal Binary Outcome Data from a SMART to Compare Adaptive Interventions
A Tutorial for Propensity Score Weighting for Moderation Analysis With Categorical Variables
Missing Data
Econometric Methods for Program Evaluation
Predictions put into practice
Weighting Regressions by Propensity Scores
A Quantile-Based g-Computation Approach to Addressing the Effects of Exposure Mixtures
Doubly Robust Estimation of Causal Effects
A tutorial on propensity score estimation for multiple treatments using generalized boosted models
Improving propensity score weighting using machine learning
Entropy Balancing is Doubly Robust
Bayesian Nonparametric Modeling for Causal Inference
Covariate balancing propensity score for a continuous treatment
Matching Methods for Causal Inference
Comparing Multiple Imputation and Propensity-Score Weighting in Unit-Nonresponse Adjustments
Early Parenting Program as Intervention Strategy for Emotional Distress in First-Time Mothers
A community-engaged randomized controlled trial of an integrative intervention with HIV-positive, methamphetamine-using men who have sex with men
A Primer on Inverse Probability of Treatment Weighting and Marginal Structural Models
Quasi-experimental designs for causal inference
Global Learning for Local Serving
Is Childhood Adversity Before Age 5 Associated with Adolescent and Young Adult Substance Use? Findings from a U.S. Prospective Cohort Study
Causal Effects of Time-Varying Exposures
The Effects of Charitable Giving on Life Satisfaction of Older Korean Adults
The Causal Impact of Segregation on a Disparity
The effects of neighborhood factors, maternal depression and parental emotional support on social – behavioral outcomes of children involved in the child welfare system
Community-driven violence reduction programs
A hidden cost of convenience
Occupational feminization, gender, and wages
On the nexus between gender diversity quota and R&D investment
Environmentalism among poor and rich countries
The effects of exposure to an enhanced preschool program on the social-emotional functioning of at-risk children
One versus two years
One or two years of participation
Effects of Single-Sex Schooling in the Final Years of High School
Assessing the Effectiveness of Projects Supporting On-Farm Conservation of Native Crops
Unpacking Complex Mediation Mechanisms and Their Heterogeneity Between Sites in a Job Corps Evaluation
Joint Effects in Cross-Lagged Panel Research Using Structural Nested Mean Models
Efficient Augmented Inverse Probability Weighted Estimation in Missing Data Problems
High-Dimensional Model-Assisted Inference for Local Average Treatment Effects With Instrumental Variables
Causal Pitfalls in the Decomposition of Wage Gaps
The Finite Sample Performance of Estimators for Mediation Analysis Under Sequential Conditional Independence
Genetic Matching for Estimating Causal Effects
New Evidence on the Finite Sample Properties of Propensity Score Reweighting and Matching Estimators
Tuning Random Forests for Causal Inference under Cluster-Level Unmeasured Confounding
A Causal View on Bias in Missing Data Imputation
A Systematic Review of Propensity Score Methods in the Social Sciences
How Bias Reduction Is Affected by Covariate Choice, Unreliability, and Mode of Data Analysis
An Evaluation of Weighting Methods Based on Propensity Scores to Reduce Selection Bias in Multilevel Observational Studies
Identifying and Estimating Causal Moderation for Treated and Targeted Subgroups
A Viable Alternative When Propensity Scores Fail
The EffectLiteR Approach for Analyzing Average and Conditional Effects
A Generalized Bootstrap Procedure of the Standard Error and Confidence Interval Estimation for Inverse Probability of Treatment Weighting
Alternate policing strategies
Toward an Analytical Criminology
The Impact of Medicare Part D on Medication Adherence Among Older Adults Enrolled in Medicare-Advantage Products
Tracing Causal Paths from Experimental and Observational Data
Estimating controlled direct effects through marginal structural models
Combining Double Sampling and Bounds to Address Nonignorable Missing Outcomes in Randomized Experiments
Residual Balancing
Why Propensity Scores Should Not Be Used for Matching
The Gap-Closing Estimand
Estimating the Effect of Perceived Risk of Crime on Social Trust in the Presence of Endogeneity Bias
Impact of Adopting Improved Seeds on Access to Broader Food Groups Among Small-Scale Sorghum Producers in Tanzania
A Propensity-Score-Weighted Population-Based Study of the Health Benefits of Dogs and Cats for Children
Applied Bayesian Modeling and Causal Inference from Incomplete‐Data Perspectives
Analysis of Incomplete Multivariate Data
Observational Studies
Bayesian Data Analysis
Model Assisted Survey Sampling
Generalized Linear Models
On the Variances of Asymptotically Normal Estimators from Complex Surveys
Doubly Robust Estimation in Missing Data and Causal Inference Models
Analysis of Semiparametric Regression Models for Repeated Outcomes in the Presence of Missing Data
Estimation of Causal Effects using Propensity Score Weighting
A Generalization of Sampling Without Replacement from a Finite Universe
Semiparametric Efficiency in Multivariate Regression Models with Missing Data
Survey Nonresponse Adjustments for Estimates of Means
The Effectiveness of Adjustment by Subclassification in Removing Bias in Observational Studies
Causal Inference Using Potential Outcomes
Stratification and weighting via the propensity score in estimation of causal treatment effects
On the Application of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9
Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group
Bayesian Analysis of Binary and Polychotomous Response Data
Estimation of Regression Coefficients When Some Regressors are not Always Observed
Bayesian Inference for Causal Effects
Inference and missing data
The central role of the propensity score in observational studies for causal effects
Estimating causal effects of treatments in randomized and nonrandomized studies.
Statistics and Causal Inference
Constructing a Control Group Using Multivariate Matched Sampling Methods That Incorporate the Propensity Score
Multiple Imputation for Nonresponse in Surveys
| Unique citing works | 69 |
|---|---|
| Citations per year | 3,83 |
| Citation span | 2008 - 2026 (19) |
| Citation velocity | current |
| Highly cited | No |
| Citation types | Neutral: 62 |