Doubly robust estimation of the local average treatment effect curve

Elizabeth Leigh Ogburn, Andrea Rotnitzky, James M. Robins

Research output: Contribution to journalArticle

Abstract

We consider estimation of the causal effect of a binary treatment on an outcome, conditionally on covariates, from observational studies or natural experiments in which there is a binary instrument for treatment. We describe a doubly robust, locally efficient estimator of the parameters indexing a model for the local average treatment effect conditionally on covariates V when randomization of the instrument is only true conditionally on a high dimensional vector of covariates X, possibly bigger than V. We discuss the surprising result that inference is identical to inference for the parameters of a model for an additive treatment effect on the treated conditionally on V that assumes no treatment-instrument interaction. We illustrate our methods with the estimation of the local average effect of participating in 401(k) retirement programmes on savings by using data from the US Census Bureau's 1991 Survey of Income and Program Participation.

Original languageEnglish (US)
Pages (from-to)373-396
Number of pages24
JournalJournal of the Royal Statistical Society. Series B: Statistical Methodology
Volume77
Issue number2
DOIs
StatePublished - Mar 1 2015

Fingerprint

Average Treatment Effect
Robust Estimation
Covariates
Curve
Binary
Causal Effect
Efficient Estimator
Observational Study
Census
Treatment Effects
Randomisation
Indexing
High-dimensional
Interaction
Model
Experiment
Average treatment effect
Robust estimation
Inference

Keywords

  • Instrumental variables
  • Local average treatment effect
  • Local efficiency
  • Multiplicative effect

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Doubly robust estimation of the local average treatment effect curve. / Ogburn, Elizabeth Leigh; Rotnitzky, Andrea; Robins, James M.

In: Journal of the Royal Statistical Society. Series B: Statistical Methodology, Vol. 77, No. 2, 01.03.2015, p. 373-396.

Research output: Contribution to journalArticle

@article{eb0065d1c270400fa4827231c734a0ed,
title = "Doubly robust estimation of the local average treatment effect curve",
abstract = "We consider estimation of the causal effect of a binary treatment on an outcome, conditionally on covariates, from observational studies or natural experiments in which there is a binary instrument for treatment. We describe a doubly robust, locally efficient estimator of the parameters indexing a model for the local average treatment effect conditionally on covariates V when randomization of the instrument is only true conditionally on a high dimensional vector of covariates X, possibly bigger than V. We discuss the surprising result that inference is identical to inference for the parameters of a model for an additive treatment effect on the treated conditionally on V that assumes no treatment-instrument interaction. We illustrate our methods with the estimation of the local average effect of participating in 401(k) retirement programmes on savings by using data from the US Census Bureau's 1991 Survey of Income and Program Participation.",
keywords = "Instrumental variables, Local average treatment effect, Local efficiency, Multiplicative effect",
author = "Ogburn, {Elizabeth Leigh} and Andrea Rotnitzky and Robins, {James M.}",
year = "2015",
month = "3",
day = "1",
doi = "10.1111/rssb.12078",
language = "English (US)",
volume = "77",
pages = "373--396",
journal = "Journal of the Royal Statistical Society. Series B: Statistical Methodology",
issn = "1369-7412",
publisher = "Wiley-Blackwell",
number = "2",

}

TY - JOUR

T1 - Doubly robust estimation of the local average treatment effect curve

AU - Ogburn, Elizabeth Leigh

AU - Rotnitzky, Andrea

AU - Robins, James M.

PY - 2015/3/1

Y1 - 2015/3/1

N2 - We consider estimation of the causal effect of a binary treatment on an outcome, conditionally on covariates, from observational studies or natural experiments in which there is a binary instrument for treatment. We describe a doubly robust, locally efficient estimator of the parameters indexing a model for the local average treatment effect conditionally on covariates V when randomization of the instrument is only true conditionally on a high dimensional vector of covariates X, possibly bigger than V. We discuss the surprising result that inference is identical to inference for the parameters of a model for an additive treatment effect on the treated conditionally on V that assumes no treatment-instrument interaction. We illustrate our methods with the estimation of the local average effect of participating in 401(k) retirement programmes on savings by using data from the US Census Bureau's 1991 Survey of Income and Program Participation.

AB - We consider estimation of the causal effect of a binary treatment on an outcome, conditionally on covariates, from observational studies or natural experiments in which there is a binary instrument for treatment. We describe a doubly robust, locally efficient estimator of the parameters indexing a model for the local average treatment effect conditionally on covariates V when randomization of the instrument is only true conditionally on a high dimensional vector of covariates X, possibly bigger than V. We discuss the surprising result that inference is identical to inference for the parameters of a model for an additive treatment effect on the treated conditionally on V that assumes no treatment-instrument interaction. We illustrate our methods with the estimation of the local average effect of participating in 401(k) retirement programmes on savings by using data from the US Census Bureau's 1991 Survey of Income and Program Participation.

KW - Instrumental variables

KW - Local average treatment effect

KW - Local efficiency

KW - Multiplicative effect

UR - http://www.scopus.com/inward/record.url?scp=84961290270&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961290270&partnerID=8YFLogxK

U2 - 10.1111/rssb.12078

DO - 10.1111/rssb.12078

M3 - Article

C2 - 25663814

AN - SCOPUS:84961290270

VL - 77

SP - 373

EP - 396

JO - Journal of the Royal Statistical Society. Series B: Statistical Methodology

JF - Journal of the Royal Statistical Society. Series B: Statistical Methodology

SN - 1369-7412

IS - 2

ER -