Sensitivity analysis for average treatment effects

Sascha O. Becker, Marco Caliendo

Research output: Contribution to journalArticleResearchpeer-review

211 Citations (Scopus)

Abstract

Based on the conditional independence or unconfoundedness assumption, matching has become a popular approach to estimate average treatment effects. Checking the sensitivity of the estimated results with respect to deviations from this identifying assumption has become an increasingly important topic in the applied evaluation literature. If there are unobserved variables that affect assignment into treatment and the outcome variable simultaneously, a hidden bias might arise to which matching estimators are not robust. We address this problem with the bounding approach proposed by Rosenbaum (Observational Studies, 2nd ed., New York: Springer), where mhbounds lets the researcher determine how strongly an unmeasured variable must influence the selection process to undermine the implications of the matching analysis.

Original languageEnglish
Pages (from-to)71-83
Number of pages13
JournalStata Journal
Volume7
Issue number1
DOIs
Publication statusPublished - Mar 2007
Externally publishedYes

Keywords

  • Matching
  • Mhbounds
  • Rosenbaum bounds
  • Sensitivity analysis
  • St0121
  • Treatment effects
  • Unobserved heterogeneity

Cite this

Becker, Sascha O. ; Caliendo, Marco. / Sensitivity analysis for average treatment effects. In: Stata Journal. 2007 ; Vol. 7, No. 1. pp. 71-83.
@article{7b83fb9144f64422a1076d06a9c96754,
title = "Sensitivity analysis for average treatment effects",
abstract = "Based on the conditional independence or unconfoundedness assumption, matching has become a popular approach to estimate average treatment effects. Checking the sensitivity of the estimated results with respect to deviations from this identifying assumption has become an increasingly important topic in the applied evaluation literature. If there are unobserved variables that affect assignment into treatment and the outcome variable simultaneously, a hidden bias might arise to which matching estimators are not robust. We address this problem with the bounding approach proposed by Rosenbaum (Observational Studies, 2nd ed., New York: Springer), where mhbounds lets the researcher determine how strongly an unmeasured variable must influence the selection process to undermine the implications of the matching analysis.",
keywords = "Matching, Mhbounds, Rosenbaum bounds, Sensitivity analysis, St0121, Treatment effects, Unobserved heterogeneity",
author = "Becker, {Sascha O.} and Marco Caliendo",
year = "2007",
month = "3",
doi = "10.1177/1536867X0700700104",
language = "English",
volume = "7",
pages = "71--83",
journal = "Stata Journal",
issn = "1536-867X",
publisher = "StataCorp",
number = "1",

}

Sensitivity analysis for average treatment effects. / Becker, Sascha O.; Caliendo, Marco.

In: Stata Journal, Vol. 7, No. 1, 03.2007, p. 71-83.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Sensitivity analysis for average treatment effects

AU - Becker, Sascha O.

AU - Caliendo, Marco

PY - 2007/3

Y1 - 2007/3

N2 - Based on the conditional independence or unconfoundedness assumption, matching has become a popular approach to estimate average treatment effects. Checking the sensitivity of the estimated results with respect to deviations from this identifying assumption has become an increasingly important topic in the applied evaluation literature. If there are unobserved variables that affect assignment into treatment and the outcome variable simultaneously, a hidden bias might arise to which matching estimators are not robust. We address this problem with the bounding approach proposed by Rosenbaum (Observational Studies, 2nd ed., New York: Springer), where mhbounds lets the researcher determine how strongly an unmeasured variable must influence the selection process to undermine the implications of the matching analysis.

AB - Based on the conditional independence or unconfoundedness assumption, matching has become a popular approach to estimate average treatment effects. Checking the sensitivity of the estimated results with respect to deviations from this identifying assumption has become an increasingly important topic in the applied evaluation literature. If there are unobserved variables that affect assignment into treatment and the outcome variable simultaneously, a hidden bias might arise to which matching estimators are not robust. We address this problem with the bounding approach proposed by Rosenbaum (Observational Studies, 2nd ed., New York: Springer), where mhbounds lets the researcher determine how strongly an unmeasured variable must influence the selection process to undermine the implications of the matching analysis.

KW - Matching

KW - Mhbounds

KW - Rosenbaum bounds

KW - Sensitivity analysis

KW - St0121

KW - Treatment effects

KW - Unobserved heterogeneity

UR - http://www.scopus.com/inward/record.url?scp=36849023898&partnerID=8YFLogxK

U2 - 10.1177/1536867X0700700104

DO - 10.1177/1536867X0700700104

M3 - Article

VL - 7

SP - 71

EP - 83

JO - Stata Journal

JF - Stata Journal

SN - 1536-867X

IS - 1

ER -