Doubly robust multiple imputation using kernel-based techniques

Chiu-Hsieh Hsu, Yulei He, Yisheng Li, Qi Long, Randall S Friese

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We consider the problem of estimating the marginal mean of an incompletely observed variable and develop a multiple imputation approach. Using fully observed predictors, we first establish two working models: one predicts the missing outcome variable, and the other predicts the probability of missingness. The predictive scores from the two models are used to measure the similarity between the incomplete and observed cases. Based on the predictive scores, we construct a set of kernel weights for the observed cases, with higher weights indicating more similarity. Missing data are imputed by sampling from the observed cases with probability proportional to their kernel weights. The proposed approach can produce reasonable estimates for the marginal mean and has a double robustness property, provided that one of the two working models is correctly specified. It also shows some robustness against misspecification of both models. We demonstrate these patterns in a simulation study. In a real-data example, we analyze the total helicopter response time from injury in the Arizona emergency medical service data.

Original languageEnglish (US)
JournalBiometrical Journal
DOIs
StateAccepted/In press - 2015

Fingerprint

Multiple Imputation
kernel
Weights and Measures
Double Robustness
Aircraft
Emergency Medical Services
Predict
Reaction Time
Misspecification
Helicopter
Missing Data
Emergency
Model
Response Time
Predictors
Directly proportional
Wounds and Injuries
Simulation Study
Robustness
Kernel

Keywords

  • Bandwidth
  • Bootstrap
  • Local imputation
  • Model misspecification
  • Nonparametric

ASJC Scopus subject areas

  • Statistics and Probability
  • Medicine(all)
  • Statistics, Probability and Uncertainty

Cite this

Doubly robust multiple imputation using kernel-based techniques. / Hsu, Chiu-Hsieh; He, Yulei; Li, Yisheng; Long, Qi; Friese, Randall S.

In: Biometrical Journal, 2015.

Research output: Contribution to journalArticle

@article{6bee027a2958407fb37df1fd606e9025,
title = "Doubly robust multiple imputation using kernel-based techniques",
abstract = "We consider the problem of estimating the marginal mean of an incompletely observed variable and develop a multiple imputation approach. Using fully observed predictors, we first establish two working models: one predicts the missing outcome variable, and the other predicts the probability of missingness. The predictive scores from the two models are used to measure the similarity between the incomplete and observed cases. Based on the predictive scores, we construct a set of kernel weights for the observed cases, with higher weights indicating more similarity. Missing data are imputed by sampling from the observed cases with probability proportional to their kernel weights. The proposed approach can produce reasonable estimates for the marginal mean and has a double robustness property, provided that one of the two working models is correctly specified. It also shows some robustness against misspecification of both models. We demonstrate these patterns in a simulation study. In a real-data example, we analyze the total helicopter response time from injury in the Arizona emergency medical service data.",
keywords = "Bandwidth, Bootstrap, Local imputation, Model misspecification, Nonparametric",
author = "Chiu-Hsieh Hsu and Yulei He and Yisheng Li and Qi Long and Friese, {Randall S}",
year = "2015",
doi = "10.1002/bimj.201400256",
language = "English (US)",
journal = "Biometrical Journal",
issn = "0323-3847",
publisher = "Wiley-VCH Verlag",

}

TY - JOUR

T1 - Doubly robust multiple imputation using kernel-based techniques

AU - Hsu, Chiu-Hsieh

AU - He, Yulei

AU - Li, Yisheng

AU - Long, Qi

AU - Friese, Randall S

PY - 2015

Y1 - 2015

N2 - We consider the problem of estimating the marginal mean of an incompletely observed variable and develop a multiple imputation approach. Using fully observed predictors, we first establish two working models: one predicts the missing outcome variable, and the other predicts the probability of missingness. The predictive scores from the two models are used to measure the similarity between the incomplete and observed cases. Based on the predictive scores, we construct a set of kernel weights for the observed cases, with higher weights indicating more similarity. Missing data are imputed by sampling from the observed cases with probability proportional to their kernel weights. The proposed approach can produce reasonable estimates for the marginal mean and has a double robustness property, provided that one of the two working models is correctly specified. It also shows some robustness against misspecification of both models. We demonstrate these patterns in a simulation study. In a real-data example, we analyze the total helicopter response time from injury in the Arizona emergency medical service data.

AB - We consider the problem of estimating the marginal mean of an incompletely observed variable and develop a multiple imputation approach. Using fully observed predictors, we first establish two working models: one predicts the missing outcome variable, and the other predicts the probability of missingness. The predictive scores from the two models are used to measure the similarity between the incomplete and observed cases. Based on the predictive scores, we construct a set of kernel weights for the observed cases, with higher weights indicating more similarity. Missing data are imputed by sampling from the observed cases with probability proportional to their kernel weights. The proposed approach can produce reasonable estimates for the marginal mean and has a double robustness property, provided that one of the two working models is correctly specified. It also shows some robustness against misspecification of both models. We demonstrate these patterns in a simulation study. In a real-data example, we analyze the total helicopter response time from injury in the Arizona emergency medical service data.

KW - Bandwidth

KW - Bootstrap

KW - Local imputation

KW - Model misspecification

KW - Nonparametric

UR - http://www.scopus.com/inward/record.url?scp=84949870661&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84949870661&partnerID=8YFLogxK

U2 - 10.1002/bimj.201400256

DO - 10.1002/bimj.201400256

M3 - Article

C2 - 26647734

AN - SCOPUS:84949870661

JO - Biometrical Journal

JF - Biometrical Journal

SN - 0323-3847

ER -