Multiple imputation for nonignorable missing data

Jongho Im, Soeun Kim

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Multiple imputation is a popular technique for analyzing incomplete data. Missing at random mechanism is often assumed when multiple imputation is performed, assuming that the response mechanism does not depend on the missing variable. However, the assumption of ignorable nonresponse may lead to largely biased estimates when in fact the missingness is nonignorable. In this paper, we propose a multiple imputation method in the presence of nonignorable nonresponse. In the proposed method, we take the selection model approach and specify the response model and the respondents’ outcome model to capture the joint model of the study variable and the response indicator. The proposed data augmentation algorithm uses the respondents’ outcome model and incorporates a semiparametric estimation of the respondents’ outcome model. The proposed multiple imputation method performs well if the specified response model is correct. Limited simulation studies are presented to check the performance of the proposed multiple imputation method.

Original languageEnglish
Pages (from-to)583-592
Number of pages10
JournalJournal of the Korean Statistical Society
Volume46
Issue number4
DOIs
Publication statusPublished - 2017 Dec 1

Fingerprint

Nonignorable Missing Data
Multiple Imputation
Non-response
Data Augmentation
Semiparametric Estimation
Missing at Random
Joint Model
Model
Selection Model
Incomplete Data
Biased
Simulation Study
Estimate

All Science Journal Classification (ASJC) codes

  • Statistics and Probability

Cite this

@article{8a8259d495a040dea814ff4fc6f90128,
title = "Multiple imputation for nonignorable missing data",
abstract = "Multiple imputation is a popular technique for analyzing incomplete data. Missing at random mechanism is often assumed when multiple imputation is performed, assuming that the response mechanism does not depend on the missing variable. However, the assumption of ignorable nonresponse may lead to largely biased estimates when in fact the missingness is nonignorable. In this paper, we propose a multiple imputation method in the presence of nonignorable nonresponse. In the proposed method, we take the selection model approach and specify the response model and the respondents’ outcome model to capture the joint model of the study variable and the response indicator. The proposed data augmentation algorithm uses the respondents’ outcome model and incorporates a semiparametric estimation of the respondents’ outcome model. The proposed multiple imputation method performs well if the specified response model is correct. Limited simulation studies are presented to check the performance of the proposed multiple imputation method.",
author = "Jongho Im and Soeun Kim",
year = "2017",
month = "12",
day = "1",
doi = "10.1016/j.jkss.2017.05.001",
language = "English",
volume = "46",
pages = "583--592",
journal = "Journal of the Korean Statistical Society",
issn = "1226-3192",
publisher = "Korean Statistical Society",
number = "4",

}

Multiple imputation for nonignorable missing data. / Im, Jongho; Kim, Soeun.

In: Journal of the Korean Statistical Society, Vol. 46, No. 4, 01.12.2017, p. 583-592.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Multiple imputation for nonignorable missing data

AU - Im, Jongho

AU - Kim, Soeun

PY - 2017/12/1

Y1 - 2017/12/1

N2 - Multiple imputation is a popular technique for analyzing incomplete data. Missing at random mechanism is often assumed when multiple imputation is performed, assuming that the response mechanism does not depend on the missing variable. However, the assumption of ignorable nonresponse may lead to largely biased estimates when in fact the missingness is nonignorable. In this paper, we propose a multiple imputation method in the presence of nonignorable nonresponse. In the proposed method, we take the selection model approach and specify the response model and the respondents’ outcome model to capture the joint model of the study variable and the response indicator. The proposed data augmentation algorithm uses the respondents’ outcome model and incorporates a semiparametric estimation of the respondents’ outcome model. The proposed multiple imputation method performs well if the specified response model is correct. Limited simulation studies are presented to check the performance of the proposed multiple imputation method.

AB - Multiple imputation is a popular technique for analyzing incomplete data. Missing at random mechanism is often assumed when multiple imputation is performed, assuming that the response mechanism does not depend on the missing variable. However, the assumption of ignorable nonresponse may lead to largely biased estimates when in fact the missingness is nonignorable. In this paper, we propose a multiple imputation method in the presence of nonignorable nonresponse. In the proposed method, we take the selection model approach and specify the response model and the respondents’ outcome model to capture the joint model of the study variable and the response indicator. The proposed data augmentation algorithm uses the respondents’ outcome model and incorporates a semiparametric estimation of the respondents’ outcome model. The proposed multiple imputation method performs well if the specified response model is correct. Limited simulation studies are presented to check the performance of the proposed multiple imputation method.

UR - http://www.scopus.com/inward/record.url?scp=85020476367&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020476367&partnerID=8YFLogxK

U2 - 10.1016/j.jkss.2017.05.001

DO - 10.1016/j.jkss.2017.05.001

M3 - Article

AN - SCOPUS:85020476367

VL - 46

SP - 583

EP - 592

JO - Journal of the Korean Statistical Society

JF - Journal of the Korean Statistical Society

SN - 1226-3192

IS - 4

ER -