A mixed-effects estimating equation approach to nonignorable missing longitudinal data with Refreshment samples

Xuan Bi, Annie Qu

Research output: Contribution to journalArticle

Abstract

Nonignorable missing data occur frequently in longitudinal studies and can cause biased estimations. Refreshment samples which draw new subjects randomly in subsequent waves from the original population could mitigate the bias. In this paper, we introduce a mixed-effects estimating equation approach that enables one to incorporate refreshment samples and recover informative missing information from the measurement process. We show that the proposed method achieves consistency and asymptotic normality for fixed-effect estimation under shared-parameter models, and we extend it to a more general nonignorable-missing framework. Our finite sample simulation studies show the effectiveness and robustness of the proposed method under different missing mechanisms. In addition, we apply our method to election poll longitudinal survey data with refreshment samples from the 2007-2008 Associated Press-Yahoo! News.

LanguageEnglish (US)
Pages1653-1675
Number of pages23
JournalStatistica Sinica
Volume28
Issue number4
DOIs
StatePublished - Oct 1 2018

Fingerprint

Mixed Effects
Estimating Equation
Longitudinal Data
Missing Data
Nonignorable Missing Data
Biased Estimation
Fixed Effects
Longitudinal Study
Survey Data
Elections
Asymptotic Normality
Simulation Study
Robustness
Longitudinal data

Keywords

  • Missing not at random
  • Non-monotone missing pattern
  • Quadratic inference function
  • Shared-parameter model
  • Survey data

Cite this

A mixed-effects estimating equation approach to nonignorable missing longitudinal data with Refreshment samples. / Bi, Xuan; Qu, Annie.

In: Statistica Sinica, Vol. 28, No. 4, 01.10.2018, p. 1653-1675.

Research output: Contribution to journalArticle

@article{0ad52dd47c4b4aa58efe9f8c98db4e3c,
title = "A mixed-effects estimating equation approach to nonignorable missing longitudinal data with Refreshment samples",
abstract = "Nonignorable missing data occur frequently in longitudinal studies and can cause biased estimations. Refreshment samples which draw new subjects randomly in subsequent waves from the original population could mitigate the bias. In this paper, we introduce a mixed-effects estimating equation approach that enables one to incorporate refreshment samples and recover informative missing information from the measurement process. We show that the proposed method achieves consistency and asymptotic normality for fixed-effect estimation under shared-parameter models, and we extend it to a more general nonignorable-missing framework. Our finite sample simulation studies show the effectiveness and robustness of the proposed method under different missing mechanisms. In addition, we apply our method to election poll longitudinal survey data with refreshment samples from the 2007-2008 Associated Press-Yahoo! News.",
keywords = "Missing not at random, Non-monotone missing pattern, Quadratic inference function, Shared-parameter model, Survey data",
author = "Xuan Bi and Annie Qu",
year = "2018",
month = "10",
day = "1",
doi = "10.5705/ss.202015.0317",
language = "English (US)",
volume = "28",
pages = "1653--1675",
journal = "Statistica Sinica",
issn = "1017-0405",
publisher = "Institute of Statistical Science",
number = "4",

}

TY - JOUR

T1 - A mixed-effects estimating equation approach to nonignorable missing longitudinal data with Refreshment samples

AU - Bi, Xuan

AU - Qu, Annie

PY - 2018/10/1

Y1 - 2018/10/1

N2 - Nonignorable missing data occur frequently in longitudinal studies and can cause biased estimations. Refreshment samples which draw new subjects randomly in subsequent waves from the original population could mitigate the bias. In this paper, we introduce a mixed-effects estimating equation approach that enables one to incorporate refreshment samples and recover informative missing information from the measurement process. We show that the proposed method achieves consistency and asymptotic normality for fixed-effect estimation under shared-parameter models, and we extend it to a more general nonignorable-missing framework. Our finite sample simulation studies show the effectiveness and robustness of the proposed method under different missing mechanisms. In addition, we apply our method to election poll longitudinal survey data with refreshment samples from the 2007-2008 Associated Press-Yahoo! News.

AB - Nonignorable missing data occur frequently in longitudinal studies and can cause biased estimations. Refreshment samples which draw new subjects randomly in subsequent waves from the original population could mitigate the bias. In this paper, we introduce a mixed-effects estimating equation approach that enables one to incorporate refreshment samples and recover informative missing information from the measurement process. We show that the proposed method achieves consistency and asymptotic normality for fixed-effect estimation under shared-parameter models, and we extend it to a more general nonignorable-missing framework. Our finite sample simulation studies show the effectiveness and robustness of the proposed method under different missing mechanisms. In addition, we apply our method to election poll longitudinal survey data with refreshment samples from the 2007-2008 Associated Press-Yahoo! News.

KW - Missing not at random

KW - Non-monotone missing pattern

KW - Quadratic inference function

KW - Shared-parameter model

KW - Survey data

UR - http://www.scopus.com/inward/record.url?scp=85054484423&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054484423&partnerID=8YFLogxK

U2 - 10.5705/ss.202015.0317

DO - 10.5705/ss.202015.0317

M3 - Article

VL - 28

SP - 1653

EP - 1675

JO - Statistica Sinica

T2 - Statistica Sinica

JF - Statistica Sinica

SN - 1017-0405

IS - 4

ER -