Asymptotic bias in the linear mixed effects model under non-ignorable missing data mechanisms

Chandan Saha, Michael P. Jones

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

In longitudinal studies, missingness of data is often an unavoidable problem. Estimators from the linear mixed effects model assume that missing data are missing at random. However, estimators are biased when this assumption is not met. In the paper, theoretical results for the asymptotic bias are established under non-ignorable drop-out, drop-in and other missing data patterns. The asymptotic bias is large when the drop-out subjects have only one or no observation, especially for slope-related parameters of the linear mixed effects model. In the drop-in case, intercept-related parameter estimators show substantial asymptotic bias when subjects enter late in the study. Eight other missing data patterns are considered and these produce asymptotic biases of a variety of magnitudes.

Original languageEnglish
Pages (from-to)167-182
Number of pages16
JournalJournal of the Royal Statistical Society. Series B: Statistical Methodology
Volume67
Issue number1
DOIs
StatePublished - 2005

Fingerprint

Nonignorable Missing Data
Missing Data Mechanism
Linear Mixed Effects Model
Asymptotic Bias
Missing Data
Drop out
Estimator
Missing at Random
Longitudinal Study
Intercept
Biased
Slope
Asymptotic bias
Missing data

Keywords

  • Clinical trial
  • Incomplete data
  • Informative drop-out
  • Longitudinal study
  • Repeated measurements

ASJC Scopus subject areas

  • Mathematics(all)
  • Statistics and Probability

Cite this

@article{d6613fb97f49480caa95c25a5bab9a6d,
title = "Asymptotic bias in the linear mixed effects model under non-ignorable missing data mechanisms",
abstract = "In longitudinal studies, missingness of data is often an unavoidable problem. Estimators from the linear mixed effects model assume that missing data are missing at random. However, estimators are biased when this assumption is not met. In the paper, theoretical results for the asymptotic bias are established under non-ignorable drop-out, drop-in and other missing data patterns. The asymptotic bias is large when the drop-out subjects have only one or no observation, especially for slope-related parameters of the linear mixed effects model. In the drop-in case, intercept-related parameter estimators show substantial asymptotic bias when subjects enter late in the study. Eight other missing data patterns are considered and these produce asymptotic biases of a variety of magnitudes.",
keywords = "Clinical trial, Incomplete data, Informative drop-out, Longitudinal study, Repeated measurements",
author = "Chandan Saha and Jones, {Michael P.}",
year = "2005",
doi = "10.1111/j.1467-9868.2005.00494.x",
language = "English",
volume = "67",
pages = "167--182",
journal = "Journal of the Royal Statistical Society. Series B: Statistical Methodology",
issn = "1369-7412",
publisher = "Wiley-Blackwell",
number = "1",

}

TY - JOUR

T1 - Asymptotic bias in the linear mixed effects model under non-ignorable missing data mechanisms

AU - Saha, Chandan

AU - Jones, Michael P.

PY - 2005

Y1 - 2005

N2 - In longitudinal studies, missingness of data is often an unavoidable problem. Estimators from the linear mixed effects model assume that missing data are missing at random. However, estimators are biased when this assumption is not met. In the paper, theoretical results for the asymptotic bias are established under non-ignorable drop-out, drop-in and other missing data patterns. The asymptotic bias is large when the drop-out subjects have only one or no observation, especially for slope-related parameters of the linear mixed effects model. In the drop-in case, intercept-related parameter estimators show substantial asymptotic bias when subjects enter late in the study. Eight other missing data patterns are considered and these produce asymptotic biases of a variety of magnitudes.

AB - In longitudinal studies, missingness of data is often an unavoidable problem. Estimators from the linear mixed effects model assume that missing data are missing at random. However, estimators are biased when this assumption is not met. In the paper, theoretical results for the asymptotic bias are established under non-ignorable drop-out, drop-in and other missing data patterns. The asymptotic bias is large when the drop-out subjects have only one or no observation, especially for slope-related parameters of the linear mixed effects model. In the drop-in case, intercept-related parameter estimators show substantial asymptotic bias when subjects enter late in the study. Eight other missing data patterns are considered and these produce asymptotic biases of a variety of magnitudes.

KW - Clinical trial

KW - Incomplete data

KW - Informative drop-out

KW - Longitudinal study

KW - Repeated measurements

UR - http://www.scopus.com/inward/record.url?scp=12844260796&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=12844260796&partnerID=8YFLogxK

U2 - 10.1111/j.1467-9868.2005.00494.x

DO - 10.1111/j.1467-9868.2005.00494.x

M3 - Article

VL - 67

SP - 167

EP - 182

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 - 1

ER -