Multilevel Multiple Imputation: A Review and Evaluation of Joint Modeling and Chained Equations Imputation

Craig K. Enders, Stephen A. Mistler, Brian T. Keller

Research output: Contribution to journalArticle

45 Citations (Scopus)

Abstract

Although missing data methods have advanced in recent years, methodologists have devoted less attention to multilevel data structures where observations at level-1 are nested within higher-order organizational units at level-2 (e.g., individuals within neighborhoods; repeated measures nested within individuals; students nested within classrooms). Joint modeling and chained equations imputation are the principal imputation frameworks for single-level data, and both have multilevel counterparts. These approaches differ algorithmically and in their functionality; both are appropriate for simple random intercept analyses with normally distributed data, but they differ beyond that. The purpose of this paper is to describe multilevel imputation strategies and evaluate their performance in a variety of common analysis models. Using multiple imputation theory and computer simulations, we derive 4 major conclusions: (a) joint modeling and chained equations imputation are appropriate for random intercept analyses; (b) the joint model is superior for analyses that posit different within- and between-cluster associations (e.g., a multilevel regression model that includes a level-1 predictor and its cluster means, a multilevel structural equation model with different path values at level-1 and level-2);

Original languageEnglish (US)
JournalPsychological Methods
DOIs
StateAccepted/In press - Dec 21 2015
Externally publishedYes

Fingerprint

Joints
Structural Models
Computer Simulation
Students

Keywords

  • Missing data
  • Multilevel modeling
  • Multiple imputation

ASJC Scopus subject areas

  • Psychology (miscellaneous)

Cite this

Multilevel Multiple Imputation : A Review and Evaluation of Joint Modeling and Chained Equations Imputation. / Enders, Craig K.; Mistler, Stephen A.; Keller, Brian T.

In: Psychological Methods, 21.12.2015.

Research output: Contribution to journalArticle

@article{55e26a14fb984ed3b3970ba2eb43eec0,
title = "Multilevel Multiple Imputation: A Review and Evaluation of Joint Modeling and Chained Equations Imputation",
abstract = "Although missing data methods have advanced in recent years, methodologists have devoted less attention to multilevel data structures where observations at level-1 are nested within higher-order organizational units at level-2 (e.g., individuals within neighborhoods; repeated measures nested within individuals; students nested within classrooms). Joint modeling and chained equations imputation are the principal imputation frameworks for single-level data, and both have multilevel counterparts. These approaches differ algorithmically and in their functionality; both are appropriate for simple random intercept analyses with normally distributed data, but they differ beyond that. The purpose of this paper is to describe multilevel imputation strategies and evaluate their performance in a variety of common analysis models. Using multiple imputation theory and computer simulations, we derive 4 major conclusions: (a) joint modeling and chained equations imputation are appropriate for random intercept analyses; (b) the joint model is superior for analyses that posit different within- and between-cluster associations (e.g., a multilevel regression model that includes a level-1 predictor and its cluster means, a multilevel structural equation model with different path values at level-1 and level-2);",
keywords = "Missing data, Multilevel modeling, Multiple imputation",
author = "Enders, {Craig K.} and Mistler, {Stephen A.} and Keller, {Brian T.}",
year = "2015",
month = "12",
day = "21",
doi = "10.1037/met0000063",
language = "English (US)",
journal = "Psychological Methods",
issn = "1082-989X",
publisher = "American Psychological Association Inc.",

}

TY - JOUR

T1 - Multilevel Multiple Imputation

T2 - A Review and Evaluation of Joint Modeling and Chained Equations Imputation

AU - Enders, Craig K.

AU - Mistler, Stephen A.

AU - Keller, Brian T.

PY - 2015/12/21

Y1 - 2015/12/21

N2 - Although missing data methods have advanced in recent years, methodologists have devoted less attention to multilevel data structures where observations at level-1 are nested within higher-order organizational units at level-2 (e.g., individuals within neighborhoods; repeated measures nested within individuals; students nested within classrooms). Joint modeling and chained equations imputation are the principal imputation frameworks for single-level data, and both have multilevel counterparts. These approaches differ algorithmically and in their functionality; both are appropriate for simple random intercept analyses with normally distributed data, but they differ beyond that. The purpose of this paper is to describe multilevel imputation strategies and evaluate their performance in a variety of common analysis models. Using multiple imputation theory and computer simulations, we derive 4 major conclusions: (a) joint modeling and chained equations imputation are appropriate for random intercept analyses; (b) the joint model is superior for analyses that posit different within- and between-cluster associations (e.g., a multilevel regression model that includes a level-1 predictor and its cluster means, a multilevel structural equation model with different path values at level-1 and level-2);

AB - Although missing data methods have advanced in recent years, methodologists have devoted less attention to multilevel data structures where observations at level-1 are nested within higher-order organizational units at level-2 (e.g., individuals within neighborhoods; repeated measures nested within individuals; students nested within classrooms). Joint modeling and chained equations imputation are the principal imputation frameworks for single-level data, and both have multilevel counterparts. These approaches differ algorithmically and in their functionality; both are appropriate for simple random intercept analyses with normally distributed data, but they differ beyond that. The purpose of this paper is to describe multilevel imputation strategies and evaluate their performance in a variety of common analysis models. Using multiple imputation theory and computer simulations, we derive 4 major conclusions: (a) joint modeling and chained equations imputation are appropriate for random intercept analyses; (b) the joint model is superior for analyses that posit different within- and between-cluster associations (e.g., a multilevel regression model that includes a level-1 predictor and its cluster means, a multilevel structural equation model with different path values at level-1 and level-2);

KW - Missing data

KW - Multilevel modeling

KW - Multiple imputation

UR - http://www.scopus.com/inward/record.url?scp=84951310155&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84951310155&partnerID=8YFLogxK

U2 - 10.1037/met0000063

DO - 10.1037/met0000063

M3 - Article

AN - SCOPUS:84951310155

JO - Psychological Methods

JF - Psychological Methods

SN - 1082-989X

ER -