A fully conditional specification approach to multilevel imputation of categorical and continuous variables

Craig K. Enders, Brian T. Keller, Roy Levy

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

Specialized imputation routines for multilevel data are widely available in software packages, but these methods are generally not equipped to handle a wide range of complexities that are typical of behavioral science data. In particular, existing imputation schemes differ in their ability to handle random slopes, categorical variables, differential relations at Level-1 and Level-2, and incomplete Level-2 variables. Given the limitations of existing imputation tools, the purpose of this manuscript is to describe a flexible imputation approach that can accommodate a diverse set of 2-level analysis problems that includes any of the aforementioned features. The procedure employs a fully conditional specification (also known as chained equations) approach with a latent variable formulation for handling incomplete categorical variables. Computer simulations suggest that the proposed procedure works quite well, with trivial biases in most cases. We provide a software program that implements the imputation strategy, and we use an artificial data set to illustrate its use.

Original languageEnglish (US)
Pages (from-to)298-317
Number of pages20
JournalPsychological Methods
Volume23
Issue number2
DOIs
StatePublished - Jun 1 2018

Fingerprint

Software
Behavioral Sciences
Aptitude
Computer Simulation
Datasets

Keywords

  • Imputation software
  • Missing data
  • Multilevel models
  • Multiple imputation

ASJC Scopus subject areas

  • Psychology (miscellaneous)

Cite this

A fully conditional specification approach to multilevel imputation of categorical and continuous variables. / Enders, Craig K.; Keller, Brian T.; Levy, Roy.

In: Psychological Methods, Vol. 23, No. 2, 01.06.2018, p. 298-317.

Research output: Contribution to journalArticle

@article{e1d3d2aefc30427a83815742b74fa074,
title = "A fully conditional specification approach to multilevel imputation of categorical and continuous variables",
abstract = "Specialized imputation routines for multilevel data are widely available in software packages, but these methods are generally not equipped to handle a wide range of complexities that are typical of behavioral science data. In particular, existing imputation schemes differ in their ability to handle random slopes, categorical variables, differential relations at Level-1 and Level-2, and incomplete Level-2 variables. Given the limitations of existing imputation tools, the purpose of this manuscript is to describe a flexible imputation approach that can accommodate a diverse set of 2-level analysis problems that includes any of the aforementioned features. The procedure employs a fully conditional specification (also known as chained equations) approach with a latent variable formulation for handling incomplete categorical variables. Computer simulations suggest that the proposed procedure works quite well, with trivial biases in most cases. We provide a software program that implements the imputation strategy, and we use an artificial data set to illustrate its use.",
keywords = "Imputation software, Missing data, Multilevel models, Multiple imputation",
author = "Enders, {Craig K.} and Keller, {Brian T.} and Roy Levy",
year = "2018",
month = "6",
day = "1",
doi = "10.1037/met0000148",
language = "English (US)",
volume = "23",
pages = "298--317",
journal = "Psychological Methods",
issn = "1082-989X",
publisher = "American Psychological Association Inc.",
number = "2",

}

TY - JOUR

T1 - A fully conditional specification approach to multilevel imputation of categorical and continuous variables

AU - Enders, Craig K.

AU - Keller, Brian T.

AU - Levy, Roy

PY - 2018/6/1

Y1 - 2018/6/1

N2 - Specialized imputation routines for multilevel data are widely available in software packages, but these methods are generally not equipped to handle a wide range of complexities that are typical of behavioral science data. In particular, existing imputation schemes differ in their ability to handle random slopes, categorical variables, differential relations at Level-1 and Level-2, and incomplete Level-2 variables. Given the limitations of existing imputation tools, the purpose of this manuscript is to describe a flexible imputation approach that can accommodate a diverse set of 2-level analysis problems that includes any of the aforementioned features. The procedure employs a fully conditional specification (also known as chained equations) approach with a latent variable formulation for handling incomplete categorical variables. Computer simulations suggest that the proposed procedure works quite well, with trivial biases in most cases. We provide a software program that implements the imputation strategy, and we use an artificial data set to illustrate its use.

AB - Specialized imputation routines for multilevel data are widely available in software packages, but these methods are generally not equipped to handle a wide range of complexities that are typical of behavioral science data. In particular, existing imputation schemes differ in their ability to handle random slopes, categorical variables, differential relations at Level-1 and Level-2, and incomplete Level-2 variables. Given the limitations of existing imputation tools, the purpose of this manuscript is to describe a flexible imputation approach that can accommodate a diverse set of 2-level analysis problems that includes any of the aforementioned features. The procedure employs a fully conditional specification (also known as chained equations) approach with a latent variable formulation for handling incomplete categorical variables. Computer simulations suggest that the proposed procedure works quite well, with trivial biases in most cases. We provide a software program that implements the imputation strategy, and we use an artificial data set to illustrate its use.

KW - Imputation software

KW - Missing data

KW - Multilevel models

KW - Multiple imputation

UR - http://www.scopus.com/inward/record.url?scp=85047658453&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047658453&partnerID=8YFLogxK

U2 - 10.1037/met0000148

DO - 10.1037/met0000148

M3 - Article

VL - 23

SP - 298

EP - 317

JO - Psychological Methods

JF - Psychological Methods

SN - 1082-989X

IS - 2

ER -