Passive imputation and parcel summaries are both valid to handle missing items in studies with many multi-item scales

Iris Eekhout, Henrica C.W. de Vet, Michiel R. de Boer, Jos W.R. Twisk, Martijn W. Heymans

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

Previous studies showed that missing data in multi-item scales can best be handled by multiple imputation of item scores. However, when many scales are used, the number of items will become too large for the imputation model to reliably estimate imputations. A solution is to use passive imputation or a parcel summary score that combine and consequently reduce the number of variables in the imputation model. The performance of these methods was evaluated in a simulation study and illustrated in an example. Passive imputation, which updated scale scores from imputed items, and parcel summary scores that use the average over available item scores were compared to using all items simultaneously, imputing total scores of scales and complete-case analysis. Scale scores and coefficient estimates from linear regression were compared to “true” parameters on bias and precision. Passive imputation and using parcel summaries showed smaller bias and more precision than imputing total scores and complete-case analyses. Passive imputation or using parcel summary scores are valid missing data solutions in studies that include many multi-item scales.

Original languageEnglish
Pages (from-to)1128-1140
Number of pages13
JournalStatistical Methods in Medical Research
Volume27
Issue number4
Early online date22 Jun 2016
DOIs
Publication statusPublished - Apr 2018

Fingerprint

Imputation
Valid
Linear Models
Missing Data
Coefficient Estimates
Multiple Imputation
Linear regression
Simulation Study

Keywords

  • item imputation
  • missing data
  • Multiple imputation
  • questionnaires
  • simulation study

Cite this

Eekhout, Iris ; de Vet, Henrica C.W. ; de Boer, Michiel R. ; Twisk, Jos W.R. ; Heymans, Martijn W. / Passive imputation and parcel summaries are both valid to handle missing items in studies with many multi-item scales. In: Statistical Methods in Medical Research. 2018 ; Vol. 27, No. 4. pp. 1128-1140.
@article{255a66f109424a1fb5a21dba8e10d876,
title = "Passive imputation and parcel summaries are both valid to handle missing items in studies with many multi-item scales",
abstract = "Previous studies showed that missing data in multi-item scales can best be handled by multiple imputation of item scores. However, when many scales are used, the number of items will become too large for the imputation model to reliably estimate imputations. A solution is to use passive imputation or a parcel summary score that combine and consequently reduce the number of variables in the imputation model. The performance of these methods was evaluated in a simulation study and illustrated in an example. Passive imputation, which updated scale scores from imputed items, and parcel summary scores that use the average over available item scores were compared to using all items simultaneously, imputing total scores of scales and complete-case analysis. Scale scores and coefficient estimates from linear regression were compared to “true” parameters on bias and precision. Passive imputation and using parcel summaries showed smaller bias and more precision than imputing total scores and complete-case analyses. Passive imputation or using parcel summary scores are valid missing data solutions in studies that include many multi-item scales.",
keywords = "item imputation, missing data, Multiple imputation, questionnaires, simulation study",
author = "Iris Eekhout and {de Vet}, {Henrica C.W.} and {de Boer}, {Michiel R.} and Twisk, {Jos W.R.} and Heymans, {Martijn W.}",
note = "{\circledC} The Author(s) 2016.",
year = "2018",
month = "4",
doi = "10.1177/0962280216654511",
language = "English",
volume = "27",
pages = "1128--1140",
journal = "Stat Methods Med Res",
issn = "0962-2802",
publisher = "SAGE Publications",
number = "4",

}

Passive imputation and parcel summaries are both valid to handle missing items in studies with many multi-item scales. / Eekhout, Iris; de Vet, Henrica C.W.; de Boer, Michiel R.; Twisk, Jos W.R.; Heymans, Martijn W.

In: Statistical Methods in Medical Research, Vol. 27, No. 4, 04.2018, p. 1128-1140.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - Passive imputation and parcel summaries are both valid to handle missing items in studies with many multi-item scales

AU - Eekhout, Iris

AU - de Vet, Henrica C.W.

AU - de Boer, Michiel R.

AU - Twisk, Jos W.R.

AU - Heymans, Martijn W.

N1 - © The Author(s) 2016.

PY - 2018/4

Y1 - 2018/4

N2 - Previous studies showed that missing data in multi-item scales can best be handled by multiple imputation of item scores. However, when many scales are used, the number of items will become too large for the imputation model to reliably estimate imputations. A solution is to use passive imputation or a parcel summary score that combine and consequently reduce the number of variables in the imputation model. The performance of these methods was evaluated in a simulation study and illustrated in an example. Passive imputation, which updated scale scores from imputed items, and parcel summary scores that use the average over available item scores were compared to using all items simultaneously, imputing total scores of scales and complete-case analysis. Scale scores and coefficient estimates from linear regression were compared to “true” parameters on bias and precision. Passive imputation and using parcel summaries showed smaller bias and more precision than imputing total scores and complete-case analyses. Passive imputation or using parcel summary scores are valid missing data solutions in studies that include many multi-item scales.

AB - Previous studies showed that missing data in multi-item scales can best be handled by multiple imputation of item scores. However, when many scales are used, the number of items will become too large for the imputation model to reliably estimate imputations. A solution is to use passive imputation or a parcel summary score that combine and consequently reduce the number of variables in the imputation model. The performance of these methods was evaluated in a simulation study and illustrated in an example. Passive imputation, which updated scale scores from imputed items, and parcel summary scores that use the average over available item scores were compared to using all items simultaneously, imputing total scores of scales and complete-case analysis. Scale scores and coefficient estimates from linear regression were compared to “true” parameters on bias and precision. Passive imputation and using parcel summaries showed smaller bias and more precision than imputing total scores and complete-case analyses. Passive imputation or using parcel summary scores are valid missing data solutions in studies that include many multi-item scales.

KW - item imputation

KW - missing data

KW - Multiple imputation

KW - questionnaires

KW - simulation study

UR - http://www.scopus.com/inward/record.url?scp=85042849758&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042849758&partnerID=8YFLogxK

U2 - 10.1177/0962280216654511

DO - 10.1177/0962280216654511

M3 - Article

VL - 27

SP - 1128

EP - 1140

JO - Stat Methods Med Res

JF - Stat Methods Med Res

SN - 0962-2802

IS - 4

ER -