Multiple imputation of covariates by substantive-model compatible fully conditional specification

Jonathan W. Bartlett, Tim P. Morris

Research output: Contribution to journalArticlepeer-review

36 Citations (SciVal)

Abstract

Multiple imputation is a practical, principled approach to handling missing data. When used to impute missing values in covariates of regression models, imputation models may be misspecified if they are not compatible with the substantive model of interest for the outcome. In this article, we introduce the smcfcs command, which imputes covariates by substantive-model compatible fully conditional specification. This modifies the popular fully conditional specification or chained-equations approach to multiple imputation by imputing each covariate compatibly with a user-specified substantive model. We compare the smcfcs command with standard fully conditional specification imputation using mi impute chained in a simulation study and illustrative analysis of data from a study investigating time to tumor recurrence in breast cancer.

Original languageEnglish
Article numberst0387
Pages (from-to)437-456
Number of pages20
JournalStata Journal
Volume15
Issue number2
Publication statusPublished - 1 Apr 2015

Keywords

  • Congenial
  • Interactions
  • Multiple imputation
  • Nonlinearities
  • smcfcs
  • st0387
  • Substantive model compatible

ASJC Scopus subject areas

  • Mathematics (miscellaneous)

Fingerprint

Dive into the research topics of 'Multiple imputation of covariates by substantive-model compatible fully conditional specification'. Together they form a unique fingerprint.

Cite this