Accounting for missing data in statistical analyses: multiple imputation is not always the answer

2019783 citationsJournal Articlehybrid Open Access

Authors

Rachael A. Hughes · University of Bristol

Jonathan A C Sterne · University of Bristol

Kate Tilling · NIHR Bristol Biomedical Research Centre

Abstract

Choice of method for dealing with missing data is crucial for validity of conclusions, and should be based on careful consideration of the reasons for the missing data, missing data patterns and the availability of auxiliary information.

Topics & Keywords

Statistical Methods and Bayesian Inference Advanced Causal Inference Techniques Statistical Methods and Inference

UN Sustainable Development Goals

Good health and well-being

Publication Details

Published in: International Journal of Epidemiology

Volume 48, Issue 4, pp. 1294-1304

DOI: 10.1093/ije/dyz032

Field-Weighted Citation Impact: 80.05