How should variable selection be performed with multiply imputed data?

454 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

MICE: Multivariate Imputation by Chained Equations in R (2011, background): “Several strategies have been proposed that count the number of times that variable is in the model (Brand 1999; Heymans et al. 2007; Wood et al. 2008).”
Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies: The CHARMS Checklist (2014, background): “These adverse effects are related to the amount of missing data [112] and the extent to which data are missing completely at random [108,111].”
Multiple imputation using chained equations: Issues and guidance for practice (2011, methodology): “An alternative [43] would be to repeat the model building (step 2) if new imputations are drawn at step 3.”
Multiple imputation of covariates by fully conditional specification: Accommodating the substantive model (2012, unknown): “Performing model selection in combination with MI is a challenging problem generally, 24 and therefore deciding on what substantive model to specify when using the SMC-FCS algorithm is not a trivial process.”