This repository contains R scripts for the illustration of collider bias in the context of multiple imputation, based on real-world data from the Stockholm Public Health Cohort (SPHC). The applied example demonstrates how including a collider variable in the imputation model can bias effect estimates in longitudinal studies.
Further methodological details are available in the manuscript:
Maya B. Mathur, Willi Zhang, and Ilya Shpitser. Imputation without nightMARs: Graphical criteria for valid imputation of missing data. Preprint available at https://osf.io/preprints/osf/zqne9.