A Cautionary Note for Plasmode Simulation Studies in the Setting of Causal Inference

This presentation discusses the importance of plasmode simulations in evaluating statistical estimators. It highlights various strategies for generating plasmode data and emphasizes the risks of not simulating treatment or missingness variables from a known conditional distribution.

It was presented at the Eastern North American Region (ENAR) International Biometric Society Spring 2025 Meeting.