Tarikul Islam

Identification and Estimation of Causal Estimands with Missing Not at Random Data

Missing not at random (MNAR) data pose significant challenges for causal inference, particularly when both confounders and the outcome are partially observed. Without additional assumptions beyond those required for causal inference, causal estimands are generally not identifiable under MNAR mechanisms. This paper firs...