Figure 3
From: Imputation strategies for missing binary outcomes in cluster randomized trials
Treatment effect estimated from random-effects logistic regression when 30% data is covariate dependent missing.
Figure 3 | BMC Medical Research Methodology