Evaluation of Four Multiple Imputation Methods for Handling Missing Binary Outcome Data in the Presence of an Interaction between a Dummy and a Continuous Variable

<div>Comparison of the performance of imputation methods for real dataset based on the average percent bias (PB) for the interaction effect. MAR, missing at random; MCAR, missing completely at random; MCR, multiple imputation by chained equation using classification and regression trees; MRF, multiple imputation by chained equation using random forests; MI, MICE-Interaction method; MS, MICE-Stratified method.</div>

Journal of Probability and Statistics

fig2

Figure 2

Figure 2: Evaluation of Four Multiple Imputation Methods for Handling Missing Binary Outcome Data in the Presence of an Interaction between a Dummy and a Continuous Variable 

Figure 2 | Evaluation of Four Multiple Imputation Methods for Handling Missing Binary Outcome Data in the Presence of an Interaction between a Dummy and a Continuous Variable 