Inferences on missing information under multiple imputation and two-stage multiple imputation

摘要: Abstract In the presence of missing values, researchers may be interested in rates information. The information are (a) important for assessing how contributes to inferential uncertainty about, Q , population quantity interest, (b) an component decision number imputations, and (c) can used test model fitting. this article I will derive asymptotic distribution two scenarios: conventional multiple imputation (MI), two-stage MI. Numerically show that proposed agrees with simulated one. also suggest imputations needed obtain reliable rate estimates each method, based on distribution.