Missing not at Random Imputation Models for Multiple Imputation
by Chained Equation
Description
Provides imputation models and functions for binary or continuous Missing Not At Random (MNAR) outcomes through the use of the 'mice' package. The mice.impute.hecknorm() function provides imputation model for continuous outcome based on Heckman's model also named sample selection model as described in Galimard et al (2018) and Galimard et al (2016) . The mice.impute.heckprob() function provides imputation model for binary outcome based on bivariate probit model as described in Galimard et al (2018).