Reply to the comment on ‘Working with population totals in the presence of missing data comparing imputation methods in terms of bias and precision’ by Bogaart et al.