Using random imputation to match a variable's distribution>