Although single imputation is widely used, it does not reflect the uncertainty created by missing data at random. So, multiple imputation is more favorable then single imputation in case of data missing at random.
