• Global imputation based on a missing attribute
If we look at the other values taken by a variable with a missing item, and we find that some of them are more
frequent than others, then we may decide to use this fact to assign the most frequent value to a missing one. We fill
the missing attribute values with a measure of central tendency (see [9–11] for a comparison of mean imputation
and the most common imputation attribute value). These methods are usually regarded as inadequate, because the
standard deviation of the sample is underestimated even when data are MCAR.