Some percentage of missing values will always occur naturally. When this happens, the missing data is known as Missing Completely at Random (MCAR). An example of this is if 2% of a survey was not recorded due to a glitch in a survey response system. If this were the case, one could assume that it was not related to any other variable in the data. MCAR variables are also not correlated with its own data values, that is it would also not be related to any other of the non-missing values that appear in the data.