statistics · cause
Missing not at random
The gaps are not a random sprinkle.
Some data are missing, and the chance that a value is missing depends on the value itself. The observed data are a distorted sample.
People with very high or very low incomes are less likely to answer the income question. What does the remaining data show?
If the blanks were scattered at random, the answers you have would still represent the whole.
When the chance of a blank depends on the income itself, the answers you keep are a biased sample.
Very high and very low incomes skip the question, so the remaining answers sit closer together.
An average calculated from the remaining answers no longer matches the true population.
The reason for the gap is related to the thing you are trying to measure.