What Absence Means

People are not missing from datasets at random. They are missing because of how the data was collected, by whom, and who was considered worth counting at the time.

This matters everywhere and it matters especially for the questions this platform exists to help people ask. Historical records systematically undercount the people the recording institution served least well. That is not a flaw in your analysis; it is a fact about the source, and it belongs in what you publish.

Three specific traps.

A category that did not exist when the data was collected produces a gap, not a zero. A count of something nobody was recording is absence of evidence.

Small groups aggregated into 'other' disappear from every analysis downstream, and no amount of careful work later recovers them.

A zero and a blank are different. Zero means measured and none. Blank means not measured, or not reported, or refused. Treating them alike changes every average you compute.

✦ Naming the gap makes the finding stronger

There is a fear that stating limits undermines the work. The opposite is true. A reader who notices a limit you did not mention stops trusting everything else you wrote. A reader who sees you name it first trusts the rest more.