True blanks

• The question was optional, did not apply, or was skipped.
• The field was added after the older rows were created.
• A value failed to load or was cleared, and nothing recorded why.

• Software disagrees about blanks. One tool reads an empty cell as empty text, another as a missing value, another as zero.
• Averages quietly leave blank rows out, so the result describes only the people who answered.
• Filters and joins on a column with blanks drop rows that nobody meant to exclude.

• Count empty cells in each column, and count cells that hold only spaces separately.
• Check whether blanks cluster by date, source, or another column. A pattern usually points to the cause.
• Compare the blank rate between columns that should be filled together.

• Make every blank the same kind of empty, so that spaces, empty text, and missing values are not three different things.
• Find out why the column has blanks before deciding what to do. The right treatment for not applicable differs from the one for unknown.
• Leave blanks blank unless there is a stated reason to fill them. A filled value looks like a real one from then on.
• If values are estimated, mark which ones in a separate column.

• Disguised missing valuesA cell is not empty, but what it holds means there is no value: N/A, a dash, 999, or the word none.
• Partially filled rowsA row has some of its fields and lacks others, so it is usable for some purposes and not for others.
• Defaults mistaken for dataA value that looks real was filled in automatically by a form or a system, and nobody actually chose it.

TermsMissing value, Null, Empty string, Completeness