How to use this calculator
- Enter or choose rows (records).
- Enter or choose columns (variables).
- Enter or choose missing cells.
- Enter or choose rows with at least one missing value.
- The missing cells updates as you type. Use Copy result or Share to keep it.
Method
Missing % = Missing cells ÷ (Rows × Columns) × 100; Complete rows = Rows − Incomplete rows
- Rows × Columns
- total cells
- Incomplete rows
- rows with any missing value
Worked example
- Total cells = 1,000 × 20 = 20,000.
- Missing = 500 ÷ 20,000 = 2.5%.
- Complete rows = 1,000 − 120 = 880, so listwise deletion drops 12% of rows.
Only 2.5% of cells are missing, but a complete-case analysis would lose 12% of the rows.
What the result means
The share of rows lost matters more than the share of cells. If it is large, or if missingness may depend on the values themselves, consider multiple imputation instead of dropping rows.
Assumptions and limits
- Counts come from the same data set.
- Percentages do not tell you why data are missing (completely at random, at random or not at random), which decides the right fix.
Questions
What percentage of missing data is acceptable?
There is no universal cut-off. Under 5% is often treated as minor, but bias depends on why values are missing, not just how many.
Why can’t the missing cells exceed rows × columns in incomplete rows?
Every missing cell sits in an incomplete row, and each such row has at most one cell per column.