The graph looks convincing, but are the numbers sound?
A student enters plant heights into a spreadsheet and accidentally records 82 cm instead of 8.2 cm. The resulting chart shows an extraordinary growth spurt. Data validation is the process of checking observations and entries before treating them as evidence.
Define expected formats
Dates, units, categories and numerical ranges should be clear. Consistent formats help detect errors.
Check source records
Compare entered values with original observations rather than assuming the spreadsheet is authoritative.
Use reasonableness checks
An extreme value may be an error or a genuine observation. Investigate instead of automatically deleting it.
Look for missing and duplicate entries
Repeated records or blank measurements can distort averages and trends.
Validate calculations
Spreadsheet formulas, references and unit conversions need checking, especially when copied across rows.
Document corrections
When a mistake is confirmed, explain what changed and preserve the original evidence where appropriate.
Break-test: a value is unusual but correct
Can students keep valid evidence while explaining why it differs from other observations?
Connected knowledge
See Research Data Integrity, Statistical Reasoning and MOE Syllabuses.