Five data quality checks to run before you report

Data quality problems rarely announce themselves. They show up as a number that looks plausible in a report and falls apart in an audit. These five checks catch most of them.

1. Check the totals against the source

Add up the raw records and compare with the figure in the report. This sounds obvious, but copy-paste errors between an Excel tracker and a Word report are one of the most common sources of inaccurate reporting.

2. Look for duplicates

Duplicate beneficiary records inflate results. Check for repeated identifiers, identical names with slightly different spellings, and records entered twice on the same day.

3. Test the disaggregation

If your report disaggregates by sex, age or location, confirm that the parts equal the whole. A total that does not match its components usually means a record is missing a category.

4. Trace one record end to end

Pick a random record and follow it from the collection tool to the database to the report. You are checking that the definition was applied consistently and that no transformation changed the meaning.

5. Compare against the previous period

A sudden jump or drop is a question, not necessarily an error. Ask what changed: a new site, a new data collector, a revised definition. Documenting the answer is part of good data quality.

Make it routine

Run these checks at the end of every quarter and keep a short note of what you found and fixed. That note becomes your data quality trail, and it makes any formal assessment much easier.

For a fuller process, including site visit checklists and a scoring workbook, see our [Data Quality Assessment Toolkits](/product/data-quality-assessment-toolkits).