Compare Public Datasets Before Trusting Their Conclusions
A spreadsheet can look precise while measuring something different from the question you asked.
The aim is to help you complete one clear task, recognize what needs verification, and avoid the most common mistakes. Work through the checks in order, then use the real-world example to apply them to your situation.
This guide is designed as a practical decision checklist. Use the relevant official guidance alongside the steps, especially when rules depend on your country, device or individual circumstances.
1. Define the unit and population
Find what one row represents: a person, household, country, school or month. Check geographic coverage and whether missing groups are excluded. An average across countries does not describe each household.
2. Read the methodology
Identify how the data was gathered, its definitions, sample design and revisions. Two series with the same label may use different years, currency adjustments or inclusion rules.
3. Check missingness and time
Inspect blank values, suppressed cells, collection dates and changes in definitions. Avoid drawing a trend from only two years when the series is irregular or measured with large uncertainty.
4. Reproduce one reported figure
Calculate one simple aggregate or percentage independently from the raw data and compare it with the published summary. Record the source URL, retrieval date and transformation steps in your analysis notes.
Verify the source before acting
For the definitions, current capabilities or rules behind this topic, use the original source below. Features, eligibility, safety requirements and local laws can change. A guide cannot replace a provider's official instructions or qualified advice for a consequential decision.