
Find the original table
Trace the post or screenshot to the document producing the number. Find date, method, definition and corrections. A quote repeated by several accounts is not a new source. If you cannot recover its scope, avoid presenting the number as an established finding.
Keep units and denominators
Specify people, accounts, responses, visits or events. A percentage needs a base: how many cases and among whom? Do not mix total, average and change. Check geography and dates before using a figure to describe another population.
Draw without distorting the message
Choose a readable representation and label the scale. Avoid volume effects and decoration that complicate comparisons. If an axis does not start at zero, make that choice visible and explain it when it may change interpretation. Keep data and calculations so the post can be checked.
Offer a text version
Accompany the chart with a sentence describing the result without exaggeration. Add a table or value link where needed. W3C recommends a complete text equivalent for complex images. Separate what data shows from your explanatory hypothesis and state known uncertainties.
Choose the right check for the situation
These editorial pointers help prepare and review your work. Choose the situation matching your content; check current features and conditions in the cited sources before applying it to a service.
| Situation | Useful action | Mistake to avoid |
|---|---|---|
| Published percentage | State the total, numerator, period and calculation rule. | Showing a rate without explaining how many cases it covers. |
| Comparing series | Check units, definition, population and collection method before juxtaposing. | Comparing indicators whose labels merely look identical. |
| Chart shared as an image | Add an explicit title, source, date and textual equivalent of essential values. | Leaving axes or the source unreadable after phone resizing. |
| Data correction | Date the correction, replace the chart and explain what changes. | Silently changing an already cited image without noting a substantive correction. |
Hypothetical example: apply the method
In a hypothetical questionnaire, a team receives few responses. It publishes respondent count and response categories without presenting the result as representative of the entire community. No example number replaces collected data.
Template to adapt
Copy this outline into your working document. It collects no information on this site; complete it for your situation.
Claim: Original source: Date and territory: Population: Unit and base: Calculation: Limits: Accessible values:
Your pre-use check
0 / 4 checks completed
Frequently asked questions
Is a screenshot enough as a source?
It can preserve a record, but the original document helps check method, date and context.
Should numbers be rounded?
Choose suitable precision and state relevant rounding. Excess precision can create a false impression of certainty.
Continue with related content
Sources and scope
Primary references consulted for factual points. Methods and templates are editorial suggestions; service conditions should still be checked in current documentation.