Explain what a rank means before publishing the table

A public ranking needs more than a numbered first column. Readers need to know what position means, who was eligible and which choices could change the order. Put the following seven disclosures beside the table or in a clearly linked methodology note. This is an editorial checklist, not a formal scoring standard.

1. What the ranking measures

Write a sentence that names the measured outcome. “Libraries ranked by annual visits” describes a count. “Best libraries” adds a quality judgment that the count alone cannot establish. If the dataset measures usage, keep usage in the title and column label.

For an invented library example, use “Recorded visits in the reporting year.” Do not quietly substitute unique visitors: one person may make more than one visit.

2. Who could appear

State the eligible population and exclusions. A table of participating institutions is not automatically a table of every institution in a country. Give excluded and unavailable entries a visible explanation so readers do not mistake absence for poor performance.

3. Which direction is better

If you assign rank, explain whether higher or lower values receive the leading position. A smaller waiting time and a larger collection size point in opposite directions. If neither is inherently better, label the operation as sorting instead of presenting an overall winner.

4. What happens to ties

Choose a tie policy before writing the headline. In a hypothetical table with values 80, 80 and 70, a shared-first convention might show ranks 1, 1 and 3. A different convention might show 1, 1 and 2. State yours. Do not use alphabetical order to imply that two equal scores differ in quality.

5. What missing values mean

Keep “not reported” visibly separate from zero. If an entry cannot be ranked, mark its rank unavailable and explain the reason. A reader should be able to tell whether the institution reported no activity or supplied no usable observation.

6. How a combined score was built

For a score made from several measures, disclose the inputs, transformations and weights. The OECD/JRC handbook treats weighting, missing data and sensitivity analysis as substantive parts of composite-indicator construction; they are not merely presentation choices. Link the method and avoid presenting a chosen weighting scheme as the only reasonable one. OECD/JRC handbook

7. Which release the ranking describes

Put the observation period and dataset version together. “Published in September” does not say whether the observations concern August or the previous calendar year. When values change, say whether the new order reflects new observations, corrected records or a revised method.

Before sharing the page, ask someone to explain why the first row is first. If they cannot answer from the page and its methodology, the ranking still needs context. A sortable table can expose several useful views without pretending that one order settles every reader’s question.

Put field definitions in a data dictionary and distinguish the observation period from the source and update dates. Those notes support the ranking method without replacing it.