The work least visible in academic databases is not the least valuable.
It is the work published where nobody indexes — and that distinction determines how entire research systems are perceived by anyone reading the figures.
Where coverage fails
- Local-language publication.
- Regional and national journals.
- Applied and practice-oriented work.
- Books and chapters in some fields.
- Conference proceedings in others.
- Reports, datasets and tools.
Who it affects
- Researchers in systems where local-language publication is normal.
- Applied and policy-facing researchers.
- Fields where books matter more than articles.
- Anyone whose work reaches practitioners rather than international specialists.
- The gaps are systematic, not random — they fall repeatedly on the same people.
What the gaps produce
- Output figures that understate whole systems.
- Comparisons distorted in one direction.
- Individual records that look sparse.
- Assessment that penalises locally essential work.
- Conclusions about research activity that measure indexing rather than activity.
What data sources should do
- State coverage by field and language, plainly.
- Say a sparse record indicates coverage, not activity.
- Include less-indexed sources where possible.
- Allow scholars to add missing outputs.
- Never present a record as complete.
What researchers can do
- Deposit everything in a repository with an identifier — this makes work findable even where it is not indexed.
- Add missing outputs to your own profile.
- Ensure complete metadata on local-language work.
- Cite less-indexed work where relevant — this is how it becomes visible at all.
What users of the data should do
- Ask what is not covered before drawing a conclusion.
- Treat sparse records as uncertain, not as evidence of inactivity.
- Avoid comparisons where coverage differs, or state the caveat prominently.
- Report the limitation with the figure.
One thing worth remembering
Absence in the data is not absence in the world.
Every figure built on indexed output is a lower bound — and it is a lower bound that is systematically lowest for exactly the research systems and researchers whose work is least represented in the indexes.
Câu hỏi thường gặp
Where does coverage fail?
Local-language publication, regional and national journals, applied and practice-oriented work, books and chapters in some fields, conference proceedings in others, and reports and datasets.
Who do coverage gaps affect?
Researchers in systems where local-language publication is normal, applied researchers, fields where books matter, and anyone reaching practitioners — systematically, not randomly.
What do the gaps produce?
Output figures understating whole systems, distorted comparisons, sparse-looking individual records, assessment penalising locally essential work, and conclusions measuring indexing rather than activity.
What can researchers do?
Deposit everything in a repository with an identifier, add missing outputs to their profile, ensure complete metadata on local-language work, and cite less-indexed work.
What should users of the data remember?
That absence in the data is not absence in the world — every figure built on indexed output is a lower bound, and lowest for the systems least represented.
