Grey literature searching is the deliberate hunt for relevant evidence that was never published in indexed journals, including trials registries, conference abstracts, theses, government and organisational reports, and preprints. Searching it matters because studies with unwelcome or null results are less likely to be formally published, so a review that ignores grey literature risks an overly optimistic picture of the evidence.
The bias grey literature is there to counter
The reason grey literature searching exists is publication bias: positive, statistically significant results are published more often and more quickly than null findings. If a review draws only on the published record, its synthesis can overstate an effect. Searching unpublished and hard-to-find sources is one of the main defences against that distortion, alongside formal checks like a funnel plot during analysis. The two work together: grey literature reduces the bias at the search stage, and the funnel plot helps detect what remains.
Where grey literature lives
Trials registries
Trials registries record studies at the point they begin, which exposes trials that were completed but never written up, and others still in progress. Checking registries also lets you spot outcome reporting differences between what a trial planned and what it published, which feeds directly into risk of bias assessment.
Conference abstracts and proceedings
A large share of research is first, and sometimes only, presented at conferences. Many conference abstracts never become full papers, so searching proceedings, often via databases such as Embase that index them, surfaces studies the journal literature misses.
Theses, reports, and preprints
Doctoral theses, government reports, non-governmental organisation publications, and preprint servers all hold evidence that never entered a bibliographic index. The relevant sources depend heavily on the field, which is why grey literature targets are chosen as part of the review protocol rather than improvised later.
How to search sources that resist structured queries
Grey literature rarely supports the precise Boolean strategies you use in bibliographic databases. Many sites offer only basic keyword boxes, so you adapt: use a small set of core terms, browse by topic or organisation, and search the websites of bodies known to publish in your area. Because these searches are less reproducible, careful note-taking matters even more here than in database searching.
A practical grey literature checklist
Because grey sources are scattered, it helps to work from a fixed list of source types and decide, for each, whether it is relevant to your question before you start. A workable default for a health review covers:
- Trials registries for prospectively registered and completed-but-unpublished studies, checked for the planned-versus-reported outcome differences that feed publication and reporting bias checks.
- Conference proceedings, often reachable through a database such as Embase that indexes abstracts the journal record omits.
- Dissertation and thesis repositories for doctoral work that was never turned into a paper.
- Government, agency, and charity reports from the bodies that fund or publish in your area, searched on their own websites.
- Preprint servers for the newest findings not yet through peer review, flagged clearly as preprints in the record.
Two stopping rules keep this from becoming open-ended. Set the source list in the protocol so you are not adding sources to chase a result, and stop searching a browse-based source once further looking returns nothing new, the same saturation logic used in iterative citation snowballing.
How grey literature reaches the unpublished evidence
The value of grey literature is not just volume but the specific bias it offsets. A registry entry can reveal a completed trial whose null result was never written up, and comparing a trial’s registered primary outcome with what it eventually published exposes selective outcome reporting that no amount of database searching would surface. This is why a review that pools results in a meta-analysis leans on grey literature so heavily: an effect estimate built only on the published record can be inflated, and the grey search is the first line of defence before any statistical check is even run.
Keeping a defensible grey literature record
For every grey source you record the platform, the exact terms or browse path, the date searched, and what you found, exactly as you would for a database, following documenting and reporting your search. Records found this way still pass through title and abstract screening against your eligibility criteria, and are counted into the PRISMA flow diagram under the appropriate source category so the audit trail stays complete.
Common grey literature mistakes
The recurring failures are not about effort but about discipline. The first is skipping grey literature entirely and relying on a funnel-plot check to reassure the reader, when the better defence is to find the missing studies in the first place. The second is searching grey sources but not recording them, which makes that part of the search unreproducible and invites criticism when the review is appraised. The third is treating a preprint as a peer-reviewed study without labelling it, which mixes evidence of different standing. Avoiding all three keeps the grey search both useful and defensible, and it is exactly the standard our literature search service works to.