State of CRediT adoption — 2026

A taxonomy by adoption layer

“Publisher X supports CRediT” turns out to be one of the most ambiguous claims in scholarly publishing. The CASRAI implementation scorecard, finalised this year and published at /for-publishers/scorecard, breaks it into five distinct things a publisher has to do. Strictly speaking, a publisher only “supports CRediT” if all five are present:

1. Structured capture at submission. Authors pick roles from a controlled vocabulary; the system stores them per-author per-role.
2. JATS XML emission with canonical NISO URIs. The published article's XML has <role vocab="credit" vocab-term-identifier="https://casrai.org/credit/..."> on every contributor.
3. Crossref deposit including the role metadata. Schema 5.5+ <contributor_role> elements deposited with the DOI.
4. ORCID propagation. Where authors provide ORCID iDs, the publisher pushes the contribution onto each ORCID record.
5. Acknowledged-contributor support. Non-author contributors (medical writers, technicians) get CRediT-style attribution somehow — author guidance, structured acknowledgements, or workflow extension.

On a 500-journal sample drawn from the 2026 Hosseini et al. dataset plus CASRAI manual review, the distribution looks like this:

• ~85% have policy: their author guidelines mention CRediT.
• ~70% have structured capture.
• ~50% have JATS emission with NISO URIs.
• ~30% have Crossref deposit with CRediT metadata.
• ~20% have ORCID propagation.
• <5% have acknowledged-contributor support.

The structured-vs-narrative gap

The widest single fault line in CRediT implementation is the difference between capturing roles in structured formand printing a narrative paragraph at the end of the manuscript. The latter is what most journal author guidelines mean by “include a CRediT statement” — a free-text paragraph that reads “A.S. conducted the experiments and analysed the data; B.D. supervised the project and revised the manuscript” and so on.

Free-text paragraphs are unreadable for machines. They cannot reliably round-trip into JATS XML, cannot deposit cleanly to Crossref with the per-role structure intact, cannot propagate to ORCID records, and cannot be ingested by institutional CRIS systems for evaluation and recognition purposes. They satisfy the policy box but defeat the standard's value proposition.

Where CRediT works as the standard intends, it's because the publisher captures the data in structured form at submission and threads it through the pipeline. That is, where the publisher is on Editorial Manager or ScholarOne, has enabled the structured CRediT module per-journal, and has the JATS-XML output wired up. PLOS, eLife, the Nature portfolio, Cell Press, MDPI, and the major Wiley + Elsevier titles all do this. Many society-published journals on older or lighter editorial platforms do not.

Where the data goes next

For CRediT data captured at submission to be useful for evaluation, hiring, narrative-CV building, or science-of-science research, it must propagate to (1) Crossref, (2) ORCID, and (3) institutional CRIS systems. Crossref Schema 5.5+ has supported CRediT in contributor metadata since 2024; most major publishers' deposit pipelines now emit it.

ORCID propagation is the weakest of the three. ORCID Auto-Update pulls citations from Crossref into ORCID records automatically, but the per-role attribution doesn't always flow with them. Where it does, it shows up under each publication on the author's ORCID record — “Investigation (lead), Writing — original draft (equal)” for that specific paper. This is the long-term value to the researcher: a defensible, verifiable, machine-readable contribution record built over time.

CRIS ingestion is more uneven. Pure (Elsevier), Symplectic Elements, VIVO and Worktribe all support CRediT as a custom vocabulary, but per-institution configuration varies. The CASRAI Dictionary v2026.1 bulk-download endpoints at /dictionary/download publish the controlled vocabulary in six formats specifically so CRIS administrators don't have to roll their own.

The acknowledged-contributor problem

CRediT was originally designed for the named-author byline. In practice, much of the actual research work — bench experiments, regulatory submissions, statistical consulting, technical writing, software development, data cleaning — is done by people who don't reach the ICMJE Vancouver authorship bar. They're thanked in the acknowledgements section. Their contribution is invisible to every downstream system: Crossref, ORCID, CRIS, narrative-CV builders.

Four years into the formal standard, this is the most-cited unresolved limitation. The Standing Committee's 2025 review surfaced it explicitly; the question is whether an extension that admits acknowledged contributors will meaningfully change publisher behaviour — given that <5% of publishers even ship Tier 1 (structured author CRediT) completely.

The funder mandate landscape

No major research funder yet mandates CRediT statements at the grant application stage. UKRI's R4RI narrative CV (mandatory since Jan 2024) implicitly rewards CRediT-aware publication records; Wellcome similarly. NIH hasn't formally mandated CRediT but the new 2024 Public Access Policy creates an opportunity to require it for PMC-deposited articles, which the NIH-funded community produces in bulk.

The interesting fault line is institutional rather than funder. As institutions adopt narrative-CV formats internally (REF 2029 in the UK, ERA in Australia, university tenure review processes in the US), the question of what evidence supports a contribution claim becomes operationally significant. CRediT data harvested from publications and propagated through CRIS systems is the most credible answer.

What CASRAI recommends for 2026 → 2027

1. Publishers: complete the implementation depth at JATS + Crossref + ORCID — the value is in the pipeline, not the policy.
2. Institutions: ingest CRediT into the CRIS via the CASRAI bulk-download endpoints; surface per-role contribution data in researcher profiles.
3. Funders: align grant-reporting templates with CRediT vocabulary; require CRediT statements on publications resulting from funded work.
4. ORCID: prioritise CRediT-aware citation ingestion across all publisher feeds.
5. Researchers: ask your editor and your publisher whether your paper's CRediT roles propagated to Crossref and ORCID. Where they didn't, flag it.

Reference + further reading

• Hosseini, M. et al. (2026). “Publisher CRediT adoption: a structured vs narrative-implementation review.” Learned Publishing.
• ANSI/NISO Z39.104-2022 — Contributor Roles Taxonomy.
• CASRAI publisher implementation scorecard
• CRediT adoption landscape — the live data page
• Bulk-download the CASRAI Dictionary in 7 formats