The CRediT adoption ledger at CASRAI tracks publisher-level CRediT support. For this audit we drilled in at the journal level for the top-100-cited journals across major disciplines, looking at three questions: does the journal collect CRediT structurally at submission, does it collect the lead/equal/supporting qualifier, and does it carry CRediT into the published JATS XML with fidelity? The picture in mid-2026 is meaningfully better than two years ago, with substantial uneven spots that the community can still address.
Methodology
We selected the top-100 journals by citation impact across STEM, biomedical, social sciences, and humanities (drawn from a mid-2025 cross-disciplinary snapshot). For each, we inspected the author-instruction pages, the submission system’s structured-data entry, a small sample of recent published articles’ JATS XML (via Crossref REST API where available), and the journal’s stated CRediT policy. Where the journal’s submission was not directly inspectable we drew on the publisher’s documented policy.
The audit is descriptive, not exhaustive. A journal may have changed its policy since our snapshot; spot-checking has confirmed broad accuracy but specific claims about individual journals should be verified against current author instructions.
The top-line numbers
Of the top-100 audited, 78 collect CRediT structurally at submission; 14 request it as prose only; 8 do not request it.
Of the 78 that collect structurally, 52 collect the lead/equal/supporting qualifier; 26 collect the roles without qualifier.
Of the 78 that collect structurally, 61 carry CRediT through to the published JATS with role URIs and (where collected) qualifier values; 17 appear to drop the structured data at the publication step, producing prose CRediT statements without machine-readable equivalents.
The publisher patterns
The big publishers have diverged in their patterns.
Elsevier: the bulk of Elsevier journals collect structured CRediT with qualifier and carry it to JATS. There are exceptions — typically older or society-published Elsevier-hosted journals — but the Elsevier baseline is strong.
Springer Nature: variable. The Nature-branded journals all collect structured CRediT with qualifier. Other Springer Nature journals are at different stages of submission-system upgrade; some collect prose only.
Wiley: most major Wiley journals collect structured CRediT but the qualifier collection is patchier. The published JATS carries the structured data where collected.
Taylor & Francis: structured CRediT collection is in place for major journals; the qualifier and JATS fidelity are weaker than at other major publishers.
Cell Press: full structured CRediT with qualifier and complete JATS fidelity across the journal portfolio.
PLOS, eLife, F1000Research: full structured CRediT with qualifier and open JATS, often with the CRediT data exposed via the publisher’s API.
OUP and CUP: structured CRediT broadly adopted; qualifier and JATS fidelity uneven by journal.
Society publishers: highly variable. Major society journals operating their own submission systems range from full structured collection to absent. Society journals on commercial publisher platforms inherit the publisher’s pattern.
The qualifier collection gap
The single largest pattern of unevenness is the qualifier collection. 26 of the 78 structured-CRediT-collecting journals do not collect the qualifier. The result is published CRediT statements that assign roles without indicating Lead, Equal, or Supporting; this is the gap that the 2022 NISO Z39.104 standard was meant to address and that the qualifier was added to resolve.
The fix is straightforward at the publisher-platform level: the submission system needs a field for the qualifier alongside the role assignment. The Editorial Manager and ScholarOne platforms both support the qualifier field; whether it is enabled is a per-journal configuration.
The CASRAI CRediT adoption ledger is being extended in v2026.3 to track qualifier-collection as a separate adoption dimension. Publishers and journal editors who want to close the gap should enable the qualifier field in their submission system.
The JATS-fidelity gap
The second pattern of unevenness is the JATS-fidelity drop. 17 of the 78 structured-CRediT-collecting journals appear to drop the structured data at the publication step. The published article carries a prose CRediT statement; the structured data that was collected at submission does not appear in the JATS or the Crossref deposit.
This is wasted effort. The structured data was collected; it should travel through to the publication. The fix is at the publication-workflow step: the JATS generation must read the structured CRediT fields and emit them as <role> elements with the appropriate URI and qualifier. The CASRAI CRediT JATS integration patterns document the correct emission.
For authors, the practical implication is that the structured data they entered may not appear in the publication; the prose statement is the public-facing record. This is a quality issue the publisher should address.
The prose-only journals
The 14 journals that collect CRediT as prose only are mostly society journals on older submission platforms or specialised journals that have not upgraded their workflow. The transition to structured collection is straightforward but requires platform investment.
For authors submitting to these journals, the practical advice is to enter the CRediT statement in the prose format that the journal expects but to mirror it in a structured form (e.g., a CSV or JSON file with role and qualifier per contributor) for the institutional record and for ORCID deposit. The prose statement that appears in the publication may be less informative than the structured form the authors have.
The no-CRediT journals
The 8 journals not requesting CRediT are concentrated in disciplines where CRediT has been slower to adopt (some humanities journals, some specialist law journals, some specialist economics journals) and in journals that have not updated their author instructions in recent years. The community-led push for adoption has slowed in the past 18 months as the easy adopters have done so; the remaining holdouts require sustained advocacy.
What to do
For publishers, three priorities. First, enable qualifier collection where it is supported but not active. Second, fix the JATS-fidelity drop where structured data is being lost. Third, complete the structured-CRediT collection rollout for the journals that are still prose-only.
For journal editors, the practical step is to verify your journal’s CRediT collection workflow and the JATS output for a recent article. If the structured data is being lost, raise it with the publisher. If the qualifier is not being collected, request it.
For authors, the practical step is to mirror your CRediT statement in a structured form that is independent of the publisher’s workflow, and to verify the JATS output post-publication. This is a small overhead that ensures your contributorship record is complete regardless of publisher workflow gaps.
For the broader community, the audit is encouraging but unfinished. The 78% structured-collection rate and the 65% JATS-fidelity rate are both meaningful improvements over the 2022 baseline; the path to 95%+ on both is operationally clear.
