Crossref schema 5.5

Background

Crossref's deposit schema is the canonical metadata interchange format for journal articles, conference proceedings, books, and a long tail of scholarly outputs. Schema 5.5, proposed in 2023 and generally available in 2024, added native support for CRediT contributor roles inside the existing contributors element. The change was announced and consulted on the Crossref blog post on proposed schema changes. The current schema reference lives on the Crossref schema documentation page.

The practical effect is that, for the first time, a single deposit can express not just who authored a paper but what each of them did. Downstream consumers — funders, institutional CRIS systems, science-of-science researchers, the ORCID registry — read those roles directly from Crossref without scraping publisher pages.

Where CRediT lives in a deposit

Inside the person_name element, each contributor may carry zero or more role children. Each role names a CRediT URI in the casrai.org/credit namespace. The vocabulary is referenced via the vocabulary and vocabulary-identifier attributes the same way as in JATS.

Crossref schema 5.5 deposit, contributors block

xml

<contributors>
  <person_name sequence="first" contributor_role="author">
    <given_name>San</given_name>
    <surname>Zhang</surname>
    <ORCID authenticated="true">https://orcid.org/0000-0001-2345-6789</ORCID>
    <affiliations>
      <institution>
        <institution_name>University of Example</institution_name>
        <institution_id type="ror">https://ror.org/04abcd123</institution_id>
      </institution>
    </affiliations>
    <role
        vocabulary="credit"
        vocabulary-identifier="https://casrai.org/credit"
        term="Conceptualization"
        term-identifier="https://casrai.org/credit/roles/conceptualization"/>
    <role
        vocabulary="credit"
        vocabulary-identifier="https://casrai.org/credit"
        term="Software"
        term-identifier="https://casrai.org/credit/roles/software"
        specific-use="lead"/>
    <role
        vocabulary="credit"
        vocabulary-identifier="https://casrai.org/credit"
        term="Writing - original draft"
        term-identifier="https://casrai.org/credit/roles/writing-original-draft"
        specific-use="supporting"/>
  </person_name>

  <person_name sequence="additional" contributor_role="author">
    <given_name>Mei</given_name>
    <surname>Liu</surname>
    <ORCID authenticated="true">https://orcid.org/0000-0002-3456-7890</ORCID>
    <role
        vocabulary="credit"
        vocabulary-identifier="https://casrai.org/credit"
        term="Investigation"
        term-identifier="https://casrai.org/credit/roles/investigation"
        specific-use="lead"/>
    <role
        vocabulary="credit"
        vocabulary-identifier="https://casrai.org/credit"
        term="Formal analysis"
        term-identifier="https://casrai.org/credit/roles/formal-analysis"/>
  </person_name>
</contributors>

The attribute conventions mirror those used in JATS 1.2+: a fixed vocabulary="credit", the root URL of the vocabulary, the human-readable role name, and the canonical per-role URI. The optional specific-use attribute carries the degree of contribution: lead, equal, or supporting.

Reading roles back out

Once a deposit has been ingested, Crossref's REST API exposes the roles on the /works/{doi} endpoint as a role array on each contributor. The shape is straightforward and stable.

GET /works/{doi}

bash

curl -sS "https://api.crossref.org/works/10.1234/example.5678" | jq '.message.author[] | {given, family, ORCID, role: .role}'

Trimmed REST API response

json

{
  "given":  "San",
  "family": "Zhang",
  "ORCID":  "http://orcid.org/0000-0001-2345-6789",
  "role": [
    { "name": "Conceptualization", "id": "https://casrai.org/credit/roles/conceptualization" },
    { "name": "Software",          "id": "https://casrai.org/credit/roles/software", "specific-use": "lead" }
  ]
}

The same data is available in bulk via Crossref's snapshot products and surfaces in Crossref Event Data, which means downstream analytics platforms can aggregate role-tagged contributions across the corpus without bespoke publisher integrations.

Deposit checklist

Validate the deposit against the schema 5.5 XSD before submitting; older XSDs reject the role element.
Use the canonical casrai.org/credit URIs, not the deprecated CASRAI-era ones.
Submit one CRediT role per role element; do not concatenate.
Include the ORCID iD on the contributor so the roles propagate into the researcher's profile via ORCID auto-update.
If the contribution degree is known and was collected at submission, populate specific-use; otherwise omit it.

When to use Crossref vs DataCite

Crossref is for narrative scholarly works — journal articles, book chapters, conference papers, preprints. For datasets and software, deposit instead at DataCite using the closely related contributorType vocabulary; the two are designed to interoperate. The 2024 joint guide from Crossref and DataCite documents the cross-walk.