CASRAI Dictionary

Tag: credit taxonomy authorship

  • Credit Taxonomy Authorship: A Case for Funder Adoption in Grant Reporting

    Opinion: grant reporting should require structured credit taxonomy authorship data alongside biosketches and final reports. Funders currently reward the named author list, not the research team that actually produced the work — and the CRediT roles already used by publishers are the readiest tool to fix that gap. This is a CASRAI perspective, not a report of confirmed funder policy: no major funder currently mandates it.

    The Contributor Role Taxonomy (CRediT) is a standardised set of 14 roles — from conceptualisation and data curation to funding acquisition and writing — used to describe who did what on a research output, distinct from the narrower question of who qualifies as an “author”. CASRAI originated CRediT in 2014; the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022, and it is licensed CC-BY 4.0 for free reuse by anyone, including funders.

    What is the CRediT taxonomy, and why does grant reporting ignore it?

    CRediT is not an authorship test. It does not decide who qualifies as an author under criteria such as those set out by the International Committee of Medical Journal Editors (ICMJE); it describes contribution type once a research output exists. Publishers including Elsevier, Wiley and Taylor & Francis now require a CRediT statement at submission, mapping each named author to one or more of the 14 roles.

    Grant reporting sits entirely outside this system. A funder’s final report typically lists a principal investigator, co-investigators, and a project narrative — not a structured breakdown of who curated the data, who wrote the software, or who administered the project day to day. That gap matters because grant reports, not journal articles, are where funders form their view of “who delivered this award”.

    The case for funder-required credit taxonomy authorship data

    Three arguments support requiring CRediT-style data in grant reporting, not just at publication.

    • Credit for non-PI staff. Research software engineers, data managers, and postdoctoral researchers frequently deliver the technical core of a funded project without ever becoming a named co-investigator on the award. A contributor-role field in the final report creates an auditable record of that work, independent of authorship politics on any resulting paper.
    • Better evidence for funders’ own decisions. Funders assess renewal applications, track record, and “who can actually deliver” partly from CVs and biosketches. A structured role history — built cumulatively across a researcher’s funded outputs — is a more reliable signal than author position, which varies wildly by discipline and negotiation.
    • Continuity with ORCID. ORCID has supported CRediT role tagging on individual “Works” records since 2019. Extending the same structured field to the grant-reporting stage would let a researcher’s contributor history accumulate consistently across both outputs and awards, rather than resetting at each reporting boundary.

    None of this requires funders to redefine authorship. It only requires them to capture, at the reporting stage, data that publishers already collect at the publication stage.

    The administrative-burden counter-argument

    The strongest objection is not conceptual, it is operational. Grant reporting is already a compliance burden for research offices, and adding another structured field is not free.

    • Duplication risk. If contributor roles are recorded once at reporting and again at publication, teams will re-key the same information twice unless the two systems are linked via ORCID or a shared identifier.
    • Multi-institutional friction. Large consortium awards, common in Horizon Europe and UKRI-funded collaborations, involve dozens of contributors across institutions with different research-information systems; agreeing roles before a report deadline adds negotiation overhead.
    • Taxonomy fit. The 14 CRediT roles were designed for journal-article contributions. Some categories of grant-funded work — public engagement, infrastructure maintenance, cohort recruitment — map awkwardly onto the existing role list without local adaptation.

    These are real costs, not reasons to abandon the idea. They are reasons to pilot it narrowly and design the reporting field so it can be pre-populated from existing ORCID or publication CRediT data rather than entered from scratch.

    How grant reporting compares with today’s publisher practice

    The asymmetry between publication-stage and award-stage contributorship data is the core of the argument. It also happens to be an information gap most coverage of CRediT does not spell out.

    Stage / stakeholder Structured contributor-role data required today? Mechanism, where it exists
    Major journal publishers (Elsevier, Wiley, Taylor & Francis) Yes, at submission CRediT author statement mapping each author to one or more of 14 roles
    Grant final/interim reports (typical funder templates) No Narrative project summary and named investigator list only
    NIH biosketch No structured field Free-text “Contributions to Science” section
    ORCID “Works” record Optional, researcher-populated CRediT role tags supported since 2019
    This proposal (CASRAI perspective) Argued position, not existing policy A CRediT-derived contributor-role block appended to funder reports, pre-populated from ORCID where possible

    Answer-first questions on CRediT and author contributions

    What is funding acquisition in author contribution?

    Funding acquisition is one of CRediT’s 14 defined roles, covering acquisition of the financial support for the project that led to the published output. It is the single CRediT role most directly relevant to grant reporting, since it explicitly separates the person who secured the award from those who executed the research — a distinction current biosketch narratives rarely make clean.

    What are the criteria for author contribution?

    Under ICMJE criteria, authorship requires substantial contribution to the work’s conception or design (or data acquisition, analysis, or interpretation), drafting or critically revising the manuscript, final approval of the published version, and agreement to be accountable for it. CRediT does not replace these criteria; it sits alongside them to describe contribution type once authorship has already been determined.

    What are examples of author contributions?

    Typical CRediT-defined contributions include conceptualisation, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, software, supervision, validation, visualisation, and the two writing roles — original draft, and review and editing. A single individual can hold several roles on one output.

    Implications for funders and institutions

    If funders moved toward requesting credit taxonomy authorship data in grant reports, research offices would need three things before a mandate could work in practice: an ORCID-linked pre-population mechanism to avoid double entry, a pilot cohort limited to a small number of funding calls, and explicit guidance that CRediT roles describe contribution, not authorship eligibility, so institutions do not over-interpret the data during promotion or tenure review.

    The honest case for funder adoption is incremental, not sweeping: pilot it on a subset of awards, link it to ORCID so it is populated once and reused, and treat early results as evidence rather than assuming the benefit before it is tested. Given that publishers already run this system at scale, the marginal cost of extending it one stage earlier, into grant reporting, is smaller than building a comparable structure from nothing.

  • CRediT Contribution Taxonomy: The Humanities Gap

    The CRediT contribution taxonomy is a 14-role vocabulary built at a 2012 biomedical-sciences workshop, and three of its roles — Investigation, Software and Resources — describe laboratory research so specifically that they routinely fail to capture what happens in archival, ethnographic or purely theoretical scholarship. That mismatch is a design artefact of CRediT’s origin, not a flaw researchers should paper over by force-fitting their work into the nearest lab-shaped box.

    The credit contribution taxonomy is best understood as a controlled vocabulary of contributor roles, not a universal grammar of scholarly labour. CASRAI originated the CRediT contributor role taxonomy in 2014; the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022. Understanding where that STEM-derived vocabulary strains against humanities and social science (HSS) practice helps journals, university presses and research offices apply it honestly rather than awkwardly.

    CRediT is a controlled, 14-role vocabulary for describing individual contributions to a research output, developed to replace ambiguous author-order conventions with discrete, attributable roles.

    What is the CRediT taxonomy and where did it come from?

    CRediT emerged from a 2012 workshop convened by the Wellcome Trust and Harvard University, bringing together biomedical scientists, publishers and funders to fix a specific problem: author-order lists that concealed who actually did what on a laboratory paper. CASRAI took over stewardship in 2014 and formalised the 14-role vocabulary in 2015.

    In 2022, CRediT was formally adopted as ANSI/NISO Z39.104-2022, with machine-readable metadata built for JATS XML manuscript pipelines. The roles were never designed with archival, ethnographic or purely theoretical research workflows in the room — a gap that was structural from the outset, not an oversight that later revisions quietly fixed.

    Which CRediT roles map poorly onto humanities and social science work?

    Three roles carry the clearest fingerprints of their laboratory origin. Each assumes a mode of working — bench experiments, code, physical materials — that has no direct equivalent in much archival, ethnographic or theoretical scholarship.

    • Investigation is defined as “performing the experiments, or data/evidence collection” — language built for wet-lab or fieldwork protocols. An archival historian spending eighteen months in a single repository, or a philosopher building an argument from primary texts, is doing investigative labour that this wording does not naturally describe.
    • Software assumes programming and code as a discrete, separable contribution. Much qualitative and theoretical scholarship has no computational layer at all, so the role sits permanently empty on the contributor statement — not because no comparable labour occurred, but because the taxonomy has no slot for it.
    • Resources lists “reagents, materials, patients, laboratory samples, animals, instrumentation” — a checklist with no analogue for archival access negotiated with a rights holder, oral-history interview subjects recruited over years, or a rare manuscript collection consulted under restricted access.

    The table below maps each role’s STEM-native definition against the closest HSS reality it is asked to cover.

    CRediT role STEM-native definition HSS scholarship it is asked to cover
    Investigation Performing experiments or data/evidence collection Archival research, ethnographic fieldwork, oral history, close textual analysis
    Software Programming, code, computational tools No equivalent in most theoretical or literary scholarship
    Resources Reagents, samples, instrumentation, materials Archival access, informant recruitment, rare-collection consultation

    What does the evidence say about CRediT outside STEM?

    The mismatch is documented, not merely anecdotal. A 2025 study published in Accountability in Research examined the contributor role taxonomy’s use in library and information science journals and found the existing 14 roles were not a comfortable fit for social-science-style contributions. Vasilevsky et al. (2021), also in Accountability in Research, argued that authorship alone is insufficient for collaborative research and called for contributor-role systems to be extended beyond their original scope.

    Matarese and Shashok, writing in Publications (2019), found that CRediT’s categories can be too coarse even within the biomedical contexts it was built for, prompting proposals for revision. A separate study of a psychology research project found that independent raters classifying the same contributions showed low agreement on both the number and type of roles involved — evidence that the taxonomy’s boundaries are harder to apply consistently than its clean 14-item list suggests.

    The International Committee of Medical Journal Editors (ICMJE) has separately noted that documenting contributions with CRediT or any similar scheme “leaves unresolved the question of the quantity and quality of contribution that qualify an individual for authorship” — a caveat that applies with equal force to HSS disciplines, where sole authorship and non-hierarchical intellectual debt are already harder to parcel into discrete roles.

    How can journals and institutions adapt CRediT for HSS scholarship?

    Adapting CRediT for archival, ethnographic or theoretical work does not require abandoning it. It requires using it honestly rather than stretching its STEM vocabulary to breaking point.

    1. Leave roles blank rather than force-fitting them. CRediT does not require every role to be filled for every output; an empty Software field on a monograph chapter is accurate, not a gap to be papered over.
    2. Pair CRediT with a free-text supplementary statement for contributions the 14 roles do not describe — archival negotiation, translation, fieldwork access-brokering — rather than mislabelling them as “Investigation” or “Resources” for the sake of completing the form.
    3. Treat single-authored HSS works as a distinct case, where the contributor/author distinction that CRediT was built to clarify may simply not apply, rather than applying it cosmetically.
    4. Track discipline-specific extension proposals emerging from library and information science and other social-science-adjacent fields, several of which have proposed additional or renamed roles rather than a wholesale replacement taxonomy.

    Answer-first Q&A on CRediT and contributor roles

    What is the CRediT taxonomy?

    The CRediT taxonomy is a standardised, 14-role controlled vocabulary for describing individual contributions to a scholarly research output, used instead of, or alongside, traditional author-order bylines. It was originated by CASRAI in 2014 and is now formalised as ANSI/NISO Z39.104-2022, with each role carrying a unique, machine-readable identifier.

    What are the 14 roles of CRediT taxonomy?

    The 14 roles are Conceptualization, Data Curation, Formal Analysis, Funding Acquisition, Investigation, Methodology, Project Administration, Resources, Software, Supervision, Validation, Visualization, Writing – Original Draft, and Writing – Review & Editing. They are organised without hierarchy, and contributors may hold multiple roles on a single output.

    What does Investigation mean in CRediT taxonomy?

    Investigation is officially defined as “conducting a research and investigation process, specifically performing the experiments, or data/evidence collection.” That phrasing centres experimental and fieldwork-style data gathering, which is why archival research, close reading and theoretical argument-building sit awkwardly inside a role written for laboratory or survey-based evidence collection.

    How do I CRediT someone in a research paper?

    Authors typically complete a CRediT statement at submission, assigning each named contributor one or more of the 14 roles, optionally with a degree qualifier (“lead,” “equal” or “supporting”). For humanities and social science submissions where roles do not cleanly apply, the more transparent approach is to leave inapplicable roles unfilled and add a brief supplementary note rather than mislabel contributions to complete the form.

    Implications for research administrators and publishers

    For research offices and publishers serving mixed STEM/HSS portfolios, the practical implication is that a single CRediT template cannot be applied uniformly across disciplines without editorial guidance. Journals in library science, digital humanities and area studies have already begun documenting where the taxonomy strains, and that evidence base — not a wholesale rejection of contributor-role systems — is the right foundation for discipline-sensitive guidance.

    The taxonomy’s own governance structure supports this kind of refinement: NISO’s ANSI/NISO Z39.104-2022 standard is maintained through open, community-based revision, meaning discipline-specific extension proposals have a legitimate path forward rather than requiring a competing standard. Institutions adopting CRediT contributor roles for mixed-discipline outputs, and those documenting broader authorship practice, should treat the STEM origin of these 14 roles as a known constraint to design around, not a hidden defect to discover after the fact.

  • CRediT Taxonomy at PLOS ONE: Mandatory Roles

    PLOS ONE does not accept a free-text paragraph of author contributions. Since adopting the CRediT taxonomy, the journal requires every author to be assigned one or more of 14 standardised, machine-readable contributor roles at submission, and those role tags are published with the article. This structured, mandatory model sits in contrast to journals that still rely on a narrative “author contributions” statement, and it is why PLOS ONE is now a reference case for what machine-readable authorship metadata looks like in practice.

    The credit taxonomy plos one implementation is one of the clearest examples of a publisher moving contributor-role reporting from prose to structured data. CRediT (Contributor Roles Taxonomy) is a fixed set of 14 role labels — such as Conceptualization, Data Curation, Formal Analysis and Writing – Original Draft Preparation — used to tag, rather than narrate, what each named author actually did on a research output.

    What is the CRediT taxonomy?

    CASRAI originated the CRediT contributor role taxonomy in 2014 as a way to replace vague authorship credit with a fixed, shared vocabulary. The standard is now stewarded by NISO as ANSI/NISO Z39.104-2022, which formally defines the 14 roles and their scope.

    Each role describes a discrete type of research labour — not seniority, not authorship order, and not a value judgement on contribution size. A single author can hold several roles; a single role can be shared by several authors. The taxonomy is designed to be tagged against a person, ideally via an ORCID iD, so that contribution data can be indexed, aggregated and machine-read rather than only read as prose.

    • Conceptualization
    • Data Curation
    • Formal Analysis
    • Funding Acquisition
    • Investigation
    • Methodology
    • Project Administration
    • Resources
    • Software
    • Supervision
    • Validation
    • Visualization
    • Writing – Original Draft Preparation
    • Writing – Review & Editing

    How did PLOS ONE make CRediT mandatory and machine-readable?

    PLOS states plainly on its authorship policy page that it “has adopted the CRediT Taxonomy to describe each author’s individual contributions to the work,” and that the submitting author is responsible for entering every author’s contributions at the point of submission. This is not an optional supplementary note — it is a required submission field, checked before peer review can proceed.

    Because the roles are selected from a closed list rather than typed freely, the resulting metadata is structured at source. PLOS publishes the completed role set with the final article as tagged data, which downstream systems, indexers and bibliometric researchers can parse without needing to interpret prose. PLOS pairs this with a mandatory ORCID iD for the corresponding author, linking machine-readable roles to a persistent researcher identifier rather than a name string alone.

    This mandatory-and-structured model is precisely what distinguishes PLOS ONE’s approach from journals that reference CRediT only as a recommended framework for a free-text “author contributions” paragraph.

    CRediT vs free-text contribution statements: what changes?

    Free-text contribution statements ask authors to describe their roles in a sentence or short paragraph, with no controlled vocabulary. The result is legible to a human reader but effectively opaque to software, and inconsistent from one journal — even one article — to the next.

    Feature PLOS ONE: mandatory CRediT tagging Free-text contribution statement
    Vocabulary Closed set of 14 defined roles (ANSI/NISO Z39.104-2022) Open, author-written prose
    Machine readability Structured, taggable per author Requires manual or NLP interpretation
    Consistency across articles Uniform role labels journal-wide Wording varies by author and article
    Submission requirement Mandatory field at Editorial Manager submission Often optional or loosely enforced
    Bibliometric usability Enables large-scale contribution analysis Poorly suited to aggregation

    The practical effect is that a mandatory, tagged taxonomy turns “who did what” into queryable data, while a free-text statement remains a one-off narrative disclosure that satisfies transparency norms without generating reusable metadata.

    What does the evidence show about CRediT data in practice?

    Because PLOS ONE’s CRediT tags are structured and published at scale, they have become a dataset in their own right. Ding et al. (2021), writing in Scientometrics, used PLOS ONE’s tagged contributor roles to build and evaluate a co-author credit-allocation method — work that would not have been possible against free-text statements alone.

    Separately, Larivière and colleagues analysed division of labour in biomedical research using CRediT-tagged data, a study now cited more than 135 times, underscoring how structured role data has become a recognised input for research-on-research and responsible-assessment work. Nature’s 2025 retrospective on the taxonomy’s first decade likewise frames CRediT’s core value as enabling “trust, integrity and responsible research assessment” — a claim that depends on contribution data being structured enough to analyse, not merely readable.

    • CASRAI originated CRediT in 2014; NISO now stewards it as ANSI/NISO Z39.104-2022.
    • PLOS requires CRediT role assignment as a mandatory submission field, not an optional note.
    • Ding et al. (2021, Scientometrics) built a credit-allocation model directly from PLOS ONE’s CRediT tags.
    • Larivière et al.’s CRediT-based division-of-labour study has been cited over 135 times.

    Answer-first Q&A

    What is the CRediT taxonomy?

    The CRediT taxonomy is a standardised system of 14 roles for describing what each author contributed to a research output. CASRAI originated it in 2014; it is now formally stewarded by NISO as ANSI/NISO Z39.104-2022. Journals that adopt it ask authors to select applicable roles rather than write free prose.

    What are the 14 roles in the CRediT taxonomy?

    The 14 roles are Conceptualization, Data Curation, Formal Analysis, Funding Acquisition, Investigation, Methodology, Project Administration, Resources, Software, Supervision, Validation, Visualization, Writing – Original Draft Preparation and Writing – Review & Editing. Authors can hold multiple roles, and roles can be shared across a byline.

    Is PLOS ONE a credible journal?

    PLOS ONE is a fully peer-reviewed, indexed journal published by the non-profit Public Library of Science. Its mandatory, structured CRediT and ORCID requirements are part of a broader editorial-integrity framework that includes ICMJE-aligned authorship criteria and COPE-based authorship-dispute handling.

    Is it good to publish in PLOS ONE?

    For authors who want transparent, machine-readable contribution records, publishing in PLOS ONE means every co-author’s role is captured in structured form and published alongside the article — a stronger provenance record than a narrative statement, though editorial fit and scope should still guide the submission decision.

    Implications and what comes next

    For research administrators and institutions, PLOS ONE’s model is a working template for what “compliance-ready” contributorship metadata looks like: mandatory at submission, tied to ORCID, and published as structured data rather than prose. Funders and institutions assessing individual contribution to collaborative outputs gain a queryable record instead of having to parse inconsistent narrative statements.

    For publishers still using an optional or free-text model, the PLOS ONE case demonstrates that a mandatory, role-based submission field is operationally achievable at very high volume — PLOS ONE has published hundreds of thousands of articles under this requirement. As more journals move toward structured contributorship, the gap between “CRediT as a suggested framework” and “CRediT as an enforced, machine-readable field” is likely to become the more meaningful dividing line in authorship transparency than whether a journal mentions CRediT at all.

    Research administrators evaluating a journal’s authorship rigour should check not just whether CRediT is referenced in author guidelines, but whether role assignment is enforced as structured, mandatory metadata — the distinction this case study sets out to make clear.

  • CRediT Taxonomy Under NISO: Inside Z39.104-2022

    The CRediT taxonomy is governed today by the National Information Standards Organization (NISO), not by the group that originally designed it. Formal stewardship sits with the credit taxonomy niso standard, ANSI/NISO Z39.104-2022, whose maintenance runs through a NISO CRediT Standing Committee that reviews proposed changes and coordinates revisions to the published standard.

    ANSI/NISO Z39.104-2022 is the American National Standard that formalises the Contributor Roles Taxonomy (CRediT) — a controlled vocabulary of 14 contributor roles used by scholarly journals to describe individual research contributions, approved by ANSI on 14 January 2022 and published by NISO on 8 February 2022.

    CASRAI originated the CRediT contributor role taxonomy in 2014. The standard is now stewarded by NISO as ANSI/NISO Z39.104-2022. Understanding where CASRAI’s design work ends and NISO’s formal governance begins matters for any publisher, institution, or developer deciding how to submit a correction, propose a new role, or cite the standard accurately.

    Who stewards the CRediT taxonomy today?

    NISO stewards the CRediT taxonomy through ANSI/NISO Z39.104-2022, a standard approved by the American National Standards Institute and published by NISO. Stewardship is distinct from origination: CASRAI and a cross-institutional pilot group designed the original taxonomy, but formal, ongoing governance now belongs to NISO’s standards infrastructure.

    This distinction is not a technicality. It determines who has authority to add, deprecate, or clarify a contributor role, and it is why publishers citing the standard should reference ANSI/NISO Z39.104-2022 rather than an unversioned “CRediT taxonomy” with no governing body attached.

    Aspect CASRAI’s original design work NISO’s formal stewardship
    Period 2012 pilot through 2015 launch 2020 work item to present
    Origin event 2012 Wellcome Trust / Harvard University workshop with ICMJE-affiliated biomedical journal editors 2020 NISO work item to register CRediT as an ANSI/NISO standard
    Governing body CASRAI-convened pilot group NISO CRediT Standing Committee
    Formal designation None — informal taxonomy ANSI/NISO Z39.104-2022
    Licence Open, community use CC-BY 4.0, per credit.niso.org
    Change authority Original design team NISO Standing Committee via ANSI balloting

    How is the Z39.104 working group structured?

    The NISO working group that produced Z39.104-2022 was deliberately cross-sector, drawing named representatives from publishers, funders, universities, and research-consulting firms rather than a single stakeholder type. That composition is itself a governance signal: no one sector controls the standard.

    Publicly listed contributors to the NISO work item included representatives from PLOS, Oxford University Press, Taylor & Francis Group, IOP Publishing, UK Research and Innovation (UKRI), Northwestern University, Université de Montréal, Carnegie Mellon University, and the Mathematical Association of America, alongside independent research consultants.

    • Publishers — PLOS, Oxford University Press, Taylor & Francis, IOP Publishing, the Mathematical Association of America
    • Funders — UK Research and Innovation (UKRI)
    • Universities — Northwestern University, Université de Montréal, Carnegie Mellon University
    • Independent consultants — Research Consulting Limited and Kerridge Research Consulting

    Once ANSI approval completed in January 2022, this working group’s role transitioned into the standing NISO CRediT Standing Committee, which now provides the ongoing forum for feedback, implementation support, and future expansion of the taxonomy into subject areas beyond its original biomedical-publishing roots.

    What is the revision cadence for the standard?

    ANSI/NISO Z39.104-2022 does not operate on a fixed annual revision schedule. Instead, it follows NISO’s continuous-maintenance model: proposed changes can be submitted at any time, but they are only incorporated into a new dated version of the standard after the Standing Committee reviews them and, where warranted, NISO runs the change through formal ANSI balloting.

    Three dates anchor the standard’s history so far:

    • 2020 — NISO launches the work item to formalise CRediT as an ANSI/NISO standard, with a small working group focused on the existing 14 roles.
    • 14 January 2022 — ANSI approves the standard.
    • 8 February 2022 — NISO publishes ANSI/NISO Z39.104-2022.

    No subsequent dated revision has been published since 2022; proposed extensions — such as recognising acknowledged (non-authorship) contributions or peer-review credit — are discussed through the Standing Committee and the associated CRediT Community of Interest before any future ballot.

    How do publishers submit change requests?

    Publishers, institutions, and individual researchers can raise a proposed change to the taxonomy at any time; the request is then triaged by the NISO CRediT Standing Committee rather than acted on unilaterally by any single publisher.

    1. Draft the request in writing, specifying the exact role, definition, or scope change proposed and the use case it addresses.
    2. Route it to NISO for referral to the Standing Committee, including your name, affiliation, and contact details.
    3. Await committee review — the Standing Committee discusses submissions as part of its regular meetings and decides whether to advance them.
    4. Formal balloting — if the committee approves a substantive change, NISO carries it through ANSI’s standards-approval process before it appears in a revised, dated version of Z39.104.

    This is why individual publishers — Sage among them — note on their own author-guidance pages that not every journal has adopted CRediT yet, and direct queries to dedicated editorial mailboxes rather than to NISO directly: implementation decisions sit with each publisher, while the taxonomy itself sits with NISO.

    For institutions building internal guidance, CASRAI’s CRediT contributor roles hub and the individual CRediT role pages summarise the 14 roles for practical reference, alongside broader context in CASRAI’s authorship resources.

    Common questions about CRediT taxonomy governance

    What is the CRediT taxonomy?

    CRediT (Contributor Roles Taxonomy) is a controlled vocabulary of 14 contributor roles used to describe the specific contributions individuals make to a research output, distinct from a simple author byline. It has been in widespread scholarly-publishing use since 2015 and was formalised as ANSI/NISO Z39.104-2022 in 2022.

    What are the 14 roles of CRediT taxonomy?

    The 14 roles are Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, and Writing – review & editing. Each role can be attributed to more than one contributor, and each contributor can hold more than one role.

    Does every publisher use the same CRediT taxonomy?

    No. The taxonomy itself is standardised under ANSI/NISO Z39.104-2022, but adoption is uneven: some journals, including certain Sage titles, have not yet implemented CRediT statements at all. Standardisation of the vocabulary does not guarantee uniform implementation across every journal or publisher.

    The practical implication for research administrators is that citing “the CRediT taxonomy” without a version reference is no longer precise enough for policy documents, institutional repositories, or funder-reporting templates. ANSI/NISO Z39.104-2022 is the citable, versioned artefact; CASRAI’s 2014 design work is the historical origin, not the current governing document. As the Standing Committee’s remit expands toward acknowledged contributions and peer-review credit, expect the next dated revision to widen the taxonomy’s scope beyond its original 14 roles rather than replace them.

  • CRediT Taxonomy Investigation: Not Misconduct

    The credit taxonomy investigation role — formally “Investigation” in CRediT — covers hands-on data and evidence collection: running experiments, gathering samples, and testing hypotheses. It has no connection to a research-misconduct investigation, which is a formal institutional inquiry into fabrication, falsification, or plagiarism. The two share a word, not a meaning, and that overlap causes recurring confusion on author contribution forms.

    CRediT — the Contributor Roles Taxonomy — is a controlled vocabulary of 14 roles used to describe how each named author contributed to a research output. CASRAI originated CRediT in 2014; the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022, and its 14 role definitions are maintained at credit.niso.org.

    Table of contents

    What does “Investigation” mean in the CRediT taxonomy?

    Under ANSI/NISO Z39.104-2022, the credit taxonomy investigation role is defined as “conducting a research and investigation process, specifically performing the experiments, or data/evidence collection.” It is one of 14 defined contributor roles, sitting alongside Conceptualization, Methodology, Formal Analysis, and ten others.

    The role covers the operational middle of a study: the point where a planned method is actually carried out and data starts to exist. NISO’s role definition lists the following as typical Investigation tasks:

    • Following or modifying methods to collect or generate quantitative or qualitative data
    • Testing research hypotheses and documenting the research process
    • Searching and reviewing literature, samples, data, and other evidence
    • Reporting findings for further discussion, analysis, and exchange of ideas

    None of this concerns wrongdoing. A contributor credited with Investigation did fieldwork, ran assays, coded interviews, or otherwise generated the study’s raw material — nothing more, nothing less.

    How is CRediT’s Investigation role different from a misconduct investigation?

    A research-misconduct investigation is a formal institutional process triggered by a credible allegation of fabrication, falsification, or plagiarism. In the United States, the Office of Research Integrity defines these three categories under 42 CFR Part 93, the federal policy governing PHS-funded research. In the UK, institutions follow the UK Research Integrity Office (UKRIO) procedure and the Concordat to Support Research Integrity, and publishers typically follow COPE’s investigation flowcharts once a concern is raised.

    The two processes could not be more different in stakes, actors, or timing. The table below sets out the distinction — and adds a third homonym that also trips up search results: the everyday financial “credit investigation” run by lenders.

    Aspect CRediT “Investigation” role Research-misconduct investigation Financial “credit investigation”
    What it is One of 14 standard contributor-role labels A formal inquiry into research integrity breaches A lender’s check of a borrower’s repayment history
    Governed by ANSI/NISO Z39.104-2022 (CRediT) Institutional policy, UKRIO/COPE (UK), 42 CFR Part 93/ORI (US) Consumer-credit and lending regulation
    Triggered by Submitting a manuscript with an author contribution statement A credible allegation of fabrication, falsification, or plagiarism A loan or credit application
    Who is involved Named authors/contributors and the corresponding author Research integrity officer, appointed committee, the accused Lender, credit reference agency, applicant
    Typical outcome A credited line in the published contribution statement Finding of misconduct, correction, retraction, or exoneration Loan approval, denial, or adjusted terms

    Why does the confusion keep happening on contribution forms?

    Editors and journal staff routinely field author queries asking whether ticking “Investigation” on a CRediT form invites scrutiny of their conduct. It does not. The confusion has three compounding causes.

    First, the word “investigation” already has a dominant everyday meaning tied to wrongdoing — police investigations, misconduct investigations, workplace investigations — so authors default to that association before reading the CRediT-specific definition. Second, publisher-facing CRediT forms often list all 14 roles as bare labels with no inline definition, forcing authors to look up what each term means mid-submission. Third, search behaviour reflects a genuine third homonym: “credit investigation” is also standard terminology in consumer lending, where it means a lender checking a borrower’s repayment history — a completely unrelated financial process that has nothing to do with either scholarly authorship or research integrity.

    This is a naming problem, not a substantive ambiguity. Once a contributor sees the full NISO definition — data/evidence collection — the confusion resolves immediately. The friction is entirely at the point of first encounter, typically an unlabelled checkbox in a submission system.

    How should authors and editors correctly apply the role?

    Authors should select Investigation whenever they personally performed experiments, collected data, ran surveys or interviews, or gathered samples and evidence for the study — regardless of whether they also held other roles such as Methodology or Formal Analysis. CRediT roles are not mutually exclusive; a single contributor commonly holds several.

    Editors and journal staff can reduce the confusion at source by adding the one-line NISO definition directly beside each role checkbox in submission systems, rather than relying on authors to consult an external reference. This single change removes almost all first-time-user hesitation around the Investigation label.

    Institutions drafting internal contribution-disclosure policies should keep CRediT role assignment procedurally separate from any research-integrity policy documentation, even where both appear in the same manuscript-submission workflow, so that the two processes are never conflated administratively.

    Frequently asked questions

    What does “Investigation” mean in CRediT taxonomy?

    In CRediT, “Investigation” is the role covering the research and investigation process itself — performing experiments or collecting data and evidence. It sits alongside 13 other defined roles under ANSI/NISO Z39.104-2022 and describes hands-on data generation, not any form of wrongdoing inquiry.

    What is the CRediT taxonomy?

    CRediT (Contributor Roles Taxonomy) is a standardised, 14-role controlled vocabulary for describing each named author’s specific contribution to a scholarly work. CASRAI originated it in 2014; NISO now stewards it as ANSI/NISO Z39.104-2022, and major publishers including Elsevier, Wiley, Sage, and Taylor & Francis request it at submission.

    What are the criteria for authorship?

    ICMJE’s Recommendations set out four authorship criteria — substantial contribution to conception/design or data acquisition/analysis; drafting or critical revision; final approval of the published version; and accountability for the work’s integrity. Some secondary sources miscount this as five by splitting the first criterion.

    Does “credit investigation” mean the same as CRediT’s Investigation role?

    No. A financial credit investigation is a lender’s check of a borrower’s repayment history before approving a loan — a consumer-lending process with no connection to scholarly authorship. It shares only the surface phrase with CRediT’s data/evidence-collection role.

    Implications for editors and institutions

    Naming collisions like this one are a small but measurable source of submission friction: every unlabelled checkbox that requires an author to context-switch away from the manuscript to look up a definition adds time and risk of miscoding to the metadata that journals, funders, and indexers eventually rely on. Contribution statements feed downstream systems — CrossRef metadata, ORCID records, institutional research-information systems — so a mislabelled or abandoned Investigation entry is not a cosmetic error; it degrades the accuracy of the scholarly record’s provenance data.

    As more funders and institutions move toward requiring structured contribution statements alongside authorship, the practical fix sits with journal and submission-system design, not with the taxonomy itself: inline definitions, tooltips, or a linked glossary at the point of role selection resolve the ambiguity before it becomes a support ticket. The taxonomy’s 14 roles remain stable under ANSI/NISO Z39.104-2022; what needs to improve is how clearly each one is presented at first encounter.

  • Author Contributions Credit: The Evidence on Authorship Disputes

    Author contributions credit statements built on the CRediT taxonomy help structure and resolve authorship disputes once they arise, but published 2025–2026 evidence does not show they reliably prevent gift authorship, ghost authorship or misattribution before it happens. Formal CRediT declarations are a documented dispute-resolution aid, not a proven dispute-prevention mechanism.

    CRediT (Contributor Roles Taxonomy) is a standardised set of 14 role labels — including Conceptualization, Investigation and Writing – Original Draft — used to describe each named contributor’s specific input to a research output. CASRAI originated the CRediT contributor role taxonomy in 2014, following a 2012 workshop convened with Harvard University and the Wellcome Trust; the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022.

    What does CRediT actually promise to fix?

    CRediT was designed to replace a single, undifferentiated author byline with a granular breakdown of who did what. The rationale, set out by Brand, Allen, Altman, Hlava and Scott in Learned Publishing (2015), was that opaque author lists make it hard to distinguish substantial intellectual contribution from honorary inclusion, and that a shared vocabulary of roles would reduce the ambiguity that fuels disagreement.

    That rationale has been widely adopted. Elsevier, Wiley, Taylor & Francis, Sage and the Royal Society all require or encourage CRediT statements, and journals frequently cite “reducing authorship disputes” as a stated aim. The open question — the one this article addresses — is whether the taxonomy’s real-world track record supports that claim, or whether it functions mainly as a transparency exercise that leaves the underlying disputes largely unchanged.

    Does CRediT help resolve disputes once they arise?

    The clearest empirical evidence so far concerns resolution, not prevention. Partin and Hosseini, writing in Accountability in Research (published online 7 December 2025), describe how the US National Institutes of Health Intramural Research Program uses CRediT as a fact-finding tool once an authorship dispute has already been raised.

    In that process, disputing parties are asked to independently complete CRediT-based contribution grids for every person involved in a project. Investigators then compare the resulting maps to identify where perceptions diverge. The NIH approach involves two broad stages:

    1. An informal stage, in which coauthors are asked to discuss and reconcile their CRediT assignments directly, ideally before submission or shortly after a disagreement surfaces.
    2. A formal fact-finding stage, used when informal discussion fails, in which a neutral investigator combines CRediT grids with interviews, manuscript drafts and laboratory records to reach a documented determination.

    Partin and Hosseini report that CRediT is genuinely useful here because it forces disputants onto a common vocabulary, reducing the scope for talking past one another. Their central finding, however, is that CRediT is a non-hierarchical taxonomy: it lists what each person did but cannot itself weigh how important a given contribution was relative to another. Deciding whether “Conceptualization” outweighs “Investigation” in a specific case still requires human judgement from the investigator, not the taxonomy.

    Does CRediT prevent gift authorship and misattribution in the first place?

    On prevention, the evidence is weaker and more mixed than the resolution evidence above. A 2025 scoping review in Accountability in Research, examining implementation barriers and improvement strategies for CRediT, found that the taxonomy’s limited applicability across research types, unresolved ethical concerns, and persistent interpersonal conflict among contributors continue to undermine its stated aims — even in journals that mandate CRediT statements at submission.

    Two further data points reinforce this picture:

    Study Claim tested What the evidence found Verdict
    Partin & Hosseini (2025), Accountability in Research CRediT helps resolve disputes once raised Structures fact-finding and shared vocabulary at NIH IRP; cannot rank contribution importance Supported, with limits
    De Peuter et al. (2025), Quarterly Journal of Experimental Psychology Contribution disclosure prevents gift/ghost authorship Among >800 psychology researchers surveyed, almost two-thirds had experienced gift authorship and roughly a quarter ghost authorship at least once; nearly half had witnessed gift authorship more than once Not supported
    Sauermann & Haeussler (2017), Science Advances Contribution statements displace author-order bias Evaluators still weighted author order more heavily than stated contributions; junior researchers reported concern over statement placement Not supported

    Sauermann and Haeussler’s survey-based study is particularly relevant to the “symbolic” critique: even where contribution statements exist, readers and evaluators continued to rely on legacy signals such as author order and position, which leaves room for CRediT statements to be completed pro forma rather than as a genuine check on inclusion. Combined with the 2025 scoping review’s findings on persistent ethical concerns, the pattern across the literature is consistent: CRediT changes how disputes are discussed far more reliably than it changes whether questionable authorship is awarded in the first place.

    None of this means CRediT statements are worthless. The ICMJE’s authorship criteria and COPE’s guidance both continue to treat granular contributorship as good practice, and NISO’s community-owned taxonomy gives institutions a shared reference point that did not exist before 2014. What the 2025–2026 literature does not support is the stronger claim, sometimes made in publisher marketing copy, that adopting CRediT measurably reduces the incidence of gift authorship or misattribution across a journal’s output.

    Common questions on CRediT and authorship disputes

    What can lead to disputes over authorship?

    Authorship disputes most often arise from unclear expectations set at the start of a project, uneven communication as roles shift, and disagreement over how to rank contributions such as data collection versus manuscript writing. Late additions or omissions of contributors, and pressure to include senior staff who did not meet authorship criteria, are also common triggers.

    How to resolve authorship disputes?

    Institutional guidance, including Harvard’s authorship guidelines, recommends that disputes are best settled directly among coauthors through structured discussion, ideally using a shared contribution framework such as CRediT. Where informal discussion fails, escalation to a neutral institutional fact-finder — as practised at the NIH — combines CRediT grids with interviews and documentary evidence to reach a determination.

    Why is it important to give credit to authors?

    Accurate attribution of credit underpins research accountability: it identifies who is answerable for which parts of a study, supports fair evaluation in hiring and funding decisions, and protects the scholarly record against both over- and under-crediting. ICMJE guidance ties authorship directly to accountability for the reported work, not merely recognition.

    How to credit authors in research?

    Journals following the CRediT taxonomy ask the corresponding author to assign each contributor one or more of the 14 standard roles — such as Methodology, Formal Analysis or Supervision — during submission, with all coauthors expected to review and agree the assignments before publication. CRediT does not itself alter a journal’s underlying authorship-eligibility criteria.

    What this means for institutions, journals and funders

    For research offices and integrity officers, the practical implication is to treat CRediT as a structured mediation tool, not a preventative control. Building CRediT-based contribution grids into project agreements from the outset — before a manuscript is drafted — gives disputes a documented baseline to be resolved against, mirroring the NIH IRP model described by Partin and Hosseini.

    For journals and publishers, the 2025 scoping review’s findings suggest that mandating CRediT statements without accompanying editorial verification is unlikely to move the needle on gift or ghost authorship rates. Verification steps — such as requiring all coauthors to individually confirm their assigned roles, rather than accepting a single corresponding-author submission — would more directly address the “pro forma completion” risk that Sauermann and Haeussler’s findings imply.

    Looking ahead, the research gap is specific and addressable: no published study yet compares gift-authorship or dispute rates between matched journals that do and do not require CRediT statements. Until that comparative evidence exists, institutions should present CRediT accurately — as originated by CASRAI in 2014 and now stewarded by NISO under ANSI/NISO Z39.104-2022 — as a transparency and resolution aid with a proven role in mediating disputes, rather than as a demonstrated fix for the authorship misconduct it was designed to curb.

  • CRediT Roles for Data Curation and Software

    CRediT roles such as Data Curation and Software give data managers and research software engineers a standardised, citable way to be formally recognised for research contributions that byline authorship rules routinely exclude. Two of the taxonomy’s fourteen roles map directly onto their work, and institutions can use those role tags — including the optional “degree of contribution” qualifier — as documentary evidence in annual review, promotion and tenure cases.

    The Contributor Role Taxonomy (CRediT) is a controlled vocabulary of fourteen contributor roles used to describe the specific contributions individuals make to a research output, independent of authorship order. CASRAI originated the CRediT contributor role taxonomy in 2014; the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022, approved in 2022.

    What Do the Data Curation and Software CRediT Roles Cover?

    Data Curation and Software are two of the fourteen roles in the CRediT contributor role taxonomy, and together they cover most of the technical infrastructure work that keeps a research output usable and reproducible. Under NISO’s definitions, Data Curation covers management activity that annotates, cleans and maintains research data — including software code where it is needed to interpret that data — for both initial use and later reuse. Software covers programming and software development: designing programs, implementing code and algorithms, and testing existing code components.

    These definitions were written broadly enough to capture roles that rarely appear on a title page:

    • Research data managers and data stewards who build metadata schemas and manage repository deposits
    • Research data librarians who oversee data management plans and long-term preservation
    • Research software engineers (RSEs) who build, maintain and test the pipelines and analysis code a study depends on
    • Bioinformaticians and computational scientists whose code is central to the result but who did not draft the manuscript

    Several adjacent roles frequently apply to the same people: Data Curation and Software combine most often with Methodology, Validation, Resources (for computing infrastructure) and Visualization. A single individual can — and typically does — hold more than one CRediT role on the same output; NISO’s implementation guidance is explicit that a role can also be assigned to multiple contributors on the same paper.

    Why Are Data Managers and Software Engineers Excluded From the Byline?

    Byline authorship is governed by criteria such as those from the International Committee of Medical Journal Editors (ICMJE), which require substantial contribution to conception, drafting or critical revision, plus final approval and accountability. Building a data pipeline or curating a dataset frequently fails one of those tests — usually drafting or revision — even when the contribution was essential to the result.

    CRediT was designed to sit alongside, not replace, those authorship criteria. NISO’s own guidance states plainly that CRediT “is not designed to determine authorship” but instead documents the specific contributions of authors and other contributors, including people who would otherwise only appear in an acknowledgements line. That separation is precisely what makes it useful for technical staff: a data manager or RSE can receive a formal, structured, machine-readable Data Curation or Software credit on an output without needing to clear the higher bar of full authorship.

    How Can Institutions Use CRediT Roles in Performance Review?

    Research institutions can treat a contributor’s accumulated CRediT role tags as structured evidence of impact, distinct from and complementary to publication counts or authorship position. Because the roles are standardised and, where publishers tag them in JATS XML, machine-readable, they can be aggregated across a person’s ORCID record rather than re-argued from scratch at every review cycle.

    Traditional CV evidence CRediT-based evidence Use in review
    Author position (2nd, 5th, etc.) Named role (Data Curation, Software) Documents what was actually done, not just list order
    General “contributed to analysis” statement Structured, standardised role plus optional degree of contribution Comparable across papers, journals and disciplines
    Acknowledgements-only mention Formal CRediT role on the same footing as co-authors Countable, citable line in annual review or promotion dossier
    Manual claim, hard to verify Machine-readable tag exportable via ORCID / Crossref metadata Independently verifiable by a review committee

    Concrete steps an institution can take:

    • Require staff to list their specific CRediT roles — not just “co-author” — against each output in annual review and promotion documentation
    • Instruct principal investigators to assign Data Curation and Software roles at the manuscript-submission stage, before contributor memory fades
    • Ask review committees to weight a sustained pattern of Data Curation or Software roles as evidence of infrastructure contribution, on comparable footing with first authorship in relevant career tracks
    • Encourage contributors to keep their ORCID record current so CRediT roles from publishers using CRediT-tagged JATS XML populate automatically

    What Does CRediT Leave Undefined? The Degree-of-Contribution Gap

    CRediT includes an optional qualifier — “lead”, “equal” or “supporting” — that can be attached where several people share the same role. This degree-of-contribution tag is exactly what a performance-review committee needs to distinguish a data manager who led curation for a multi-year cohort study from one who supported it for a single deposit. It is the single most under-used lever institutions have available inside a taxonomy most already partially adopt.

    Two constraints matter for institutions relying on it. First, NISO’s implementation guidance is explicit that degree-of-contribution is not currently part of the CRediT standard itself — individual publishers decide whether to request it, so its presence in a contributor statement is inconsistent across journals. Second, adoption of CRediT overall is uneven: PLOS has made CRediT its sole standard across all journals, while others offer it optionally through Editorial Manager (integrated since 2016) or Clarivate’s ScholarOne (integrated since 2018). Crossref has stated it will add CRediT to its publisher metadata schema in 2026, which should make role and degree-of-contribution data far easier to aggregate at the institutional level once adopted — but until then, institutions auditing technical contributions cannot assume degree-of-contribution data exists for every paper in a portfolio, and should ask contributors to supply it directly where the publisher record is silent.

    Frequently Asked Questions About CRediT Roles

    What is a credit role?

    A CRediT role is one of fourteen standardised labels — such as Data Curation, Software, Methodology or Investigation — used to describe a specific type of contribution an individual made to a research output. Roles are assigned independently of authorship order and can be combined, so one contributor may hold several roles on the same paper.

    What are the 14 credit contributor roles?

    The fourteen roles are Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, and Writing – review & editing, as defined under ANSI/NISO Z39.104-2022.

    What does role taxonomy mean?

    A role taxonomy is a fixed, controlled vocabulary that classifies types of contribution rather than free-text description. CRediT’s role taxonomy standardises how contributions are labelled across journals and disciplines, so a “Software” credit means the same thing whether the paper is in genomics or economics.

    Implications for Research Assessment

    As Crossref’s planned 2026 metadata integration and continued publisher adoption make CRediT data more machine-readable, the practical barrier to using it in institutional review shifts from availability to policy: whether promotion committees are willing to treat a Data Curation or Software role, especially one tagged “lead”, as comparable evidence to first authorship in the relevant career track. Institutions that update review criteria now — ahead of that data becoming routine — will be positioned to credit data managers and research software engineers on the strength of a standard that, unlike free-text acknowledgements, is built to be counted.