CASRAI Dictionary

Tag: UKRI

  • Funder ORCID iD requirements: 2026 landscape

    The trajectory of ORCID iD requirements across major funders has been steady through 2018-2025: voluntary for PIs, recommended for PIs, required for PIs, and now in 2026 increasingly required for co-investigators and other named project personnel. This post is a practical map of the requirements as they stand in mid-2026, with attention to the operational implications for institutions handling cross-funder applications.

    The PI requirement

    For principal investigators, ORCID iDs are now effectively required across the major Western funders. UKRI requires the PI’s ORCID iD at application; the seven research councils’ submission system pre-fills bibliographic data from the ORCID record where available. The Wellcome Trust, the major UK charities, and the major European national funders (DFG, ANR, NWO, the Swiss SNSF, the Spanish AEI) have similar requirements.

    The US federal funders have moved more cautiously. NIH requires ORCID iDs in the biosketch via the SciENcv tool; NSF requires ORCID iDs for personnel in current-and-pending support; DOE, DOD, NASA, USDA have varying requirements with a convergent direction. The 2025 update to the Common Forms work standardised the ORCID requirement across federal-funder personnel forms, with the binding date in 2026.

    EU funding (Horizon Europe, ERC) requires ORCID iDs for PIs and an increasingly complete coverage of co-investigators. The Funding & Tenders Portal integrates with ORCID for bibliographic data retrieval.

    For PIs, the practical posture is that an ORCID iD is now table stakes. The CASRAI funder applicant guide tracks the per-funder requirements.

    The co-investigator expansion

    The 2026 development is the expansion of ORCID requirements to co-investigators and other named project personnel.

    UKRI’s 2025-2026 transition expanded the ORCID requirement to all named investigators on an application. The Funding Service expects ORCID iDs for everyone listed.

    NIH’s pilot to require ORCID iDs for all senior/key personnel began in 2025 and is now standard practice. Co-investigators without ORCID iDs cannot be listed.

    The EU Horizon Europe expansion is in progress, with the Funding & Tenders Portal expected to require ORCID iDs for all consortium personnel by end-2026.

    For institutions, this means the ORCID-iD adoption work that previously focused on PIs now needs to cover the entire research staff. Most research-intensive universities are at or near saturation for PIs; the broader staff coverage is lower and the expansion requires sustained adoption work.

    The structured-data side

    The funder ORCID requirements are about more than identifier collection. The funders increasingly consume structured data from ORCID at application time and write structured data back at award time.

    At application: funders pull biographical sketches, affiliation history, prior funding, and publication lists from ORCID. The applicant’s ORCID record becomes the source-of-truth for these data; the funder’s application system displays them with allow-edit-but-prefer-ORCID semantics. This is a substantial UX improvement over the previous pattern of researchers re-typing their bibliographic data into each funder’s system.

    At award: funders write the award itself to ORCID as a funding-record entry, and increasingly the project as a RAiD reference. The researcher’s ORCID record accumulates a complete funding history without manual entry.

    The ORCID 4.0 contribution model (discussed in our earlier post on the IDR roadmap) is the underlying schema that supports this. Funders that have implemented the 4.0 deposit patterns can write rich structured data; funders still on 3.x produce thinner records.

    Operational implications for institutions

    Three operational priorities for research-administration offices.

    First, institutional ORCID-iD coverage. Run a coverage audit: what fraction of your research staff have ORCID iDs, and are those iDs current and verified? Where coverage is incomplete, run an institutional ORCID-iD adoption campaign. Many institutions have done this for PIs and need to extend it to early-career researchers, postdocs, and research-staff categories that were previously deprioritised.

    Second, ORCID-integration in CRIS. Your CRIS should be reading from and writing to ORCID for affiliated researchers. The CRIS-ORCID integration patterns are well-documented; the CASRAI CRIS integration guide walks through them per major CRIS vendor. The integration removes substantial duplicate-data-entry burden from researchers.

    Third, application-support workflow. Your application-support workflow should be ORCID-anchored: the researcher’s ORCID record is the source for biographical data; the application’s data-entry tools should pull from ORCID by default; any application-specific data should be entered once and propagated to ORCID where appropriate.

    The edge cases

    Three edge cases deserve flagging.

    First, researchers without ORCID iDs. As funder requirements tighten, researchers without iDs become unable to apply. The case is straightforward to resolve (registration takes minutes) but the institutional support to ensure no researcher is blocked at application time is non-trivial.

    Second, researchers with duplicate or fragmented ORCID records. Some researchers have multiple ORCID iDs from different registration events; some have records that are out of date or incomplete. The ORCID record merge process and the institutional support for record cleanup are not always smooth. Institutional ORCID-administrator workflows should handle these cases.

    Third, international researchers. ORCID is a global infrastructure but adoption varies by country and discipline. International collaborators on a US- or EU-funded project may not have ORCID iDs; the application workflow needs to accommodate getting them registered. Many institutions have built bridge-support for this scenario.

    The 2027 trajectory

    Looking forward, the trajectory through 2027 is clear. ORCID iDs will be required not just for personnel but for funded outputs (publications, datasets, software) at the deposit step. The funder’s compliance tracking will run on the ORCID-anchored graph: did the awarded researcher produce the outputs they committed to, were the outputs FAIR-compliant, were the contributors recognised with CRediT roles? The institutional infrastructure to support this picture is most of what current adoption work is building.

    For institutions, the strategic implication is to treat ORCID adoption not as a compliance item but as research-infrastructure investment. The return on the investment compounds: each researcher with a complete ORCID record reduces friction across every subsequent application, every output deposit, every reporting cycle. The institutions that did the work in 2022-2025 are reaping the benefit in 2026; those that have not are still doing the manual work.

    Related dictionary entries

  • Project IDs in 2026: RAiD adoption update

    The Research Activity Identifier (RAiD) crossed several adoption thresholds in 2024-2025. ISO 23527:2022 standardisation completed; the Australian Research Data Commons reached operational scale; UKRI integrated RAiD into its funding workflow; the EU’s HORIZON identifier work began aligning with RAiD; the ARDC-led international RAiD Steering Group brought together national service providers from Australia, New Zealand, UK, Canada, and several EU member states. The May 2026 picture is meaningfully different from the May 2025 picture. This post is an adoption update.

    Where RAiD is now

    RAiD is operational at substantial scale in Australia, where the ARDC operates the national RAiD service and integration with the Australian Research Council and the National Health and Medical Research Council has matured. Every ARC and NHMRC grant from 2024 onward has a RAiD; the integration is a routine compliance item.

    RAiD is operational in New Zealand via a national service implementation aligned with the ARDC model.

    RAiD is operational in the UK via a UKRI-operated service, with integration into the Je-S successor (the UKRI Funding Service that launched in 2023). UKRI grants from 2024 onward have RAiDs; backfilling of historical grants is in progress.

    RAiD is in pilot in Canada via a CRDCN-led initiative, with the Tri-Agencies (CIHR, NSERC, SSHRC) participating in design.

    RAiD has affiliated national service providers in the Netherlands, Germany, and Finland; full EU integration is in development through the EOSC Federation work.

    RAiD is not yet operational at scale in the US. NIH’s evolving project-identifier work overlaps with but is not identical to RAiD; harmonisation discussions are ongoing.

    What RAiD actually carries

    A RAiD record carries: the project’s name and description; its participants with ORCID iDs; its institutional affiliations with ROR IDs; its funding sources with Funder Registry or ROR identifiers and grant identifiers; its outputs with DOIs and other PIDs; its temporal span; its status. The record is mutable: as the project evolves, the RAiD record is updated to reflect new participants, new outputs, new affiliations.

    The mutability is the design choice that distinguishes RAiD from a per-event identifier like a DOI. A project is a living entity for years; its identifier needs to grow with it. The RAiD service architecture supports this via versioning: each update produces a new version of the RAiD record, with the old versions preserved as historical states.

    The interlock with other PIDs

    RAiD’s value is largely in the interlock layer. A RAiD record references the ORCID iDs of its participants; ORCID 4.0 carries the RAiDs of its researcher’s projects. A RAiD references the DOIs of its outputs; Crossref and DataCite metadata reference the RAiD via the relationship blocks. A RAiD references the Funder Registry IDs of its funders and the grant DOIs (where they exist) of its grants; the Crossref Grant Linking System grants reference the RAiD via the contributes-to relationship.

    The result is a structured graph: from a RAiD, an integrator can traverse to participants (ORCID), institutions (ROR), funders (Funder Registry/ROR), grants (grant DOIs), and outputs (DOIs). The graph is queryable. The OpenAIRE Graph already operationalises this for European projects; the CASRAI persistent identifiers domain tracks the broader integration.

    What’s working well

    Three operational patterns deserve flagging.

    First, funder-issued RAiDs. The pattern of the funder issuing the RAiD on award and the awardee inheriting it has worked well. The funder has the structured grant data; the awardee has the operational knowledge of the project. The funder issues the RAiD with the structured data they have; the awardee updates it as the project evolves. This minimises the burden on researchers and ensures RAiD coverage is complete for funded work.

    Second, institutional-CRIS integration. CRIS systems that ingest RAiDs from their researchers’ projects and propagate them to outputs as a metadata field have closed the project-to-output linkage that previously required string-matching grant numbers. The integration is straightforward; the value compounds over time as the historical record accumulates.

    Third, cross-funder collaboration. A project with multiple funders (typical in large clinical trials and EU consortia) can have a single RAiD referencing all the funders’ grants. This addresses a longstanding accounting friction where multi-funder projects appeared as multiple disconnected projects in funder reporting systems.

    What’s not working yet

    Three issues remain open.

    First, retroactive RAiDs for historical projects. RAiD coverage is forward-looking from each jurisdiction’s start date. Historical projects (pre-2022 or so) do not have RAiDs; building the historical record is a substantial data-engineering effort that no jurisdiction has fully completed.

    Second, international coordination. Different jurisdictions have different RAiD service providers, different operational arrangements, and slightly different metadata profiles. The RAiD Steering Group is working on harmonisation but the work is incomplete. A project that crosses jurisdictions may have RAiDs from multiple providers, with the integration between them not yet seamless.

    Third, the unfunded-project case. RAiD was designed around funded projects, with the funder as the natural issuer. Unfunded research activity (self-funded, doctoral student projects without grants, community-research projects without traditional funders) does not have a clear RAiD-issuance path. The RAiD service architecture supports researcher-issued RAiDs; the institutional and funder workflows have not fully accommodated this case.

    What integrators should do

    For institutions running a CRIS, the priorities are: ingest RAiDs into the project record; propagate to outputs as metadata; reconcile with ORCID’s funding and contribution data; surface in research-administration reporting.

    For publishers, the priority is to accept RAiDs in submission systems as a funding-reference option alongside Funder Registry entries and grant DOIs, and to deposit RAiDs to Crossref via the relationships block. Several publishers have done this; broader adoption through 2026 would be welcome.

    For funders that have not yet issued RAiDs, the priority is to evaluate the operational integration. ARDC’s documentation and the UKRI implementation are useful reference points. The integration is non-trivial but not large; institutions that have done it report it pays back within 18 months in reduced cross-system reconciliation effort.

    The broader pattern

    RAiD adoption is the latest instance of the persistent-identifier pattern: a structured identifier for a class of research entities, with an operational service to mint and resolve them, with metadata that interlocks with other PIDs, with adoption that takes years to reach scale but compounds in value as it does. ORCID took a decade to reach saturation; ROR took five years to reach the equivalent in its space; RAiD is plausibly on a five-to-seven-year trajectory to comparable coverage.

    For the CASRAI community, the practical posture in 2026 is to incorporate RAiD into integration designs from the outset, to track adoption by jurisdiction, and to advocate for adoption where the operational case is strong. The PID quartet of ORCID-ROR-RAiD-DOI is increasingly the foundation on which research-information integration is built; the more complete that foundation, the more useful the integration layer becomes.

    Related dictionary entries

  • The case for narrative CVs beyond UKRI

    The UK Research and Innovation Resume for Research and Innovation (R4RI), launched in pilot in 2019 and now the default CV format across UKRI’s seven research councils, has accumulated enough operational experience to draw lessons. The narrative-CV approach has spread internationally: the Dutch Research Council, the Swiss National Science Foundation, the Royal Society’s CV format, parts of the EU’s Horizon Europe evaluation, and several US private funders have adopted variants. Where it has not spread is the major US federal funders. This post argues the case for broader adoption with reference to the UKRI experience.

    What R4RI is

    R4RI replaces the traditional publication-list CV with a structured narrative covering four modules: how the researcher has contributed to the generation of new ideas, tools, methodologies, or knowledge; how they have contributed to the development of others; how they have contributed to the wider research community; how they have contributed to broader research and innovation users and audiences, and the wider environment. Each module has a 250-word limit and the researcher provides specific evidence.

    What R4RI explicitly does not ask for: lengthy publication lists, journal impact factors, h-indexes, citation counts, exhaustive grant histories. The Resume for Research and Innovation can include publications but as evidence of contribution, not as a metric.

    Five years of operational experience

    UKRI’s commissioned evaluations of R4RI (published 2022, 2024) and the broader literature on narrative-CV use have produced a reasonably clear picture.

    First, reviewer time. The early concern was that narrative CVs would take longer to review than conventional ones. The evaluation data show modest increases for first-time reviewers, settling to comparable or shorter review times once reviewers were familiar with the format. Reviewers report that R4RI gives them a clearer picture of the applicant’s actual contribution.

    Second, applicant time. Writing an R4RI takes longer than updating a publication list, and applicants without writing support are at a disadvantage. The equity implication is real: a researcher with access to research-administration support to help draft R4RI does better than one without. UKRI has invested in writing-support resources and several institutions have built internal capacity.

    Third, career-stage equity. Narrative CVs perform better for early-career researchers whose publication record is short but whose contribution is significant; they perform better for researchers with non-traditional career paths; they perform better for researchers in disciplines where high-impact publication is not the norm. They perform less well for researchers with very strong conventional records who feel the narrative format does not adequately recognise their publications. On balance the evaluation suggests narrative CVs reduce systemic bias against under-represented career patterns.

    Fourth, inter-rater reliability. The concern that narrative CVs would produce more variable assessment than metric-based CVs has been partially borne out: inter-rater reliability is somewhat lower for R4RI than for conventional CVs. This is in part a feature, not a bug — different reviewers genuinely weight different contributions differently, and the narrative CV surfaces those judgements. UKRI has responded with reviewer-training resources and structured rubrics.

    Why other funders should adopt narrative CVs

    Four reasons.

    First, narrative CVs operationalise DORA and CoARA commitments in a concrete way. A funder that has signed DORA but continues to use publication-metric CVs is operating in contradiction with its commitment; a funder that adopts a narrative CV format is operationalising it. The CASRAI responsible assessment domain tracks the gap between policy and practice across major funders.

    Second, narrative CVs make CRediT more useful. A CV that reports CRediT roles for the applicant’s recent papers gives the reviewer specific information about contribution; a publication list without CRediT gives only the byline order. The integration is operationally simple: narrative CVs cite specific contributions, CRediT statements describe what the contribution was.

    Third, narrative CVs reduce the metric-feedback loop. The pernicious cycle in which researchers chase high-impact-factor publications because funders weight them and funders weight them because researchers seek them is one of the system pathologies that responsible-assessment reform aims to break. A narrative CV format breaks the funder side of the loop, which gives researchers permission to optimise differently.

    Fourth, narrative CVs encode the broader range of contributions that modern research actually requires. Software, datasets, public engagement, peer review, mentorship, leadership of community-led infrastructure, contributions to open standards — none of these show up well in a conventional CV. They show up in a narrative CV. The CASRAI institutional responsible-assessment guide includes a checklist of contribution types that narrative formats can capture.

    The objections, addressed

    Three objections recur and deserve direct responses.

    Narrative CVs are subjective and unscientific. The metric-based CV is also subjective: someone decided which metrics to weight, what the weighting should be, and what counts as success. Narrative CVs make the subjectivity explicit and reviewable; metric-based CVs hide it behind a number.

    Narrative CVs disadvantage non-English-speakers. The concern is real and the mitigation is to allow CVs in the applicant’s working language with translation support funded by the funder. UKRI does not currently allow non-English R4RIs because UKRI operates in English; an international funder with multilingual operations would need to.

    Narrative CVs are too long for high-volume review. The 250-word-per-module limit and the four-module structure produce a CV that is no longer than a conventional 5-page academic CV; in many cases shorter. The objection is empirically wrong as stated.

    Practical recommendations

    For funders considering adoption, the practical steps are: pilot with one or two grant streams; train reviewers with worked examples; develop a structured rubric for scoring; provide writing-support resources for applicants; commit to an evaluation at year three; iterate the format based on the evaluation. UKRI’s experience suggests this approach yields a usable format within three years.

    For institutions supporting applicants, the practical steps are: build internal capacity to support narrative-CV drafting; offer it equitably across career stages and disciplines; treat it as part of the research-administration support package, not as an exceptional service for senior researchers only.

    For applicants, the practical advice is to start drafting in a narrative format now, even where the funder does not require it. The discipline of articulating contribution rather than enumerating publications produces a richer self-understanding of one’s own research and is useful for promotion, tenure, and personal career planning regardless of funder requirements.

    The trajectory

    We expect narrative-CV adoption to accelerate through 2026-2028. The CoARA commitment to reforming research assessment, combined with the maturity of the R4RI model, gives funders a credible template to adopt. The remaining holdouts are the major US federal funders (NIH and NSF), whose biosketch and current-and-pending-support formats are partial moves toward narrative but retain substantial metric content. The next round of US-funder review-criteria revision will be the test.

    Related dictionary entries