CASRAI Dictionary

Tag: REF 2029 units of assessment

  • REF 2029 Units of Assessment: What’s Changing From REF 2021

    REF 2029 keeps the same 34 units of assessment (UoAs) and four main panels used in REF 2021 — the structural break is inside the scoring model, where outputs, impact and environment have been renamed, rebalanced and, for staff counting, decoupled from individual researchers. Institutions do not need to remap subject groupings; they need to replan how work within each UoA is scored and evidenced.

    The REF 2029 units of assessment are the 34 subject-based categories into which every UK higher education institution (HEI) submits research for the Research Excellence Framework, the UK-wide system that informs roughly £2 billion a year of block-grant research funding, run by Research England on behalf of the four UK funding bodies.

    What structure does REF 2029 keep from REF 2021?

    REF 2029 retains the REF 2021 unit-of-assessment structure in full. Submissions are still made into 34 UoAs, each with its own expert sub-panel, sitting under four main panels, according to the official REF 2029 guidance published by Research England.

    • Main Panel A — Medicine, Health and Life Sciences: 6 UoAs, including Clinical Medicine and Biological Sciences.
    • Main Panel B — Physical Sciences, Engineering and Mathematics: 6 UoAs, including Chemistry, Physics, Computer Science and Informatics.
    • Main Panel C — Social Sciences: 12 UoAs, including Economics and Econometrics, Business and Management Studies, and Education.
    • Main Panel D — Arts and Humanities: 10 UoAs, including History, English Language and Literature, and Music, Drama, Dance, Performing Arts, Film and Screen Studies.

    Main panels set overall criteria and ensure consistency; sub-panels develop discipline-specific assessment criteria and carry out the detailed review of submissions. Full expert panel membership for all 34 UoAs was appointed and announced in September 2025, and panels began meeting to set criteria in early 2026.

    How are the three assessment elements changing?

    The structural change that matters for planning purposes is not the UoA list — it is the renaming and rebalancing of the three elements every submission is scored against. All three have been renamed for REF 2029, and two of the three weightings have moved.

    Assessment element REF 2021 REF 2029 What changed
    Research outputs Outputs — 60% Contribution to Knowledge and Understanding (CKU) — 55% Renamed; weighting cut by 5 percentage points; minimum-output rule removed
    Non-academic benefit Impact — 25% Engagement and Impact (E&I) — 25% Renamed; weighting unchanged; 2* quality threshold for underpinning research removed
    Research culture Environment — 15% Strategy, People and Research Environment (SPRE) — 20% Renamed twice (via an interim “People, Culture and Environment” label); weighting raised by 5 points

    The Environment element had the most turbulent path to REF 2029. Research England’s four funding bodies initially proposed a “People, Culture and Environment” (PCE) element weighted at 25%, ran a sector-wide PCE pilot, then paused criteria-setting in September 2025 on the instruction of UK Science Minister Lord Vallance. When criteria-setting resumed on 10 December 2025, the element was renamed Strategy, People and Research Environment and its weighting was set at 20%, with CKU confirmed at 55% and E&I at 25% — the figures that now stand for REF 2029.

    What’s changing in how staff and outputs are counted?

    REF 2029 continues the shift, begun in REF 2021, away from assessing named individuals. Institutions will no longer submit an individual staff census; instead, the “volume measure” that determines how many outputs and impact case studies a unit must submit is calculated from an average of HESA staff record data rather than a single REF census date.

    Within Contribution to Knowledge and Understanding, the REF 2021 requirement of a minimum of one output per researcher has been removed, reducing pressure on individuals who published little during the assessment period. A recommended (non-mandatory) maximum of five outputs per researcher has been reinstated for clarity, matching REF 2021 practice. Outputs remain linked to units via a “substantive link” policy rather than to named individuals.

    What’s changing for impact case studies and portability?

    Two burden-reduction changes affect Engagement and Impact. First, the minimum number of impact case studies required has been reduced to one for units with fewer than 9.99 full-time-equivalent (FTE) staff, with revised thresholds for larger units. Second, the REF 2021 requirement that underpinning research reach at least 2-star quality to support an impact case study has been removed entirely.

    REF 2029 also introduces limited portability for long-form and extended-process research outputs, such as monographs, alongside simplified requirements — a direct response to concerns raised by arts, humanities and social science submitters about the practical effect of decoupling outputs from individual staff moves.

    How should institutions plan REF 2029 submissions?

    Because the 34 UoAs are unchanged, institutions do not need to reorganise which subject groupings they submit into. Planning effort should instead focus on three areas the funding bodies have flagged as high-impact for REF 2029 preparation:

    1. SPRE evidence-gathering — institution-level and unit-level statements on strategy, people and research environment now carry a fifth of the total score, up from a 15% Environment weighting in REF 2021, so institutions should start collating research-culture evidence well ahead of the 2027 submission window.
    2. HESA data governance — because the volume measure now derives from HESA staff record averages rather than a census date, research administration and HR data teams need aligned processes for confirming eligible contracts and cost centres.
    3. Output selection strategy — with the minimum-output rule gone but a five-output ceiling retained, units should reassess how they allocate limited high-quality outputs across contributing staff under Contribution to Knowledge and Understanding.

    The REF 2029 submission window opens in autumn 2027, the submission deadline falls in autumn 2028, and results are planned for publication in December 2029, according to the REF 2029 timetable maintained by Research England. Institutions that treat SPRE, HESA-based volume measures and output-selection policy as the priority work-streams — rather than re-litigating UoA boundaries that have not moved — will be better positioned for the assessment phase.

    Answer-first Q&A on REF 2029

    What are the REF 2029 units of assessment?

    The REF 2029 units of assessment are the same 34 subject-based categories used in REF 2021, organised under four main panels covering medicine and life sciences, physical sciences and engineering, social sciences, and arts and humanities. Each UoA has its own expert sub-panel that develops discipline-specific criteria.

    What are the key changes for REF 2029?

    The key changes are not to the units of assessment but to how work within them is scored: outputs become Contribution to Knowledge and Understanding (55%), impact becomes Engagement and Impact (25%), and environment becomes Strategy, People and Research Environment (20%), alongside HESA-based staff-output counting.

    What publications are eligible for REF 2029?

    Eligible outputs are original research publications with a substantive link to an eligible staff member’s employment, assessed under Contribution to Knowledge and Understanding. REF 2029 removes the REF 2021 minimum of one output per researcher, reinstates a recommended maximum of five, and introduces limited portability for long-form outputs such as monographs.

    What does REF mean in a university context?

    REF stands for the Research Excellence Framework, the UK-wide system for assessing research quality in higher education institutions. It informs the allocation of around £2 billion a year in block-grant research funding and is run by Research England on behalf of the four UK higher education funding bodies.

    What comes next for REF 2029 planning

    With weightings confirmed and full panel criteria due by autumn 2026, the structural picture for REF 2029 is now settled enough for research administration teams to build submission timelines around it. The open questions that remain — detailed panel criteria, Code of Practice approval windows, and the mechanics of SPRE evidence templates — will be published in modules through 2026, making the REF 2029 guidance pages, not third-party summaries, the authoritative reference for any submission-critical detail.

  • REF 2029 Outputs Decoupling: What It Signals for Contribution Recognition

    REF 2029 decouples research outputs from named researchers: institutions submit outputs to a Unit of Assessment rather than to an individual, judged instead on a “substantive link” between the institution and the work. This shifts REF evaluation from researcher performance to institutional research environment, raising the stakes for how contribution is separately evidenced.

    Decoupling is the REF 2029 policy mechanism that removes the formal link between a submitted research output and the named staff member who produced it, so that outputs are assessed as belonging to a Unit of Assessment (UoA) rather than to an individual author.

    What does “decoupling” mean under REF 2029?

    Under REF 2029, outputs are submitted to a Unit of Assessment without staff details attached to individual pieces of work. No researcher name is carried through the submission record, and no output is presented as belonging to one specific author for assessment purposes.

    This reshapes the submitted category itself: what was previously called “outputs” is now Contributions to Knowledge and Understanding (CKU), which carries a 55% weighting in the overall institutional score, according to REF 2029’s official Section 1 overview. The remaining weighting splits between Strategy, People and Research Environment (20%) and Engagement and Impact (roughly 25%).

    Practical consequences of decoupling include:

    • No minimum or maximum number of outputs required from any individual staff member.
    • A recommended (not mandatory) ceiling of five outputs per researcher, reinstated in REF 2029’s December 2025 update after an earlier proposal for no cap at all.
    • Eligibility broadened to outputs produced by a wider range of roles, including technicians and research managers, not only conventionally “REF-able” academic staff.

    Instead of an author-output link, REF 2029 requires institutions to demonstrate a substantive link between the submitting institution and the output. A substantive link is generally established through an eligible employment relationship of at least 12 months at a minimum of 0.2 full-time equivalent (FTE), per REF 2029’s guidance on Contributions to Knowledge and Understanding.

    Where employment alone is insufficient or the researcher has since left, institutions may point to supporting evidence such as:

    • Internal research support, including funding for materials, technical assistance, or conference attendance.
    • Evidence of work-in-progress presentations, internal or external.
    • An external grant supporting a relevant programme of research held during the employment period.

    Outputs cannot be claimed where the substantive link occurred only after the output was made public and the author was subject to compulsory redundancy — a safeguard REF 2029 added following sector feedback on the risk of institutions retaining the outputs of staff they had made redundant.

    How does this differ from REF 2014 and REF 2021?

    Decoupling is not new to REF 2029; it extends a direction of travel set out in the 2016 Stern Review of the REF, which recommended non-portability of outputs to reduce “poaching” incentives that favoured wealthier institutions. Each REF cycle has progressively loosened the tie between researcher identity and institutional claim.

    REF cycle Output–researcher link Portability on staff move
    REF 2014 Output captured entirely by the institution employing the researcher at the census date Full transfer with the researcher
    REF 2021 Output could be captured by both the origin and destination institution on a move Partial (dual claim)
    REF 2029 Output captured by the institution demonstrating a substantive link; no named author attached Restricted; long-form outputs (e.g. monographs) retain five-year portability

    The volume of outputs an institution must submit is unchanged in formula terms: it remains the institution’s staff volume measure (FTE) multiplied by 2.5 at UoA level, consistent with the REF 2021 approach reported by REF 2029 planning guidance published by the University of Reading and others.

    What does decoupling signal for evidencing contribution?

    REF 2029’s decoupling addresses institutional-level attribution — which organisation gets credit for an output — but it does not answer a separate, longstanding question: which individuals, and in what capacity, actually contributed to producing it. That question sits squarely in the domain of contributorship taxonomies rather than research assessment exercises.

    This is where the REF 2029 shift and the contributor-role movement intersect without colliding. CASRAI originated the CRediT contributor role taxonomy in 2014 to make individual contribution to scholarly outputs explicit and machine-readable; the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022. Where REF 2029 deliberately removes the researcher’s name from the assessment record, CRediT statements retained in the published output itself remain the mechanism by which an individual’s specific role — conceptualisation, methodology, writing, data curation — stays evidenced and citable, independent of how any national assessment exercise chooses to allocate institutional credit.

    For institutions, the practical implication is that internal recognition, promotion, and workload evidence can no longer lean on REF submission data as a proxy for individual contribution, because REF 2029 submissions will not carry that data. Institutions building internal case files for tenure, promotion, or grant applications need contribution evidence that exists independently of the REF submission — structured CRediT role statements attached to outputs, ORCID-linked publication records, and clear internal documentation of the “substantive link” evidence (funding, supervision, work-in-progress records) that REF 2029 itself now requires institutions to compile.

    Guidance on research administration practice and on the underlying CRediT taxonomy is a reasonable starting point for research offices building this parallel evidence base ahead of the REF 2029 submission window.

    Answer-first questions on REF 2029 outputs

    What are the changes in REF 2029 for outputs?

    REF 2029 renames outputs as Contributions to Knowledge and Understanding (CKU), weighted at 55% of the overall score, removes individual minimum and maximum output requirements, reinstates a recommended cap of five outputs per researcher, and requires a substantive link rather than a named author for eligibility.

    Why is REF 2029’s decoupling of outputs important?

    It marks a formal shift in what REF measures: institutional research environment and support, not individual researcher performance. Funding allocation logic follows institutions, so REF 2029 aligns assessment evidence with who receives the funding — the institution — rather than the individual author of an output.

    What outputs are eligible for REF 2029?

    Eligible outputs must be brought into the public domain between 1 January 2021 and 31 December 2028, meet REF 2029’s open access requirements, and demonstrate a substantive link to the submitting institution. Outputs solely authored by PhD students or teaching-only staff are generally not eligible.

    Are REF 2029 outputs portable when staff move institution?

    Portability is now restricted rather than automatic. Long-form outputs, such as monographs, retain five-year portability so they stay attached to the author after a move; most other outputs are captured by whichever institution holds the substantive link at the point of submission.

    Implications and outlook for institutions

    Research offices preparing for REF 2029 face two parallel evidencing tasks rather than one. The first is REF-facing: documenting the substantive link — employment records, internal research support, grant funding — for every output an institution intends to submit. The second is internal: maintaining contribution records that support promotion, recognition, and researcher career narratives now that REF submissions themselves will not do this job.

    Sector commentary, including analysis from Wonkhe’s research and innovation desk, has framed this as REF revealing its true purpose: an institutional funding mechanism rather than a personal-merit exercise. Institutions that treat the two evidencing tasks as genuinely separate — REF eligibility on one track, individual contribution recognition on another — are better placed to avoid a governance gap where good research goes on the REF return but the people who did it go unrecorded anywhere durable.