REF 2029 impact case studies (ICS) keep the same three-part narrative template as REF2021 but tighten the evidence standard: a hard 2,200-word cap replaces the old five-page limit, the 2* quality threshold for underpinning research is removed, the smallest units may now submit just one ICS instead of two, and funded-research submissions must add ROR and ORCID identifiers. These changes were confirmed in Research England’s official REF 2029 guidance, published 10 December 2025, and apply to submissions covering the 2020–2028 impact period.
An impact case study is the structured evidence record — summary, underpinning research and details of impact, corroborated by external sources — through which a REF submitting unit demonstrates the reach and significance of a specific, research-derived benefit to the economy, society, culture, public policy, health or the environment beyond academia.
Contents
- What has actually changed for REF 2029 impact case studies?
- How does the REF 2029 word limit compare with REF2021’s five-page rule?
- How many impact case studies does each unit need to submit?
- What’s different about the evidence and underpinning research standard?
- Does the REF 2029 impact case study template change?
- Answer-first Q&A
- What this means for research administrators
What has actually changed for REF 2029 impact case studies?
REF 2029 preserves REF2021’s basic architecture — impact case studies scored on “reach and significance,” submitted through the same three-section narrative — but revises four load-bearing rules. Word count moves from an approximate page limit to a hard 2,200-word cap. The 2* research-quality threshold for underpinning research is scrapped entirely. The minimum case-study count for the smallest units drops from two to one. And funded ICS must now supply Research Organization Registry (ROR) and ORCID identifiers alongside grant data.
Research England frames this as evolutionary, not structural: the guidance states it wanted “to limit the overall degree of change for this exercise.” Institutions that ran REF2021 will recognise the process shape; the real shift is in what counts as eligible evidence.
How does the REF 2029 word limit compare with REF2021’s five-page rule?
REF2021 capped impact case studies at five pages, with indicative (non-binding) word counts per section. REF 2029 replaces that with an explicit, enforced ceiling: 2,200 words across the three substantive narrative fields — Summary of the impact, Underpinning research, and Details of the impact — entered directly into the REF Submission system rather than uploaded as a formatted document.
Research England’s own analysis found REF2021 submissions typically used around 2,000 words within the five-page allowance — so the 2,200-word ceiling is a modest increase on observed practice, not a real-terms cut, despite the headline shift from “pages” to “words.”
| Element | REF2021 | REF 2029 |
|---|---|---|
| Format limit | 5 pages (indicative word counts) | Hard 2,200-word cap (3 narrative sections) |
| Suggested section split | Not formally specified | Summary ~100 / Underpinning research ~600 / Details of impact ~1,500 |
| Underpinning research quality threshold | 2* (“recognised internationally”) | Removed — must only meet the REF definition of research |
| Minimum ICS for smallest units (<9.99 FTE) | 2 | 1 (2 optional) |
| References to research | Maximum 6, no word limit | Maximum 6, no word limit (unchanged) |
| Corroborating sources | Maximum 10 | Maximum 10, max 5 named individual contacts (unchanged) |
| Funded-research data fields | Funder, grant number, amount | Adds ROR identifier and ORCID for each named researcher |
How many impact case studies does each unit need to submit?
The number of required ICS still scales with a submitting unit’s volume measure (FTE), but the smallest bracket now has a choice. Units below 9.99 FTE may submit either one or two ICS — Research England “encourages” two but permits one where submitting two would place undue pressure on individual staff in very small units.
- Up to 9.99 FTE: 1 or 2 ICS (submitter’s choice; REF2021 required 2)
- 10–19.99 FTE: 2 ICS
- 20–39.99 FTE: 3 ICS
- 40–59.99 FTE: 4 ICS
- 60–89.99 FTE: 5 ICS
- 90–119.99 FTE: 6 ICS
- 120–169.99 FTE: 7 ICS
- 170+ FTE: 8, plus one further ICS per additional 50 FTE
Submitting fewer than the required number results in an “unclassified” grade for each missing case study. Units cannot submit more than the required number — institutions must select their strongest evidenced examples rather than represent the full spread of unit activity.
What’s different about the evidence and underpinning research standard?
The single biggest substantive change is removal of the 2* quality threshold. Under REF2021, underpinning research had to meet a quality bar “recognised internationally in terms of originality, significance and rigour.” REF 2029 guidance states this requirement was “burdensome, exclusive of valuable local research and impact and limiting to the diversity of the underpinning research” — it now only needs to satisfy the general REF definition of research.
REF 2029 also explicitly invites narrative discussion of engagement strategies and responsible research practices, addressing feedback that units were previously reluctant to describe engagement-led impact because REF2021’s template did not name it as a valid evidence category, even though panels rewarded it when present.
Corroboration rules are largely unchanged: institutions may cite up to 10 external sources and name up to five individual contacts, and URLs may only verify claims already made in the text — panels will not follow links to gather supplementary evidence.
Does the REF 2029 impact case study template change?
Structurally, no — REF 2029 “will maintain the REF 2021 submission template structure for ICS.” The three narrative fields, the references section (max six outputs) and the sources-to-corroborate section (max ten) carry over unchanged. Two additions are new:
- Diverse contributor recognition: an optional, unassessed field to name students, engagement professionals and research-technical staff who contributed to the impact — published in the ICS database with consent.
- Expanded funded-research data: units must supply the funder’s ROR identifier, ORCID per named researcher, funding programme, grant number, grant amount (GBP), formal partners and the impact’s country/nation, where research was externally funded. This supports funder post-assessment analysis, not panel scoring.
Eligibility windows reset for the new cycle: impact must occur between 1 August 2020 and 31 July 2028, underpinned by research first made public between 1 January 2008 and 31 December 2028. Unlike REF2021, there is no COVID extension window — impacts submitted under REF2021’s optional COVID window cannot be resubmitted except as context for a continuing case study.
Answer-first Q&A
How many impact case studies are needed for REF 2029?
The number is set by a unit’s volume measure (FTE), from a minimum of one for units under 9.99 FTE up to eight-plus-one-per-additional-50-FTE for units of 170 FTE or more. REF2021 required a minimum of two; REF 2029 allows the smallest units to submit just one.
How much is a REF impact case study worth?
Impact carries a 25% weighting in the REF quality profile. Per a 2023 Research Professional analysis cited by the LSE Impact of Social Sciences blog, a single 4-star ICS from a very large Unit of Assessment could be worth over £2 million in QR funding across a REF cycle — illustrating why evidence quality, not volume, drives strategy.
What is an impact case study for REF?
It is a structured written submission, entered via the REF Submission system, describing a specific research-derived effect on the economy, society, culture, policy, health or environment. Each case study cites underpinning research outputs and external evidence corroborating the claimed reach and significance.
What is the REF 2029 definition of impact?
REF 2029 defines impact as an effect on, change to, or benefit for the economy, society, culture, public policy, health, the environment or quality of life beyond academia — including harm reduction — occurring in any geographic location. Purely academic impacts on research knowledge are explicitly excluded from this element.
What this means for research administrators
The practical effect is a lower barrier to impact stories built on locally significant or applied research that would previously have failed the 2* threshold, plus more room to credit engagement work and professional-services contributors often invisible in REF2021 submissions. The trade-off is stricter word discipline — 2,200 words is a hard, system-enforced limit, not a page guideline that can be stretched with formatting.
Institutions should build ROR and ORCID capture into impact-tracking workflows now: these identifiers become mandatory for any ICS underpinned by externally funded research, and retrofitting historic funding records close to the deadline is a known source of REF-cycle administrative burden.
REF 2029 guidance was published in draft form from December 2025 and is scheduled for formal finalisation during 2026, so submitting units should track Research England’s published change log rather than treat any single module as final until then.
