“Research excellence” is not itself a formal standard with one defined meaning — it is the umbrella term research institutions, funders, and national assessment bodies use to describe research that is judged to be original, rigorous, and significant. In practice, the phrase has two related but distinct uses: a general institutional aspiration (a university’s strategic plan promising to “pursue research excellence”), and a formally assessed, criteria-based outcome produced by a national research-assessment exercise, the best-known of which is the UK’s Research Excellence Framework (REF).
This guide covers both senses, how the major national frameworks operationalize “excellence” into scoreable criteria, and the growing reform movement — anchored by DORA, CoARA, and the Leiden Manifesto — that pushes back against reducing “excellence” to citation-based metrics alone.
The two senses of “research excellence”
Used loosely, research excellence describes a quality standard: work that is methodologically sound, makes an original contribution, and is significant to its field. Used formally, it describes the output of a structured assessment process with defined criteria, expert panel review, and — often — funding consequences. The second sense is where research administrators most need precision, because a national exercise’s specific scoring criteria determine how an institution prepares submissions, allocates internal resources, and reports outcomes to leadership.
How national research-assessment frameworks define it
Several countries run, or have run, a formal national exercise that assesses university research and ties the result to a portion of government block-grant funding. They differ in mechanics, but most share a similar structure: a mix of expert peer review and metrics, applied to research outputs, plus some accounting for wider impact and the environment that produced the work.
REF (United Kingdom)
The Research Excellence Framework is the most fully developed and most-referenced example. Run jointly by Research England, the Scottish Funding Council, the Higher Education Funding Council for Wales, and the Department for the Economy Northern Ireland, it succeeded the earlier Research Assessment Exercise (RAE) with the first REF cycle held in 2014. Each submitted unit of assessment is scored by discipline-specific expert panels on three components: outputs (the quality of publications and other research outputs), impact (documented effects of the research beyond academia, evidenced through impact case studies), and environment (the strategy, resources, and infrastructure supporting the research). Each element is graded on a five-point scale — 4*, 3*, 2*, 1*, and unclassified — with 4* representing work judged world-leading in originality, significance, and rigor. Results directly affect the block-grant research funding UK universities receive. See the site’s dedicated entries on REF 2021 and the upcoming REF 2029 cycle for the specifics of each assessment period.
ERA (Australia) — discontinued, replacement in development
Excellence in Research for Australia (ERA), run by the Australian Research Council, assessed the quality of research across Australian universities in rounds held in 2010, 2012, 2015, and 2018. The planned 2023 round did not proceed: the Australian Minister for Education announced ERA’s discontinuation in August 2022, citing the administrative burden of the exercise. The Australian Universities Accord’s final report (2024) recommended developing a future approach to national research evaluation, and the Australian Research Council has been working on a proposed replacement since. Anyone researching ERA should treat it as a discontinued exercise, not an active one, as of this writing.
Other national exercises
New Zealand’s Performance-Based Research Fund (PBRF), administered by the Tertiary Education Commission, assesses research quality at the level of individual researchers via periodic Quality Evaluations, informing a portion of tertiary funding. Italy’s Valutazione della Qualità della Ricerca (VQR), run by ANVUR, is a similar national research-quality evaluation tied to public university funding. Both follow the same general logic as REF and the former ERA — expert or metric-informed review of outputs, translated into a funding-relevant score — even though the specific criteria, cycle length, and scoring mechanics differ by country. Because the details of each scheme change between cycles, research administrators working with a specific non-UK framework should confirm current requirements directly against that scheme’s own current guidance rather than relying on a general description.
What the criteria actually reward
Across these frameworks, three components recur, even where the exact labels differ:
- Output quality — originality, rigor, and significance of individual publications or other research outputs, judged by expert peer review (not solely by citation counts or journal placement).
- Impact or engagement — documented, evidenced effects of the research beyond academia: on policy, practice, industry, public understanding, or health and social outcomes.
- Environment — the strategy, infrastructure, staffing, and support structures that make the research possible and sustainable.
This three-part structure matters operationally: an institution preparing a REF-style submission is not simply selecting its most-cited papers. It is assembling evidence across all three dimensions, which is why research-assessment preparation is a cross-institutional exercise involving research offices, individual researchers, and impact-evidence gathering well before a submission deadline.
The metrics critique: DORA, CoARA, and the Leiden Manifesto
A parallel, and partly overlapping, conversation concerns how “excellence” gets measured when formal national exercises are not in play — specifically, the use of journal-level metrics like the Journal Impact Factor as a proxy for the quality of an individual researcher, article, or grant application in hiring, promotion, and funding decisions. Three reference points anchor the reform movement pushing back on that practice:
- DORA (San Francisco Declaration on Research Assessment, 2012) — a widely signed commitment by institutions, funders, and publishers to stop using journal-based metrics as a stand-in for the quality of individual outputs or researchers.
- CoARA (Coalition for Advancing Research Assessment, 2022) — a broader, EU-originated coalition of research organizations working toward reformed, multidimensional assessment practices that go beyond a single metrics-based fix.
- The Leiden Manifesto (published in Nature, 2015) — ten principles for the responsible use of quantitative indicators in research evaluation, including that quantitative evaluation should support, not replace, qualitative expert assessment, and — notably for this topic — that evaluation should “protect excellence in locally relevant research” rather than only rewarding output that scores well on globally-oriented citation metrics.
The throughline across all three is not that formal assessment or metrics are illegitimate, but that “excellence” reduced to a single number (an impact factor, a citation count) is a poor and sometimes actively misleading proxy for research quality. National exercises like REF partly anticipate this critique by using expert panel peer review rather than pure metrics as the primary assessment mechanism, though metrics still inform panel judgment in practice. For a fuller comparison of what DORA and CoARA each specifically ask signatories to do, see DORA and CoARA: Comparing Research Assessment Reform Commitments.
Why this matters for research administrators
Understanding how “research excellence” is defined and assessed in a given context is not academic trivia — it shapes concrete institutional decisions: which outputs and impact evidence to prepare for a national exercise submission, how tenure and promotion criteria are written (and whether they lean on journal-level metrics DORA and CoARA signatories have committed to move away from), and how research strategy documents set goals that can actually be evidenced later. A research office preparing for a REF-style cycle, or drafting an institutional research-assessment statement informed by CoARA membership, is working with the same underlying tension this guide describes: expert judgment of quality and significance, evidenced impact, and supportive environment, versus the pull toward simpler but less accurate metric shortcuts.
Frequently asked questions
Is “research excellence” the same thing as the REF?
No. The REF is one specific, formally defined UK assessment exercise that uses “Research Excellence Framework” as its name. “Research excellence” more broadly is a general term for research quality that predates and extends beyond that specific exercise — used by universities worldwide in mission statements, by other national assessment schemes with different names, and in general academic usage.
Who decides what counts as research excellence?
It depends on context. Within a formal national exercise like REF, subject-specific expert panels apply published assessment criteria to submitted evidence. Outside a formal exercise, individual institutions, funders, tenure and promotion committees, and disciplinary communities each apply their own (often less formally documented) standards, which is part of why the DORA/CoARA/Leiden Manifesto reform movement exists — to make those standards more consistent and less dependent on metric shortcuts.
Is a high journal impact factor the same as research excellence?
No, and DORA specifically recommends against treating it that way. A journal’s impact factor is a journal-level metric describing average citations to that journal’s articles; it says nothing definitive about the quality of any single article published in it. Signatories of DORA commit to not using journal-based metrics as a substitute for assessing the quality of an individual article, researcher, or grant application.
Does open access affect research-excellence assessments?
Open-access status is not itself a quality criterion in frameworks like REF, but funder and institutional open-access mandates increasingly interact with assessment eligibility — for example, some national exercises have made open-access deposit a condition for an output to be eligible for submission at all. Requirements are cycle-specific and should be confirmed against the current guidance for the relevant exercise.
