The OECD’s new working paper, “New expectations and demands from science: Rethinking research assessment frameworks,” maps the tensions, actors, and drivers reshaping how research is judged worldwide. Published as OECD Science, Technology and Industry Working Papers, No. 2026/07, it is a companion to the OECD’s flagship policy brief “Reforming research assessment for better science” — and it does a different job: rather than prescribing solutions, it diagnoses why current assessment systems are misaligned with what science is now expected to deliver.
New expectations and demands from science is the OECD’s shorthand for a widening gap: funders, governments, and the public increasingly expect research to be open, collaborative, and socially useful, while most assessment frameworks still reward narrow, quantitative output counts. The working paper is a system-level literature review — it identifies who holds power over assessment criteria, what forces are pushing reform, and where the friction actually sits.
- What the OECD’s companion paper actually says
- How it differs from the flagship OECD report
- Mapping the actors and tensions in research assessment
- Answer-first Q&A: what people are asking
- Implications and what comes next
What the OECD’s companion paper actually says
The working paper is a 43-page system-level overview, not a set of new rules. OECD (2026) states that research assessment frameworks “play a central role in shaping the priorities, direction, and culture of scientific research,” but that they have grown misaligned with “evolving policy priorities, public expectations, and new demands from science.” The core diagnosis is over-reliance on narrow performance measures — publication counts, journal impact factors, and citation metrics — which generates what the paper calls “perverse incentives and undesirable behaviours.”
Critically, the paper argues these narrow measures systematically undervalue activities that funders and the public now expect: collaboration across disciplines and borders, open science practices, societal engagement, and direct support to policymaking. The paper does not stop at critique — it distils a set of common reform principles from a comparative review of the international literature, intended to help policymakers and institutions design frameworks that better reflect these expectations.
How it differs from the flagship OECD report
The two documents were released as a pair but serve distinct purposes, and treating them as interchangeable misses the news value of the working paper. The flagship document — “Reforming research assessment for better science,” OECD Policy Briefs, No. 56 — is the prescriptive, policymaker-facing output: a short brief calling for balanced, cost-effective, transparent assessment approaches “supported by open data and carefully governed use of AI.”
“New expectations and demands from science” is the underlying evidence base. It is longer, more academic in register, and organised around system mapping rather than recommendations. Where the policy brief tells institutions what to do, the working paper explains why the system is under strain and who the competing actors are — making it the more useful read for anyone designing an institutional assessment policy rather than just citing OECD guidance.
| Assessment dimension | Traditional practice (narrow metrics) | OECD’s identified shift |
|---|---|---|
| Research outputs recognised | Publications, citation counts, journal impact factor | Datasets, software, policy contributions, teaching, public engagement |
| Data sources | Proprietary bibliometric databases | Open, interoperable data infrastructure |
| Collaboration | Individual authorship credit only | Team science and cross-sector collaboration valued explicitly |
| Societal role | Largely absent from formal criteria | Societal and policy impact incorporated |
| Use of AI in evaluation | Ad hoc, ungoverned | Carefully governed, transparent use |
Mapping the actors and tensions in research assessment
The working paper’s distinguishing contribution is its system-level actor map — it names who sets, applies, and is judged by assessment criteria, and where their interests conflict. This is the part a policy brief cannot do in a few pages.
- Funders and governments, who set the policy priorities that assessment frameworks are meant to serve but often lag them in criteria design.
- Universities and research institutions, which apply assessment for hiring, promotion, and tenure and are often the slowest layer to change.
- Publishers and indexing services, whose proprietary metrics (as flagged in expert commentary on the paper, including from bibliometrics researcher Ludo Waltman of Leiden University’s CWTS) still dominate despite the OECD’s call to shift toward open alternatives.
- Individual researchers, whose career incentives are shaped by all of the above and who bear the practical cost of misalignment.
- Research infrastructure and standards bodies, which build the open data and interoperable systems needed to support broader, fairer assessment criteria.
This tension between what funders say they want and what institutional reward systems actually measure is not a new observation — the Coalition for Advancing Research Assessment (CoARA), building on the European Commission’s Agreement on Reforming Research Assessment finalised 20 July 2022, has been pushing the same reform agenda for four years. What the OECD paper adds is a comparative, OECD-wide synthesis rather than a Europe-centred coalition commitment, giving research administrators outside the EU a reference point that isn’t tied to CoARA membership.
The timing also lines up with parallel European activity: Science Europe published its own position statement, “Connecting Open Science and Research Assessment Reform,” in April 2026, arguing that open science advances and assessment reform are “mutually reinforcing and inter-dependent drivers of research cultures.” Read together, the OECD working paper and the Science Europe statement show the reform agenda converging on the same point from two different institutional angles — global policy synthesis versus funder-coalition advocacy.
Answer-first Q&A: what people are asking
What is the OECD’s “New expectations and demands from science” paper?
It is OECD Science, Technology and Industry Working Paper No. 2026/07, published 29 April 2026 as a companion to the OECD’s policy brief “Reforming research assessment for better science.” It provides a system-level literature review mapping the actors, tensions, and drivers behind global research assessment reform, without itself issuing binding recommendations.
Why does this OECD paper matter for research administrators?
It gives research administrators and institutional leaders a non-EU-specific, evidence-based reference for redesigning hiring, promotion, and funding-review criteria. Because it maps competing actor interests explicitly, it is more useful for internal policy justification than a short recommendations-only brief.
What does the OECD say about quantitative indicators in research assessment?
The paper identifies over-reliance on narrow quantitative indicators — publication counts, citation metrics, journal impact factors — as the central structural problem, arguing it produces perverse incentives and undervalues collaboration, openness, and societal engagement that funders now expect.
How does this relate to the CoARA reform agreement?
CoARA’s Agreement on Reforming Research Assessment, finalised by the European Commission on 20 July 2022, is a European funder-and-institution coalition commitment. The OECD’s 2026 paper covers similar ground but at OECD-wide scope, functioning as an evidence synthesis rather than a signatory pledge.
Implications and what comes next
For research administrators, the practical takeaway is not to wait for a single global standard. The OECD paper’s actor map is a useful diagnostic tool for institutions auditing their own promotion and funding-review criteria against the gap between stated priorities (openness, societal impact, collaboration) and what is actually measured (publication counts and journal placement).
The convergence of OECD, CoARA, and Science Europe positions in 2026 suggests assessment reform is moving from advocacy toward implementation detail — governance of AI in evaluation, and the shift away from proprietary bibliometric data, are likely to be the next flashpoints. Institutions building or revising assessment frameworks, including those documenting contributor roles through standards such as CRediT, should treat this OECD synthesis as a system-level map to check institutional policy against, not a checklist to copy verbatim. For teams working through the practical mechanics of research administration and assessment criteria, CASRAI’s research administration resources track how these standards intersect with day-to-day institutional practice.