South Korea is in the middle of an unresolved argument about how it evaluates research, and the argument is becoming harder to ignore. Domestic universities publish heavily, but a widening body of evidence — funder data, independent bibliometrics, and a direct survey of researchers — shows that publication volume is diverging sharply from measured research quality and return on investment. The debate is not settled and there is no single finalized national policy to report; what exists is a real, intensifying conversation among funders, universities, and publishers about whether Korea’s evaluation culture needs to move away from counting papers.
The forum that named the problem
Springer Nature convened the inaugural Korea Research Advisory Forum (KRAF) on 20 August 2024, joining similar advisory councils it already runs in the US, Japan, Africa, and Europe. Participants included representatives from the Korean ARPA-H Initiative, Seoul National University, POSTECH, Korea University, Yonsei University, the Institute for Basic Science (IBS), and GISTeR, with Nature editor-in-chief Magdalena Skipper and Springer Nature VP Nick Campbell in the room.
The forum’s stated concern was a mismatch between input and output: South Korea’s R&D spending is high relative to peer economies, but measured research output and impact have not kept pace. Its recommendations pointed toward qualitative evaluation methods that weight social impact, innovation, and interdisciplinary work more heavily than raw publication counts, alongside commercialization metrics — patents, licenses, spinouts — sitting next to academic ones rather than being treated as secondary.
What the numbers actually show
The case for reform is easiest to see in the gap between how many papers Korean universities publish and how much those papers are cited. Leiden University’s Centre for Science and Technology Studies (CWTS) rankings, which report both raw publication counts and share of papers in the global top 10% by citation, show a consistent pattern: Seoul National University ranks 21st globally by paper count but falls to 74th by top-10%-cited share. The same drop repeats across other flagship institutions — Yonsei University (60th by volume, 122nd by top-10% share), Sungkyunkwan University (92nd to 199th), and Korea University (103rd to 209th).
A survey of 130 professors conducted by the Korean Academy of Science and Technology found that roughly half — about 60 respondents — identified sheer paper count as the primary factor driving research evaluation at their institutions, ahead of citation impact, translational outcomes, or teaching.
The National Research Foundation of Korea (NRF), the country’s principal research-funding agency, has itself published data feeding the debate: Korean universities’ R&D investment recovery rate — essentially, commercial and licensing return relative to R&D spend — stood at 1.62% as of 2024, well below the roughly 3.5% reported for US universities and roughly 2.1% for UK universities. That figure has become a recurring reference point for administrators arguing that a paper-count-centered evaluation structure is misallocating incentives away from commercializable, industry-relevant research.
What reform would actually change
None of this amounts to a finalized NRF regulation as of this writing — the debate is playing out through funder data releases, advisory forums, institutional survey results, and domestic reporting (Korean outlets including the Seoul Economic Daily were still framing this as an open “numbers race” as recently as mid-2026), not through a single announced policy. But the direction under discussion is consistent across venues: less weight on raw publication and citation counts in tenure and grant-renewal decisions, more explicit weight on commercialization outcomes, and calls for evaluation criteria that are differentiated by field and by a university’s regional industrial role rather than applied uniformly.
That direction puts Korea in the same conversation as the international DORA and CoARA movements, which press institutions to stop using journal-level metrics like the impact factor as a proxy for the quality of an individual researcher’s work. CASRAI’s comparison of DORA and CoARA’s commitments and its guide to how the two frameworks compare on paper cover the international baseline Korean institutions are implicitly being measured against; a companion guide looks at how DORA/CoARA signatories are actually rewriting promotion-and-tenure criteria in practice, which is the operational question Korean universities now face if the KRAF-style recommendations gain traction.
South Korea is not alone in having built extensive publication-metrics infrastructure and now needing to decide how much weight to keep giving it. Korea’s own Korea Citation Index (KCI) and the KRI researcher database are the domestic evaluation backbone this debate is really about — the question is not whether that infrastructure disappears, but how the metrics it produces get weighted against qualitative and commercialization measures going forward. Taiwan has run a parallel version of this conversation in the humanities and social sciences through its own Taiwan Citation Index (TCI-HSS), for institutions looking at how a neighboring system has approached the same quantity-versus-quality tension.
What to watch
Nothing here should be read as an announced regulatory change. What’s verifiable right now is: a Springer Nature-convened advisory forum recommending a shift toward qualitative and commercialization-inclusive evaluation; NRF’s own R&D-recovery figures being cited as evidence for that shift; independent bibliometric data (Leiden CWTS) and a domestic professor survey both pointing at the same quantity-quality gap; and continuing domestic press coverage through at least mid-2026 treating the “numbers race” as unresolved. Institutions and funders outside Korea tracking responsible-assessment reform globally should treat this as an active, not a closed, case — and watch for whether NRF or Korea’s Ministry of Science and ICT (MSIT) issues formal evaluation-criteria guidance building on these recommendations.
