What the Leiden Ranking Is
The CWTS Leiden Ranking is an annual bibliometric ranking of universities produced by the Centre for Science and Technology Studies (CWTS) at Leiden University in the Netherlands, first published in 2007.
Its defining stance is methodological restraint: the ranking reports research performance indicators only, based on Web of Science data, and famously declines to combine them into a single overall score.
What the Leiden Ranking Ranks
The ranking orders universities worldwide by research output and impact — institutions above a publication threshold in the Web of Science database, roughly 1,500 in recent editions.
Rankings can be viewed by broad scientific field, acknowledging that comparing a medical school with a technical university across all fields is misleading.
Core Inputs Used by the Leiden Ranking
All indicators derive from Web of Science publications:
- Publication output — number of indexed publications
- Top-percentile impact — the share of a university’s publications in the global top 1%, 5%, 10%, or 50% most cited in their field and year
- Collaboration metrics — share of publications with international, industrial, or short/long-distance co-authors
- Open access metrics — share of publications available open access
How the Leiden Ranking Works (High-Level)
There is deliberately no composite calculation:
- Publications are attributed to universities using full counting by default (each co-authoring institution gets credit).
- Citation impact is field- and year-normalized — a physics paper and a sociology paper are compared against their own field’s citation distribution.
- Users select an indicator (e.g., “share of top 10% publications”) and the table re-sorts by it. Different indicators produce different orders — which is the point.
Conceptual model: Not a league table but an instrument panel — every dial is field-normalized, and no dial claims to be “the” ranking.
Key Design Choices
- No composite score — composite indices, CWTS argues, hide arbitrary weights behind a false precision
- Field normalization — citations are benchmarked within fields and publication years
- Fractional counting option — credit can be split among co-authoring institutions to reduce collaboration inflation
- Size-dependent and size-independent views — raw counts and proportional measures are both offered
Update Frequency
The Leiden Ranking is published annually, each edition covering a rolling multi-year publication window from Web of Science.
Known Limitations and Criticisms
- Bibliometrics only — teaching, innovation, and institutional quality are out of scope by design
- Database coverage — Web of Science underrepresents non-English and non-journal output
- Less media-friendly — without a single ranking, press uptake is weaker than composite tables
- Attribution choices matter — counting methods and threshold settings still shape results, even without a composite
Where the Leiden Ranking Is Used
The ranking is used for:
- Research policy and benchmarking by universities and governments
- Bibliometric research on citation-based evaluation
- “Responsible metrics” initiatives, which cite CWTS work as a model
- Analysts who want field-fair comparisons without survey noise
Summary
The Leiden Ranking is the methodologist’s ranking: field-normalized citation indicators, transparent counting rules, and a principled refusal to boil universities down to one number. It answers a narrower question than QS or ARWU — and answers it with unusual care.
References and Sources
- CWTS Leiden University. Leiden Ranking — Methodology (official documentation).
- Wikipedia. CWTS Leiden Ranking.
- Waltman, L., et al. — CWTS publications on responsible research metrics.