In short: our rankings rest on open data, primary documents and expert review. Existing rankings carry a small, capped weight, and we never copy or redistribute other publishers' ranking tables.
Principles
- Verifiable. Every piece of evidence used in an assessment can be traced to a source.
- Open first. We prefer data that is openly licensed, so that others can check our work.
- Lawful. We respect copyright, database rights, licences and website terms.
- Proportionate. No single source, and especially no existing ranking, can dominate a result.
Sources we use
Open data
- OpenAlex, an open catalog of scholarly works, authors and institutions, released under CC0.
- CWTS Leiden Ranking Open Edition, whose results and data files are released under CC0.
- DEQAR, the EQAR Database of External Quality Assurance Results, available under the Open Data Commons Public Domain Dedication and Licence (PDDL). We credit EQAR when we use it.
- ROR, the Research Organization Registry, for consistent institutional identifiers.
Primary and institutional sources
- Institutional publications, such as program pages, curricula, course catalogs, annual reports and published outcome data.
- Accreditation and quality assurance reports published by recognized agencies.
- Official registers used for eligibility checks, as described in the Eligibility Policy.
Material sent to us
We do not solicit submissions from the entities we rank. Material sent to us nonetheless is verified with its source before any use, and is handled under the Terms of Use and the Privacy Notice.
Research literature
Assessors and editors draw on peer-reviewed research about education, AI, and the fields we rank, both to design indicators and to interpret evidence.
Use of existing rankings
Established rankings capture years of reputation and data. Ignoring them entirely would discard useful information; relying on them heavily would make RankingLab dependent and slow. Our approach is a single, capped pillar called the External Reference Signal.
Weight and cap
| Setting | Value |
|---|---|
| Default weight | 5% of the overall score |
| Hard cap | 10% of the overall score |
| If no permissible source exists | 0%, with the weight redistributed proportionally to the other pillars |
The cap cannot be exceeded in any ranking or edition.
When a source is permissible
We use another organization's ranking only where at least one of the following applies:
- The ranking data is published under an open licence that permits reuse, such as the CC0 release of the Leiden Ranking Open Edition.
- We have obtained written permission or a licence from the publisher.
- The use is limited to factual references that are consistent with the publisher's published terms.
What we never do
- Reproduce, republish or redistribute third-party ranking tables.
- Scrape websites in breach of their terms of use.
- Extract a substantial part of a protected database.
- Use third-party names or trademarks except to identify a source.
- Suggest that any ranking publisher endorses or participates in RankingLab.
Why these limits exist
Ranking publishers protect their work in different ways. For example, the QS terms for topmba.com require prior written consent to reproduce content and do not allow the data to be reconfigured. The Times Higher Education terms direct requests to reproduce or adapt rankings data to their data team and protect their trademarks. In the European Union, the Database Directive (96/9/EC) gives database makers a sui generis right against extraction of substantial parts of a database. The IAU does not allow copying all or substantial parts of WHED without its consent. We design our process to respect these rules. This summary is not legal advice; we review it with counsel as terms change.
Data quality
- Figures are checked against original documents on a sample basis every edition.
- Anomalies flagged by automated checks are investigated by people before use.
- Where evidence is weak or conflicting, assessors record lower confidence, and this is reflected in the published uncertainty bands described on the Methodology page.
Our own data
Unless stated otherwise, RankingLab text and ranking data are published under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). Third-party material keeps its own licence. See the Citation, Name & Badge Use Policy.
Requests from data owners
If you own data or content and believe we have used it outside the terms above, please contact hi@rankinglab.org. We will review the concern promptly and remove or correct material where appropriate.