QS vs Times Higher Education: What Each Ranking Family Actually Measures

Side-by-side breakdown of QS and THE indicators, weights and reputation mechanics, plus ARWU, so their positions can be read without false equivalence.

QS vs Times Higher Education: What Each Ranking Family Actually Measures

QS and Times Higher Education can produce different orders because they are not interchangeable measurements of institutional performance. Each position summarizes a selected set of indicators, weights and data procedures. Shared labels such as reputation, citations or international do not make those quantities equivalent. The comparison must therefore move below the overall place and align the actual measurements: what is observed, how much weight it carries and how the data are obtained.

The scoring architectures are not the same

QS calls its groupings “Lenses” and its measurements “Indicators.” Research and Discovery carries 50%, divided into Academic Reputation at 30% and Citations per Faculty at 20%. Employability and Outcomes carries 20%, split between Employer Reputation at 15% and Employment Outcomes at 5%. Learning Experience carries 10% and consists of Faculty Student Ratio. Global Engagement carries 15%, with International Faculty Ratio, International Research Network and International Student Ratio at 5% each. Sustainability carries 5%. These categories define the published architecture described in the QS methodology guide.

THE describes its universe as research-intensive and organizes the score around five pillars: Teaching at 29.5%, Research Environment at 29%, Research Quality at 30%, International Outlook at 7.5% and Industry at 4%. The shared label reputation does not mean the same input: in QS, academic and employer reputation sit under Research and Discovery and Employability and Outcomes, while THE places teaching and research reputation inside their respective pillars. Likewise, citations occupy a defined QS indicator but sit within THE’s broader research-quality structure.

Reputation is a major input in both

QS’s Academic Reputation carries 30% and Employer Reputation carries 15%. Both are obtained through reputational surveys whose contact lists institutions supply; an institution can submit up to 400 contacts for each survey, and QS distributes the surveys to those selected contacts. QS says respondents assess the reputation of global institutions, not necessarily only the institutions that nominated them. QS also describes its ranking as the only ranking that includes employer reputation. In the flagship rankings, it says it assesses over 8,000 institutions in a usual cycle but ranks only around 1,500. The survey process is documented in QS’s participation guidance.

THE assigns 15% to Teaching Reputation and 18% to Research Reputation. It combines two survey years to give more than 108,000 responses and adds a measure based on the number of institutions whose academics vote for a university. A large response total does not make the systems equivalent: the candidate pool, institutional coverage and the way reputation enters each score remain separate parts of the measurement.

The nomination mechanism also makes data provenance important. QS makes clear that institutions help supply the contacts receiving its surveys. That does not establish whether the resulting opinions are accurate or biased, but it does show why reputation scores cannot be understood independently of the survey population that produced them. THE’s much larger reported response volume should be read with the same methodological attention.

Teaching enters through different proxies

QS does not operate a teaching pillar. Its Learning Experience allocation is 10%, represented only by Faculty Student Ratio. This is a staffing-load measure relative to enrolment; it does not directly include teaching reputation, doctoral activity or institutional income. A change in that ratio therefore says more about the relationship between students and faculty than about the demonstrated quality of teaching.

THE gives Teaching a dedicated 29.5% pillar, while Research Quality carries 30%. Teaching comprises Teaching Reputation at 15%, Staff-to-student ratio at 4.5%, Doctorate-to-bachelor’s ratio at 2%, Doctorates-awarded-to-academic-staff ratio at 5.5% and Institutional income at 2.5%. Perception thus operates alongside staffing, relative degree-award volumes, doctoral output per academic staff and institutional finance. THE devotes a composite domain to teaching; QS uses a much narrower resource-ratio proxy. The complete allocation is set out in THE’s World University Rankings methodology.

Research volume, quality and income are separated differently

QS assigns 20% to Citations per Faculty. It uses an Elsevier Scopus extract and scales citation counts by faculty headcount, relating published citation volume to staffing. The measure compresses citation patterns into a staffing-adjusted total rather than describing the whole research environment, and it sits beside Academic Reputation within Research and Discovery.

THE separates Research Environment at 29% from Research Quality at 30%. Research Environment contains Research Reputation at 18%, Research Income at 5.5% and Research Productivity at 5.5%. Research Quality contains Citation Impact at 15%, Research Strength at 5%, Research Excellence at 5% and Research Influence at 5%. This separates reputation, income and productivity from the citation-related assessment of quality.

THE says Research Income is scaled against academic staff and adjusted for PPP, while calling it controversial because national policy and economic circumstances can influence it. That qualification does not establish that the indicator is useless. It identifies a structural difference between measuring research performance and measuring financial activity shaped by a national setting.

International reach and industry money receive different treatment

QS gives Global Engagement 15%. International Faculty Ratio, International Research Network and International Student Ratio each carry 5%, so international research links have a direct place in the overall structure alongside staff and student composition. THE gives International Outlook 7.5%, divided into Proportion of international students at 2.5%, international staff at 2.5% and international collaboration at 2.5%. THE normalizes collaboration by subject mix, stating that large countries have otherwise been disadvantaged compared with small countries. The systems therefore ask different international questions: research networks and institutional composition in QS, and staff, students and normalized collaboration in THE.

Industry creates another divergence. THE’s Industry pillar carries 4%, split between Industry Income at 2% and Patents at 2%. Its industry-income metric is adjusted for PPP and scaled against academic staff, with the stated purpose of capturing knowledge-transfer activity. QS includes no industry-income indicator. THE’s criticism of research income should not be confused with a claim that all research activity has equal value; it explains why an income measure tied to national policy and economic circumstances can complicate cross-institution comparison.

Shanghai and ARWU prove a third construction exists

ShanghaiRanking’s Academic Ranking of World Universities provides a third family rather than a variation of either framework. ARWU ranks through academic and research prizes and papers: Nobel laureates and Fields medalists among alumni or staff, Highly Cited Researchers selected by Clarivate, papers published in Nature or Science, and a significant volume of papers indexed in SCIE and SSCI in Web of Science. It has no reputation survey, no teaching measure and no international indicator. More than 2500 universities are actually ranked, and the best 1000 are published. That construction, described in the ARWU methodology, is direct evidence that ranking families are not interchangeable products.

How to read a disagreement

The three families measure different constructs, so their positions cannot be compared as if they shared a common scale. A high place in one does not correct, validate or invalidate a lower place in another. Cross-system agreement is not evidence that a publisher is more accurate, and disagreement is not evidence that either publisher has made an error.

When two tables disagree, the disagreement is information about the measurement, not about the universities. Read indicator-level scores and definitions rather than relying on overall positions. Identify whether a result comes from reputation, citations, staffing, income, collaboration or prizes, then consider the weight, data source and treatment of that indicator. A ranking position is an ordinal summary of a particular construction; it is not a universal measurement of institutional quality that remains unchanged when the construction changes.