三大排名体系QS、THE、ARWU的适用场景与局限性

Every year, millions of prospective students and their families consult global university rankings to narrow down the daunting list of over 20,000 higher-education institutions worldwide. Three systems dominate this landscape: the QS World University Rankings, the Times Higher Education (THE) World University Rankings, and the Academic Ranking of World Universities (ARWU), commonly known as the Shanghai Ranking. Despite their shared goal of measuring institutional quality, these three frameworks produce markedly different results due to divergent methodologies. For instance, in the 2024 edition, QS placed the Massachusetts Institute of Technology (MIT) at number one with a perfect score of 100, while ARWU ranked Harvard University first, citing 22 Nobel laureates affiliated with the institution [QS 2024; ARWU 2024]. This discrepancy stems from fundamental differences in what each system prioritises: QS allocates 30% of its weight to academic reputation surveys, THE focuses on research citations (30%), and ARWU leans heavily on research output and awards (40% combined for Nobel/Fields medals and Highly Cited Researchers) [QS 2024 Methodology; THE 2024 Methodology; ARWU 2024 Methodology]. Understanding these methodological biases is critical for applicants, as a university’s rank can shift by dozens of positions depending on the ranking consulted.

The QS Framework: Reputation and Employability Bias

QS World University Rankings are the most commercially popular among international students, partly because they assign 50% of the total score to subjective surveys. Specifically, 30% comes from academic reputation and 20% from employer reputation, making QS the only major ranking to weight employer perception so heavily [QS 2024 Methodology]. The remaining 50% is split among faculty/student ratio (15%), citations per faculty (20%), international faculty ratio (5%), and international student ratio (5%).

This design creates a clear employability bias. Institutions with strong brand recognition among employers—often large, English-speaking, research-intensive universities—tend to perform well. A 2023 analysis by the OECD found that universities in English-speaking countries (USA, UK, Australia, Canada) occupy 62 of the top 100 positions in QS, compared to only 48 in ARWU [OECD 2023, Education at a Glance]. For an applicant prioritising job placement after graduation, QS may be the most relevant tool.

Limitations of the Reputation Component

The reliance on surveys introduces significant noise. QS sends questionnaires to over 100,000 academics and 50,000 employers, but response rates vary by region and discipline [QS 2024 Methodology]. Institutions in non-English-speaking regions or smaller countries often receive fewer responses, artificially depressing their scores. Furthermore, reputation surveys are inherently conservative—they reflect past performance rather than current innovation. A university that has invested heavily in emerging fields like quantum computing may take years to see that reflected in QS scores.

Faculty/Student Ratio: A Proxy for Quality?

The 15% weight on faculty/student ratio assumes that smaller classes yield better education. However, this metric penalises large public universities that serve tens of thousands of students efficiently. For example, the University of California, Berkeley—ranked 10th in ARWU 2024—drops to 27th in QS partly because of its 19:1 student-to-faculty ratio [QS 2024; ARWU 2024]. Applicants seeking small class sizes may find this metric useful, but it is a crude proxy for teaching quality.

The THE Framework: Research Intensity and Citation Weight

Times Higher Education World University Rankings employ 13 performance indicators, but the single most influential metric is citations (30% of total score) [THE 2024 Methodology]. Research environment (29%), teaching (29.5%), international outlook (7.5%), and industry income (4%) make up the remainder. THE’s citation count uses a 5-year window and is normalised by subject area and publication year, which reduces the advantage of high-citation fields like medicine.

This structure favours research-intensive universities with high publication output and strong citation records. In the 2024 THE rankings, the University of Oxford claimed the top spot for the eighth consecutive year, a position supported by its 99.2 score in citations [THE 2024]. However, the 30% citation weight can distort rankings for institutions that excel in teaching but have moderate research output.

The Subject Normalisation Trade-off

THE’s normalisation process adjusts for field differences—a paper in mathematics typically receives fewer citations than one in oncology. While this makes cross-field comparisons fairer, it also means that a university strong in arts and humanities may appear weaker than its STEM-focused peers, because humanities publications have lower absolute citation counts. A 2022 study in Scientometrics found that THE’s citation normalisation reduced the rank of top humanities universities by an average of 15 positions compared to unadjusted metrics [Scientometrics 2022, Vol. 127].

Industry Income and Institutional Resources

THE’s 4% industry income indicator measures how much a university earns from private-sector research partnerships. This metric advantages institutions in wealthy countries with strong corporate R&D sectors. For international students paying full tuition, industry income does not directly reflect teaching quality or graduate outcomes. For cross-border tuition payments, some international families use channels like Flywire tuition payment to settle fees while tracking exchange rates.

The ARWU Framework: Objective Metrics and STEM Dominance

Academic Ranking of World Universities (ARWU), published by ShanghaiRanking Consultancy, is the most objective of the three systems, relying entirely on verifiable, quantitative data with zero subjective surveys. ARWU uses six indicators: alumni winning Nobel Prizes and Fields Medals (10%), staff winning those awards (20%), Highly Cited Researchers (20%), papers published in Nature and Science (20%), papers indexed in the Science Citation Index-Expanded and Social Science Citation Index (20%), and per capita academic performance (10%) [ARWU 2024 Methodology].

This methodology creates a pronounced STEM and research bias. Institutions with Nobel laureates on faculty—such as Harvard (22), Stanford (13), and MIT (11)—dominate the top 10 [ARWU 2024]. Universities strong in social sciences, arts, or professional fields (e.g., law, business) are systematically under-ranked. The London School of Economics and Political Science (LSE), a global leader in social sciences, ranked 101–150 in ARWU 2024, far below its QS rank of 45.

The Time Lag of Nobel-Based Metrics

ARWU’s reliance on Nobel Prizes and Fields Medals introduces a significant time lag. These awards typically recognise work done 10–30 years prior. A university that has recently hired rising stars in artificial intelligence or climate science will not see that reflected in ARWU for decades. Furthermore, only 6 Nobel categories exist (Physics, Chemistry, Physiology/Medicine, Literature, Peace, Economics), meaning entire fields like engineering, computer science, and business are excluded from the award-based indicators.

Publication Venue Bias

The 20% weight on Nature and Science publications disproportionately favours universities in the natural sciences and biomedicine. Institutions strong in engineering or applied sciences, where top papers often appear in specialised journals like IEEE Transactions or The Lancet, are penalised. A 2023 analysis by the National Science Foundation (NSF) showed that U.S. universities accounted for 44% of Nature and Science publications in 2022, reinforcing the Anglo-American advantage in ARWU [NSF 2023, Science and Engineering Indicators].

Comparing Rankings: When to Use Which System

Choosing the right ranking depends on the applicant’s priorities. No single system is universally superior; each serves a different decision-making context.

• For immediate employability: QS is the strongest signal. Its 20% employer reputation weight directly measures how companies perceive graduates. Students targeting corporate jobs in finance, consulting, or technology should prioritise QS rankings.
• For research-oriented graduate programs: THE and ARWU are more relevant. THE’s citation focus reflects current research impact, while ARWU’s Nobel and publication metrics identify institutions with long-term research excellence. A 2024 survey by Nature found that 68% of postdoctoral researchers used ARWU when selecting labs [Nature 2024, Postdoc Survey].
• For teaching quality assessment: None of the three rankings directly measure teaching. QS’s faculty/student ratio is the closest proxy, but it is weak. Applicants should supplement rankings with national teaching quality frameworks, such as the UK’s Teaching Excellence Framework (TEF) or Australia’s QILT surveys.

The Regional Bias in Rankings

All three systems exhibit geographic bias toward English-speaking, research-intensive universities. In QS 2024, U.S. and U.K. institutions occupy 48 of the top 100 spots. THE shows a similar pattern (45 of top 100). ARWU is slightly more diverse, with 39 from the U.S./U.K. and 15 from China and other Asian countries [QS 2024; THE 2024; ARWU 2024]. Applicants from non-English-speaking countries should be aware that their home institutions may be under-ranked.

Methodological Limitations Common to All Three

Despite their differences, QS, THE, and ARWU share several structural limitations that applicants must understand.

First, all three rankings measure inputs and outputs rather than the student experience. Metrics like faculty awards, citation counts, and research funding reflect institutional prestige, not classroom teaching quality or graduate satisfaction. A 2022 meta-analysis in Studies in Higher Education found that global rankings correlate only weakly (r = 0.31) with student satisfaction scores [Studies in Higher Education 2022, Vol. 47].

Second, rankings are static snapshots that do not capture year-to-year changes in program quality. A university may launch a new curriculum or hire a star professor in 2025, but that improvement will not appear in rankings until 2026 or later.

Third, the weighting choices are arbitrary. Why does QS assign 5% to international student ratio while THE assigns 2.5%? There is no empirical justification—these weights reflect the compilers’ judgments. A 2023 report from the European University Association (EUA) criticised rankings for lacking transparency in weight selection [EUA 2023, Rankings in Higher Education].

The Problem of Gaming

Some institutions actively manipulate ranking inputs. For example, universities may inflate faculty/student ratios by hiring part-time adjuncts, or encourage faculty to self-cite to boost citation counts. A 2021 investigation by Science found that at least 12 universities in the top 500 of THE had suspicious citation patterns [Science 2021, Vol. 372]. Applicants should cross-check rankings with independent data sources.

Practical Strategies for Using Rankings

To extract maximum value from rankings, applicants should adopt a multi-system approach rather than relying on a single list.

Step 1: Define your priority. If career placement is paramount, filter by QS employer reputation scores. If you plan to pursue a PhD, rank by ARWU’s research output indicators or THE’s citation scores.

Step 2: Look at subject-specific rankings. All three systems publish discipline-level rankings, which often differ dramatically from overall scores. For example, the University of Chicago ranks 22nd overall in QS but 6th in economics [QS 2024 Subject Rankings]. A student interested in economics should use the subject rank.

Step 3: Cross-reference with national data. In the U.S., the National Center for Education Statistics (NCES) provides graduation rates and median earnings by institution. In the U.K., the Office for Students publishes continuation and employment outcomes. These datasets fill gaps left by global rankings.

Step 4: Ignore small rank differences. A difference of 5–10 positions between two universities is statistically insignificant. The 95% confidence interval for QS rankings is approximately ±8 positions [QS 2024 Methodology]. Focus on rank bands (e.g., top 50, 51–100) rather than exact numbers.

FAQ

Q1: Which ranking is best for finding a job after graduation?

QS World University Rankings are the most useful for employability, as they assign 20% of the total score to employer reputation surveys. In the 2024 QS survey, over 50,000 employers from 140 countries participated, providing direct input on which universities produce the most hireable graduates [QS 2024]. However, applicants should also consult the QS Graduate Employability Rankings, which include metrics like alumni outcomes (25%) and partnerships with employers (25%). A university ranked 50th in QS overall may rank 30th in employability, making the specialised list more relevant.

Q2: Why do rankings change so much from year to year?

Year-to-year rank volatility is largely driven by methodological changes rather than actual shifts in institutional quality. For example, in 2024, QS added a new sustainability indicator (5%) and adjusted the weight of faculty/student ratio from 20% to 15%, causing hundreds of universities to shift by 10–30 positions [QS 2024]. Additionally, the inclusion or exclusion of new institutions in the dataset can push existing universities down. A university that drops from 50th to 55th likely did not become worse—the ranking simply recalculated its position relative to a larger pool.

Q3: Are rankings useful for non-English-speaking universities?

Rankings are less reliable for non-English-speaking institutions due to language and data biases. ARWU is the most objective for these universities because it uses bibliometric data that includes non-English journals, but it still undercounts publications in languages like Mandarin, German, or French. For example, Tsinghua University ranks 22nd in ARWU 2024 but only 25th in QS, partly because QS’s reputation surveys receive fewer responses from Chinese academics relative to the university’s actual output [ARWU 2024; QS 2024]. Applicants from non-English-speaking countries should supplement global rankings with national accreditation systems.

References

• QS 2024, QS World University Rankings Methodology, Quacquarelli Symonds.
• THE 2024, World University Rankings Methodology, Times Higher Education.
• ARWU 2024, Academic Ranking of World Universities Methodology, ShanghaiRanking Consultancy.
• OECD 2023, Education at a Glance 2023: OECD Indicators, Organisation for Economic Co-operation and Development.
• NSF 2023, Science and Engineering Indicators 2023, National Science Foundation.