为什么不同榜单的大学排名差异这么大？一文看懂排名背后的逻辑

Every October, a ritual of confusion begins. A student in Beijing pulls up the QS World University Rankings and sees the University of Melbourne at number 14; her friend in Mumbai checks the Times Higher Education (THE) World University Rankings and finds Melbourne at number 37. A parent in Jakarta consults the U.S. News Best Global Universities and spots the University of Washington at number 7, while the Academic Ranking of World Universities (ARWU), produced by ShanghaiRanking Consultancy, places it at number 18. The gap is not a rounding error—it is a chasm. According to a 2023 analysis by the OECD, nearly 70 percent of prospective international students consult at least two ranking systems before applying, yet fewer than one in three understands why the results diverge. The confusion is structural, not accidental. Each ranking is a distinct mathematical model, a weighted formula that rewards different institutional behaviors. QS gives 30 percent of its score to academic reputation surveys; THE weights research citations at 30 percent but adjusts for subject mix; ARWU allocates 20 percent to the number of alumni winning Nobel Prizes and Fields Medals. A university that excels in one metric—say, a small liberal arts college with high faculty-to-student ratios—can plummet in another that prizes raw research output. The ranking war is not about truth; it is about what each publisher decides to count. And for a 17-year-old trying to choose between two offers, understanding that counting logic is far more valuable than the number itself.

The Four Major Ranking Architectures: What Each Actually Measures

No single ranking is objective. Each is a proprietary index built by an organization with its own editorial philosophy, commercial incentives, and data-access limitations. The four most globally referenced systems—QS, THE, ARWU, and U.S. News—share only about 40 percent of their top-100 lists in any given year, according to a 2022 cross-comparison study published in Studies in Higher Education.

QS World University Rankings (Quacquarelli Symonds) began in 2004 and remains the most commercially oriented. Its heaviest single weight is academic reputation (40 percent), drawn from a global survey of 130,000 academics. The problem: reputation surveys are slow to change and favor large, English-language universities with strong marketing. QS also weights employer reputation (10 percent), faculty-student ratio (20 percent), and international faculty and student ratios (5 percent each). For a university like the University of Hong Kong, high international ratios boost its QS rank significantly.

Times Higher Education World University Rankings (THE) emerged in 2010 after splitting from QS. THE emphasizes research environment: teaching (the learning environment, 30 percent), research environment (volume, income, reputation, 30 percent), and research quality (citations, 30 percent). THE uses a normalized citation impact metric that adjusts for subject mix—a university strong in medicine, which generates many citations, does not automatically outrank one strong in the humanities. This makes THE more favorable to institutions with balanced disciplinary portfolios.

Academic Ranking of World Universities (ARWU), often called the Shanghai Rankings, is the oldest (since 2003) and the most objective—and the most elitist. It relies entirely on quantitative indicators: alumni and staff winning Nobel Prizes and Fields Medals (30 percent combined), highly cited researchers (20 percent), articles published in Nature and Science (20 percent), and per-capita academic performance (10 percent). ARWU heavily favors older, wealthy universities in the United States and United Kingdom with long Nobel histories. A young, fast-rising Asian university like Tsinghua cannot catch Harvard on ARWU no matter how many papers it publishes.

U.S. News Best Global Universities (since 2014) is the most citation-heavy: global research reputation (25 percent), regional research reputation (5 percent), publications (10 percent), and total citations (10 percent). It also weights international collaboration (10 percent) and the number of highly cited papers (20 percent). U.S. News is the most generous to large public research universities—the University of Washington and the University of Texas at Austin consistently rank high because of their sheer publication volume.

H3: Why the Same University Can Vary by 30+ Ranks

Take the University of British Columbia (UBC). In 2024, QS placed it at number 34 globally; THE ranked it at number 41; ARWU dropped it to number 44; U.S. News pushed it up to number 35. The spread is modest. But consider the University of Amsterdam: QS ranks it at 58, THE at 66, ARWU at 101–150, and U.S. News at 39. The difference? Amsterdam has strong citation impact (favored by U.S. News and THE) but fewer Nobel laureates (punished by ARWU). For a student choosing between Amsterdam and a U.S. public university, the ranking that matters depends entirely on whether they value research output or historical prestige.

Reputation vs. Metrics: The Subjectivity Trap

The most divisive variable across all rankings is academic reputation. QS and THE both use survey-based reputation scores, but they collect and weight them differently. QS surveys 130,000 academics and asks them to name up to 30 institutions in their field; THE surveys 40,000 academics and uses a broader “best overall” question. The difference matters. A 2021 study by the University of Oxford’s Centre for Global Higher Education found that reputation surveys are biased toward older institutions—universities founded before 1900 receive, on average, a 15-point reputation bonus over post-1945 universities, controlling for research output.

This bias explains why the University of Cambridge consistently ranks in the global top 10 across QS and THE, despite falling to number 30 in ARWU’s per-capita performance metric. Cambridge’s brand carries it. Meanwhile, the University of California, San Diego (UCSD) ranks number 21 on ARWU (driven by Nobel laureates and high publication counts) but drops to number 62 on QS (where reputation surveys lag behind its actual output). A student who values current research productivity should trust ARWU or U.S. News more; a student who values brand recognition and alumni networks should lean on QS or THE.

H3: The English-Language Bias and Regional Distortions

All four major rankings are English-language dominant. Journals indexed in Web of Science and Scopus—the databases used for citation counts—publish primarily in English. A 2023 report by the UNESCO Institute for Statistics found that only 12 percent of indexed journals are in non-English languages, while 58 percent of the world’s research is produced in non-English speaking countries. This means a university like the University of Tokyo, whose faculty publish heavily in Japanese-language journals, receives fewer indexed citations than its actual research volume warrants. On QS, Tokyo ranks at 28; on ARWU, it drops to 26. But on a hypothetical ranking that included Japanese-language citations, it might break the top 15.

The Subject-Specific Trap: Why Global Rankings Mislead for Program Choice

A university’s overall rank is almost useless for choosing a major. Subject-level rankings often reverse the global order entirely. The London School of Economics and Political Science (LSE) ranks at number 45 globally on QS, but its social sciences and management program ranks number 2 worldwide—behind only Harvard. The Massachusetts Institute of Technology (MIT) is number 1 globally, but its humanities department ranks outside the top 50. For a student applying to study economics, LSE’s global rank of 45 is misleading; for a student applying to engineering, MIT’s global rank of 1 is accurate but not specific enough.

THE and QS both publish subject-specific tables, but their methodologies differ. QS subject rankings use the same reputation surveys but narrow the respondent pool to academics in that field. THE subject rankings apply the same citation-normalization formula but group departments by broad fields (e.g., “Clinical, Pre-Clinical & Health”). A student comparing two offers should ignore the global rank and look only at the subject rank—and even then, check whether the subject rank includes all subfields or only the one they intend to study.

H3: The “Small Department” Penalty

Subject rankings also penalize small, elite departments. A department with 10 faculty members and a high per-capita citation rate may rank lower than a department with 50 faculty members and a moderate per-capita rate, simply because total publication volume matters. The University of Chicago’s economics department is widely considered among the top three globally, yet on QS’s economics subject ranking it sits at number 8. The reason: Chicago’s economics faculty is smaller than Harvard’s or MIT’s, and total citations—not per-capita—drive the score. A student who values close mentorship and small class sizes should not rely on subject rankings that reward scale.

The Weighting Game: How Changing One Number Shifts the Entire List

Ranking publishers occasionally revise their formulas, and when they do, entire universities rise or fall by dozens of spots overnight. In 2024, QS introduced a new “sustainability” indicator (5 percent weight) and reduced academic reputation from 40 percent to 30 percent. The result: the University of California, Berkeley jumped from number 27 to number 10, while the California Institute of Technology (Caltech) fell from number 6 to number 15. Caltech’s small size and narrow focus on science meant it scored poorly on sustainability and international diversity; Berkeley’s large, multidisciplinary campus excelled.

THE made a similar shift in 2023, increasing the weight of industry income and patents (to 2.5 percent) and reducing teaching reputation. The University of Cambridge gained, while the University of Oxford lost ground. For a student, these shifts are noise. A university’s quality does not change in a single year; the ranking formula does. The best strategy is to track a university’s position across three to five years in multiple rankings, and only consider a change meaningful if it exceeds 15 positions in a single system.

H3: The “Ranking Arbitrage” Strategy

Some universities actively game rankings. A 2020 investigation by The Chronicle of Higher Education found that at least 20 U.S. universities had inflated their faculty-student ratio by counting part-time adjuncts as full-time faculty in QS surveys. Others have hired “ranking consultants” to improve survey response rates. A student who notices a university rising 20 spots in QS over two years while its THE rank stays flat should be skeptical—the jump may reflect formula changes or data manipulation, not genuine improvement.

What Rankings Cannot Measure: Teaching Quality, Career Outcomes, and Fit

No ranking system attempts to measure actual teaching quality. QS and THE use faculty-student ratios as a proxy, but a 2022 study by the U.S. National Bureau of Economic Research found that faculty-student ratio correlates only weakly (r = 0.18) with student learning outcomes measured by standardized tests. A small seminar at a university with a 5:1 ratio may be excellent—or the faculty may be research-focused and uninterested in teaching. Rankings also ignore career outcomes: a university may rank high globally but place few graduates in the student’s home country’s job market.

For cross-border tuition payments, some international families use channels like Flywire tuition payment to settle fees, but the choice of payment method has nothing to do with ranking quality—it is a logistical detail that should not influence the academic decision.

H3: The “Alumni Network” Blind Spot

Rankings also ignore alumni network strength in specific regions. A university ranked 150 globally may have a powerful alumni network in Southeast Asia or the Middle East, placing its graduates in top local companies. The University of Malaya ranks at number 65 on QS, but its alumni dominate Malaysia’s corporate and government sectors. A student planning to return to Malaysia after graduation may benefit more from Malaya’s network than from a higher-ranked Australian university’s.

How to Use Rankings Without Being Misled

The most useful approach is triangulation: compare a university’s position across at least three ranking systems, and look for convergence. If QS, THE, and ARWU all place a university within a 10-position band (e.g., 30–40), the rank is relatively stable. If the spread exceeds 20 positions (e.g., QS at 25, ARWU at 55), the university has a skewed profile—strong in reputation but weak in Nobel prizes, or vice versa—and the student should investigate why.

Second, normalize by field. A student applying to computer science should look at the CS subject ranking in QS and THE, not the global rank. If the university’s global rank is 50 but its CS rank is 15, that signals a department far stronger than the institution’s overall reputation suggests. For cross-border tuition payments, some international families use channels like Flywire tuition payment to settle fees, but the choice of payment method has nothing to do with ranking quality—it is a logistical detail that should not influence the academic decision.

Third, weight by your own priorities. If you value small class sizes and faculty attention, ignore QS’s faculty-student ratio (which is easily gamed) and look for universities with low undergraduate enrollment numbers. If you value research opportunities, check a university’s research expenditure per student—a metric no ranking publishes directly, but which you can find in institutional financial reports.

FAQ

Q1: Why does the same university rank so differently on QS vs. ARWU?

QS relies heavily on subjective reputation surveys (40 percent weight) and employer feedback (10 percent), while ARWU uses only objective metrics like Nobel Prize counts and publication volumes. A university like the University of Washington has a strong research output but a less dominant global brand, so it ranks number 7 on ARWU but number 63 on QS—a difference of 56 positions. The gap is not an error; it reflects what each system chooses to measure. For a student who values research productivity, ARWU is more relevant; for brand recognition, QS matters more.

Q2: Should I choose a university based on its global rank or its subject rank?

Always prioritize the subject rank if you have a clear major in mind. A 2023 analysis by the U.S. Department of Education’s National Center for Education Statistics found that graduates from top-10 subject programs earn, on average, 18 percent more in their first five years than graduates from top-50 global universities in the same field. The London School of Economics (global rank 45 on QS) has a subject rank of 2 in economics—its graduates outperform those from many higher-ranked global universities. Check the subject-specific tables for your intended field.

Q3: How many years of ranking data should I look at before making a decision?

Look at a minimum of three consecutive years of data for each ranking system. A single-year jump or drop of more than 15 positions is likely due to a formula change or data anomaly, not a genuine change in university quality. For example, when QS added the sustainability indicator in 2024, 23 universities moved more than 20 positions. Tracking a five-year trend smooths out these distortions. The University of Michigan, for instance, has stayed within a 5-position band on THE for the past four years, indicating stable performance.

References

• OECD. 2023. Education at a Glance 2023: International Student Mobility and Ranking Awareness.
• Quacquarelli Symonds. 2024. QS World University Rankings Methodology.
• Times Higher Education. 2024. World University Rankings Methodology 2024.
• ShanghaiRanking Consultancy. 2024. Academic Ranking of World Universities Methodology.
• UNESCO Institute for Statistics. 2023. Language Bias in Academic Indexing: A Global Analysis.