How the Rankings Are Calculated — Indicator Weights, Methodology Disputes, and HKU's Strategic Response

One-sentence verdict: QS, THE, and U.S. News measure "how good a university is" with completely different rulers^※: nearly half of QS's weight comes from reputation surveys, THE spreads 29 fine-grained indicators across teaching, research, industry, and internationalisation, and U.S. News is almost pure bibliometrics. The University of Hong Kong (HKU) currently sits at QS world No. 11^※, THE world No. 33^※, and U.S. News world No. 40^※ (all in their latest 2025–2026 editions) — a spread of no fewer than 29 places that is exactly what happens when three different rulers are measuring three different things.

---

# What Exactly Is Each Ranking Measuring? A Guide to the Indicator Systems

Before reading any individual HKU ranking number, you have to understand how that number was produced. The three tables below are the starting point for everything that follows.

# QS's Nine Indicators (from the 2024 edition)

QS expanded its methodology from six indicators to nine in the 2024 edition^※. The five "lenses" and their weights are:

The single most important number in the table: the two reputation indicators together carry 45% of the total weight — nearly half the ranking. Citations per faculty is 20%, and the faculty–student ratio has been cut to 10%. This weight distribution changed markedly in the 2024 edition — more on that below.

# THE's Five Pillars (2026 edition)

THE's 2026 World University Rankings use 18 fine-grained indicators^※, grouped into five pillars. A comparison of core weights:

THE and QS differ most visibly in two places. First, THE's "research reputation" weight (18%) is far below QS's academic reputation (30%); in THE the influence of reputation is diluted across many more dimensions. Second, THE includes a staff–student ratio (4.5%) and an institutional income indicator, yet internationalisation accounts for only 7.5% — well below the 15% QS gives its three internationalisation indicators combined.

# U.S. News's 13 Indicators (Almost Pure Bibliometrics)

U.S. News's Best Global Universities is powered by 13 indicators built on the Clarivate Web of Science database^※; the 2026-27 edition draws its data window from publications in 2020–2024^※. No staff–student ratio, no international student numbers, and "reputation" limited to research reputation alone. The three largest weights:

To compress the three systems into one sentence: QS asks "how do scholars and employers around the world see you?", THE asks "how well do you perform overall across teaching, research, industry, and internationalisation?", and U.S. News asks "how many papers did you publish in the past five years, and how deeply were they cited?" One and the same HKU answers the three questions with three strikingly different answers — and there is no contradiction in that.

---

# The Big QS Methodology Rethink of 2024: How Changes in Method Shook HKU's Ranking

The QS 2024 edition was the biggest methodological overhaul in nearly two decades — and it is the key to understanding HKU's violent ranking swings between 2023 and 2025.

# What Changed in the Weights? Old vs New

The QS 2024 edition expanded six indicators into nine and substantially redistributed the weights^※:

Twenty percentage points of old weight were squeezed (academic reputation −10%, staff–student ratio −10%) to make room for the three new indicators. On the surface this reads as "more dimensions"; in substance it is "half the ruler swapped out".

# Why Did HKU Slip Five Places in a Single QS 2024 Edition?

HKU ranked world No. 21 in the QS 2023 edition (released June 2022)^※ and dropped abruptly to No. 26 in the QS 2024 edition (released June 2023)^※ — a five-place fall in a single year that the media seized on to proclaim "HKU in decline". But the drop had structural causes rather than any loss of real strength: the three new indicators take time to accumulate data, while the down-weighted staff–student ratio happened to be one of HKU's traditionally stable scoring assets.

What makes the point more forcefully is that Hong Kong's other universities slid in lockstep over the same period — HKUST 40→60, CityU 54→70, CUHK 38→47. Four of the city's top universities falling together in one year could hardly be a simultaneous decline in actual quality; it can only be "the ruler changed". The situation forms a sharp contrast with an even more extreme case from the same period: 52 South Korean universities jointly protested the QS 2024 methodology, forming a "University Rankings Forum" and declaring a collective boycott^※ — because the staff–student ratio being cut from 20% to 10% hit Korean universities, historically blessed with low ratios, hard, with some institutions losing two to three hundred places in a single year. QS's response was: "The methodology adjustments were transparently disclosed in advance, and similar approaches had already been used in the Asian rankings for years."

# After the 2024 Edition: Why Did HKU Bounce Back So Quickly?

The new ruler carried one structural advantage for HKU that determined its rebound over the following three years: the International Research Network (IRN) indicator measures the geographic diversity of cross-border partners in a university's publications^※. HKU's international research collaborations are exceptionally broad; more than 74% of professor-track academic staff are non-local, and students from 97 countries and territories are enrolled^※ — precisely the constitution the IRN rewards. Add to that a solid, real accumulation of reputation (see the next section), and HKU rebounded continuously from No. 26 in the 2024 edition to No. 17 in QS 2025^※, No. 11 in QS 2026^※, holding No. 11 again in QS 2027.

---

# The Reputation Surveys: The Largest Weight, and the Most Contested Ground

# How Do the Reputation Surveys Work?

QS's academic reputation survey is sent each year to roughly 70,000 scholars worldwide, asking respondents to nominate "the best universities globally in their field"; the employer reputation survey puts a similar question to corporate recruiters. Together these two carry 45% of total weight — the single largest block in the QS methodology^※. THE's "teaching reputation" (15%) plus "research reputation" (18%), at 33% combined, likewise form the biggest single chunk of its system.

The problem is this: a reputation survey is not measuring "how good this university's teaching was this year" — it is measuring "the impression this university leaves in respondents' minds". And impressions lag behind reality, while being heavily shaped by historical accumulation, linguistic exposure, and brand recognition.

# Four Core Criticisms from the Academy

First: self-referentiality. A peer-reviewed study published in Frontiers in Education in 2025, using Australian university data, found that QS's academic reputation correlates with the overall score at r = 0.98^※, and employer reputation at r = 0.97. Reputation alone can almost predict the total ranking; the independent explanatory power of the other seven indicators is extremely limited. In a sense the ranking becomes a "reputation ranking" — and reputation itself largely comes from the previous edition of the ranking. A loop.

Second: reinforcing established hierarchies. Reputation surveys are naturally biased towards institutions with long historical prominence and strong English-language visibility. Higher-education scholar Ellen Hazelkorn has argued that rankings are in essence "not objective" — the very choice of indicators and weightings reflects the producer's value judgements, not a neutral standard^※. Universities in non-English-speaking countries operate under a systematic disadvantage in academic language and media exposure.

Third: the potential for gaming. Reputation surveys are exposed to being "steered" by institutions. The most notorious case comes from University College Dublin: its president publicly urged professors to nominate the university favourably in QS's reputation survey; after the episode came to light, QS changed its rules to prohibit such conduct^※. QS has since introduced machine-learning tools to detect anomalous voting, but critics argue that the inherently subjective nature of reputation surveys cannot be rooted out.

Fourth: the questionable validity of the staff–student ratio as a proxy. The Frontiers in Education study also found that QS's staff–student ratio correlates with the overall score at just r = 0.11^※ — the lowest of all indicators. In other words, an indicator originally intended to stand in for "teaching quality" barely predicts ranking performance in the actual data. It measures institutional scale and headcount allocation far more than the depth of genuine teaching input.

---

# "Optimising for the Rankings": Strategic Responses and the Goodhart's Law Trap

# What Are Universities Doing?

When rankings shape reputation, reputation shapes student recruitment, and recruitment shapes funding, universities find it very hard not to pour effort into "raising the ranking score". Documented strategic behaviours include:

• Aggressive recruitment of highly cited scholars: HKU attracted more than 120 distinguished international scholars from 13 countries and territories in 2023 alone^※, directly lifting its QS academic reputation, international faculty, and citations-per-faculty scores. In 2024, 53 HKU scholars were named Clarivate Highly Cited Researchers, placing the institution 10th globally by that count^※; in 2025 the number rose to 54, lifting HKU to 13th globally^※.
• Expanding the international research collaboration network: THE's 2025 International Ranking data show that co-authored publications between Hong Kong and mainland Chinese universities grew by about 70% between 2019 and 2023^※, and Hong Kong institutions have benefited visibly on THE's internationalisation indicators. But the scholar Futao Huang has pointed out that mainland scholars' "close research ties with their home institutions" are a major driver of that figure — which raises the question of whether growth in collaboration numbers represents genuine international scholarly exchange or "proximity-driven network inflation" from geographic adjacency.
• Prioritising English-language publishing and international visibility: ranking systems systematically favour English-language papers (Scopus and Web of Science index English-language output at a higher rate), giving institutions an incentive to steer research towards English-language journals — sometimes at the expense of depth in locally relevant policy research and of Chinese-language publication.

# Goodhart's Law: "When a Measure Becomes a Target, It Ceases to Be a Good Measure"

A 2025 study in Scientometrics frames all of this as Goodhart's Law playing out in academia^※: once citation counts and publication volumes are tied directly to rankings, institutions develop such behaviours as "hyper-prolific authors", paper mills, gift authorship, and closed in-house citation rings — behaviours that superficially inflate the indicators while eroding research integrity at the core. Documented violations include mass publishing in journals with very low acceptance thresholds, multi-institution affiliations to score points repeatedly, gift co-authorship, and purchased ghost-writing. This is not simply a "moral problem"; it is an institutional vulnerability built into the ranking system itself.

For HKU specifically, the issues above are background risk rather than documented specific violations. Its reputation accumulation and recruitment of highly cited scholars are, on the public evidence, legitimate academic paths. But the broader context is a reminder: a rising ranking number cannot be equated one-for-one with proportionate gains in teaching quality or student experience. Research suggests that indicator-driven internationalisation can sometimes "divert resources from genuine teaching improvements and community engagement"^※ — a tension that both HKU's administration and its students would do well to keep in view.

---

# HKU's Differing Results in the Three Rankings: Three Rulers, Three Logics of Scoring

# Latest Positions at a Glance

Below are HKU's latest positions in the three global comprehensive rankings (editions released in 2025/2026):

A spread of 29 places across the three rankings maps directly onto the core disagreements between the three systems. The QS No. 11 is driven by HKU's reputation accumulation and deep internationalisation; the THE No. 33 reflects a balanced overall performance across teaching, research, and citations; the U.S. News No. 40 is the plain bibliometric result of five years (2020–2024) of paper output and highly-cited share.

# HKU's Relative Strengths and Weaknesses on Each Indicator

Across the three systems, HKU has a clear set of structural strengths and one persistent weakness.

Structural strengths: a deeply international constitution. Over 74% of professor-track academic staff are non-local, drawn from 97 countries and territories^※ — which scores directly on QS's international faculty (5%), international research network (5%), and international students (5%), and on THE's international outlook (7.5%). Research reputation is equally solid: 54 Highly Cited Researchers in 2025 (Clarivate criteria, November 2025)^※ keep feeding QS's citations-per-faculty and several of U.S. News's bibliometric indicators.

The persistent weakness: scale. As a "small but excellent" institution, HKU is outgunned in volume-dependent measures — U.S. News's total publications (10%) and THE's research productivity (5.5%) — when facing comprehensive universities from the United States, the United Kingdom, and mainland China. This is the core structural reason its U.S. News position (No. 40) lags its QS position (No. 11) by 29 places: U.S. News strips away internationalisation and reputation entirely and compares papers and citations alone, and there, HKU's smallness shows.

---

# Three Critical Lenses for Reading Rankings

# The Precision of the Numbers Is an Illusion

Higher-education scholars have noted that of roughly 18,000 universities worldwide, only about 3% make it into the QS top 500^※; and the further down the list an institution sits, the smaller the differences between adjacent positions become, and the more easily tiny data fluctuations trigger swings of several dozen places. Even within the top 100, neighbouring positions often do not constitute a statistically significant difference. Reading "No. 11" and "No. 15" as two universities of drastically different calibre is to misuse the precision of the digits.

# Year-on-Year Movement Is Often Methodological Noise, Not Change in Strength

The QS 2024 methodology overhaul sent HKU from No. 21 down to No. 26^※, and in the same year 52 Korean universities mounted a joint protest. Any single-year change in position should first be checked for a methodology adjustment before being read as evidence of real change in strength. By the same logic, some of the annual "record high" headlines are merely a favourable comparison against the depressed base a methodology change created the year before.

# Rankings Selectively Define What "a Good University" Means

THE, QS and U.S. News all centre their methodologies on research output; virtually none of the systems includes student satisfaction, undergraduate teaching quality, median graduate salaries, or contributions to social mobility^※. These rankings answer the question "how does this university perform as a research institution?", not "what is it like to study here?". For readers whose primary concern is undergraduate teaching quality, student support systems, or depth in a specific discipline, the global comprehensive rankings are of fairly limited value — consulting the subject-specific rankings, or looking at the institution's own detailed indicators, is usually more informative than comparing totals.

---

Further reading: a deep dive into two rankings over five years QS and U.S. News: A Five-Year Deep Dive · an overview combining four rankings World Comprehensive Rankings · subject-level analysis Subject Rankings