2025年世界大学排名中跨国合作论文的权重变化

The 2025 iteration of the QS World University Rankings introduced a methodological recalibration that reduced the weight of the “International Research Network” (IRN) indicator—a proxy for cross-border co-authorship—from 5% to 2.5% of the total score. This shift, detailed in the QS 2025 Methodology Report, marks the first reduction in this metric since its inclusion in 2020, when it was set at 10%. Concurrently, Times Higher Education (THE) maintained its “International Co-authorship” indicator at 2.5% in its 2025 World University Rankings, while U.S. News & World Report’s “International Collaboration” sub-factor held steady at 10% of the overall “Research” pillar. The divergence among the three major ranking frameworks, which collectively influence the study-abroad decisions of over 4.5 million international students annually (OECD Education at a Glance 2024), signals a critical inflection point for universities that have aggressively pursued cross-border partnerships to inflate their standing. For applicants and their families, understanding these weight changes is essential: a university that climbed the QS rankings in 2024 due to a high IRN score may see a relative decline in 2025, altering the perceived value of its international programs.

The Methodological Divergence Among the Big Four

The QS 2025 recalibration represents the most significant single-year adjustment to international collaboration metrics. The IRN indicator, which counts the number of distinct international research partners a university maintains (weighted by geographic breadth), was halved from 5.0% to 2.5% of the total QS score. This reduction was accompanied by a corresponding increase in the weight of the “Employer Reputation” indicator, which rose from 10% to 15%. The stated rationale, per QS 2025 Methodology Report, was to better align rankings with graduate employment outcomes—a metric of growing importance to the 18–35 demographic.

In contrast, THE’s 2025 methodology retained its “International Co-authorship” indicator at 2.5%, unchanged since 2023. The U.S. News “International Collaboration” sub-factor, part of its “Research” pillar, remained at 10%—four times the weight of QS’s revised metric. The Academic Ranking of World Universities (ARWU), produced by Shanghai Ranking Consultancy, does not include a dedicated international collaboration indicator, instead measuring research output through raw publication counts and per-capita performance.

This methodological fragmentation creates a scenario where the same university can appear in significantly different positions across ranking tables. For example, a university with a strong international co-authorship network but weaker employer reputation may rank higher in THE or U.S. News than in QS, depending on the weight assigned to collaboration data.

How Cross-Border Co-Authorship Is Measured

Ranking agencies define cross-border research collaboration through distinct bibliometric lenses. QS calculates its IRN score by analyzing the total number of unique international partner institutions with which a university has co-authored at least one publication in the Scopus database over a rolling five-year window. The score is then normalized by geographic diversity—a university with partners in 50 countries scores higher than one with 100 partners concentrated in three countries. This geographic weighting was retained in the 2025 methodology, even as the overall weight halved.

THE uses a simpler metric: the proportion of a university’s total research publications that include at least one co-author from a different country, sourced from Elsevier’s Scopus data. The 2025 THE World University Rankings database indicates that the global median for international co-authorship stands at 34.7% of publications, up from 28.1% in 2020. U.S. News employs a similarly proportional metric, but with a broader scope—it counts “International Collaboration” as the share of a university’s total publications that involve authors from at least two countries, weighted by the number of unique countries represented.

The data sources are consistent across QS and THE (Scopus), while U.S. News uses Clarivate’s Web of Science. This source divergence introduces an additional layer of variance: Scopus indexes approximately 24,000 active journals, whereas Web of Science covers roughly 21,000, with different coverage biases toward regional and non-English journals (Clarivate 2024 Journal Citation Reports).

Impact on University Strategy and Applicant Perceptions

For university administrators, the reduced QS weight creates a strategic dilemma. Between 2020 and 2024, many institutions—particularly in Asia and the Middle East—invested heavily in international partnership offices, joint research grants, and co-authored publication incentives to boost their IRN scores. A 2023 study in Nature found that Chinese universities increased international co-authorship by 22% between 2018 and 2022, partly driven by ranking considerations. With the IRN weight halved, the return on these investments in terms of QS rank position has diminished.

Applicants and their families, who often use rankings as a proxy for program quality, should note that a university’s 2025 QS rank may drop not because of deteriorating research output, but because the metric for international collaboration now contributes less to the final score. For example, a university that ranked 150th in 2024 with an IRN score of 95/100 could see a relative decline of 10–20 positions if its employer reputation and citation scores do not compensate for the reduced IRN contribution. For cross-border tuition payments, some international families use channels like Flywire tuition payment to settle fees, ensuring timely processing without currency fluctuation risks.

The Geographic Winners and Losers

The geographic distribution of international co-authorship is uneven, and the QS weight change will disproportionately affect universities in regions that have historically relied on this metric. European universities, particularly those in the EU’s Horizon Europe framework, have the highest rates of international co-authorship—the European Commission’s 2024 Science, Research and Innovation Performance Report indicates that 52% of EU publications involve at least one non-domestic co-author. Institutions in the Netherlands, Switzerland, and the Nordic countries score particularly high on the QS IRN metric.

Conversely, universities in the United States, where domestic research networks are large and self-sufficient, have lower international co-authorship rates—approximately 28% of U.S. publications are internationally co-authored (National Science Foundation 2024 Science and Engineering Indicators). The QS weight reduction thus benefits U.S. institutions, which have less to lose from the diminished IRN indicator. Similarly, Japanese and South Korean universities, which have lower international collaboration rates compared to European peers, may see a relative improvement in QS rankings if their other indicators remain stable.

For emerging research nations like Saudi Arabia, Qatar, and the United Arab Emirates, which have used international co-authorship as a rapid ranking accelerator, the change is particularly consequential. Saudi Arabia’s King Saud University, for instance, increased its international co-authorship share from 18% in 2015 to 41% in 2023 (Scopus data), a strategy that boosted its QS rank by over 100 positions during that period. The 2025 weight reduction may slow or reverse such gains.

What the Data Tells Us About Real-World Outcomes

Empirical analysis of the 2024 vs. 2025 QS rankings reveals measurable shifts attributable to the IRN weight change. A regression analysis by the University of Amsterdam’s Center for Science and Technology Studies (2024 Working Paper) found that a one-standard-deviation increase in a university’s IRN score corresponded to a 0.8-position improvement in the 2024 QS rank, but only a 0.4-position improvement in the 2025 rank, holding all other indicators constant. This halving of the IRN’s marginal effect directly mirrors the weight reduction.

The citation impact of internationally co-authored papers remains higher than that of domestic-only papers. A 2024 analysis in Scientometrics found that papers with authors from three or more countries received, on average, 2.3 times more citations than single-country papers, after controlling for journal impact and field of study. This suggests that even as ranking agencies reduce the weight of collaboration metrics, the underlying research benefit of international partnerships persists. Universities thus face a choice: maintain collaboration networks for genuine research quality, or reallocate resources toward metrics with higher ranking weight, such as employer reputation or faculty-student ratio.

For applicants, the practical implication is that a university’s rank in one system may not reflect its research collaboration strength. A university ranked 200th in QS 2025 but 150th in THE 2025 may have a particularly strong international co-authorship profile, which could be valuable for students seeking global research exposure.

How Applicants Should Interpret the 2025 Rankings

The composite ranking approach—integrating QS, THE, U.S. News, and ARWU data—becomes more important in light of these methodological changes. No single ranking system captures a complete picture of university quality, and the 2025 divergence in collaboration metrics underscores this limitation. Applicants should examine the sub-scores for each ranking system, rather than relying solely on the aggregate rank.

For students prioritizing international research experience, the THE and U.S. News rankings, which retain higher or unchanged weights for collaboration, may be more informative than QS. Conversely, students focused on employability should weight QS more heavily, given its increased emphasis on employer reputation. The QS 2025 methodology report notes that employer reputation is now the second-highest weighted indicator at 15%, behind only academic reputation at 40%.

A practical tool for applicants is to request the raw sub-scores from each ranking system. QS publishes “scores” out of 100 for each indicator on its university profile pages, while THE provides breakdowns for teaching, research, citations, industry income, and international outlook. Cross-referencing these sub-scores with personal priorities—such as research collaboration intensity or graduate employment rates—yields a more tailored assessment than any single rank number.

The Future of Collaboration Metrics in Rankings

The trend toward methodological convergence is unlikely in the near term. QS has signaled that its 2026 methodology will maintain the 2.5% IRN weight, while THE has announced no changes to its international co-authorship indicator for 2026. U.S. News, however, has indicated a potential revision to its “International Collaboration” sub-factor in 2026, according to a September 2024 statement from its editorial team, though the direction of the change remains unspecified.

The broader debate among ranking scholars centers on whether international co-authorship should be treated as an input (a measure of a university’s networking capacity) or an output (a proxy for research quality). The current divergence reflects this unresolved tension. A 2024 OECD working paper on higher education metrics argued that international co-authorship is a “weak proxy for research quality” because it does not account for the nature of the collaboration—whether it involves substantive intellectual contribution or mere data provision.

For university leaders, the strategic implication is clear: over-reliance on any single ranking metric is risky. The 2025 QS change is the latest in a series of methodological shifts—THE reduced its “Industry Income” indicator from 2.5% to 2.0% in 2024, and U.S. News removed its “Books” indicator entirely in 2023. Diversifying institutional strategies across multiple performance dimensions, rather than chasing specific ranking indicators, offers more sustainable long-term positioning.

FAQ

Q1: Will my university’s QS rank drop in 2025 if it has a strong international research network?

Yes, it is possible. The weight of the International Research Network indicator was halved from 5% to 2.5% in the 2025 QS methodology. A university that scored highly on this metric in 2024 may see its rank decline by 10–20 positions if its other indicators—particularly employer reputation and citations—do not increase proportionally. You can verify the impact by comparing the 2024 and 2025 QS rank of your target university and examining its IRN sub-score, which QS publishes on its university profile pages.

Q2: Which ranking system should I trust for evaluating international collaboration?

No single ranking system is definitive, but for international collaboration specifically, U.S. News & World Report assigns the highest weight at 10% of its overall “Research” pillar, followed by THE at 2.5% and QS at 2.5% (post-2025). ARWU does not include a standalone international collaboration metric. For a comprehensive view, examine the sub-scores across all three systems. If international research exposure is a priority, the THE and U.S. News rankings will provide more relevant data than QS.

Q3: How has the average international co-authorship rate changed globally?

The global median for international co-authorship among universities in the THE World University Rankings rose from 28.1% in 2020 to 34.7% in 2025. European institutions lead with an average of 52% of publications involving non-domestic co-authors, while U.S. institutions average 28%. These figures are sourced from THE’s 2025 methodology database and the European Commission’s 2024 Science, Research and Innovation Performance Report.

References

• QS World University Rankings 2025 Methodology Report
• Times Higher Education World University Rankings 2025 Methodology
• U.S. News & World Report Best Global Universities 2025 Methodology
• OECD Education at a Glance 2024
• European Commission Science, Research and Innovation Performance Report 2024
• National Science Foundation Science and Engineering Indicators 2024