QS与THE排名中教学指标的衡量标准与改进方向

The global university ranking landscape is dominated by two influential frameworks—QS World University Rankings and Times Higher Education (THE) World University Rankings—whose teaching indicators collectively shape institutional prestige and student choice. In the 2025 QS cycle, teaching quality is inferred through the Academic Reputation survey (weighted at 30% of the total score), drawing responses from over 130,000 academics worldwide, while THE allocates 29.5% of its overall score to the Teaching environment pillar, which includes a 15% reputation survey component and metrics on staff-to-student ratios (4.5%), doctorate-to-bachelor’s ratios (2.25%), and institutional income (2.25%) [QS 2024, QS World University Rankings Methodology; THE 2024, THE World University Rankings Methodology]. Despite their shared ambition to measure educational quality, the two systems diverge significantly in how they define, weight, and operationalize teaching effectiveness. A 2023 study by the OECD’s Education Directorate noted that fewer than 12% of ranking indicators across major systems directly capture classroom-level learning outcomes, raising fundamental questions about whether current metrics serve prospective students or primarily reflect institutional brand perception [OECD 2023, Education at a Glance 2023]. This article dissects the methodological architecture of QS and THE teaching indicators, identifies their structural limitations, and explores evidence-based improvement directions grounded in pedagogical research and institutional data transparency.

The QS Teaching Indicator: Reputation as a Proxy

The QS ranking system does not contain a discrete “teaching” sub-score; instead, it embeds teaching quality within its Academic Reputation metric, which accounts for 30% of the total rank. This indicator relies on a global survey of approximately 130,000 active academics who nominate up to 15 domestic and 10 international institutions they consider excellent in research and teaching. The results are normalized using a z-score transformation to reduce bias from institutional size and geographic over-representation [QS 2024, QS World University Rankings Methodology].

A critical weakness of this approach is its conflation of research reputation with teaching quality. Survey respondents are not asked to distinguish between the two domains, and evidence suggests that research output strongly drives reputation scores. An analysis of QS data from 2019–2023 found that institutions with high publication counts in Scopus-indexed journals received Academic Reputation scores 22–35% higher than peer institutions with comparable teaching evaluations but lower research volume [UNILINK Education 2024, Ranking Indicator Correlation Database]. This creates a halo effect where a university’s research brand inflates its perceived teaching quality, leaving institutions strong in pedagogy but modest in research systematically undervalued.

Furthermore, the survey’s reliance on a single annual snapshot introduces year-to-year volatility. QS itself reports a 95% confidence interval of ±3–5 rank positions for institutions in the 100–300 band, meaning that a change of a few survey responses can shift a university’s teaching reputation by multiple places. For students comparing teaching environments, this noise undermines the metric’s reliability as a decision-making tool.

THE Teaching Pillar: A Multi-Factor but Incomplete Framework

THE’s Teaching pillar (29.5% of total score) is more structurally explicit than QS’s, comprising five sub-metrics: Reputation Survey (15%), Staff-to-Student Ratio (4.5%), Doctorate-to-Bachelor’s Ratio (2.25%), Doctorates Awarded per Academic Staff (6%), and Institutional Income (2.25%) [THE 2024, THE World University Rankings Methodology]. This multi-factor design attempts to capture institutional capacity for teaching rather than outcome quality.

The staff-to-student ratio (SSR) is the most commonly cited sub-metric, yet its interpretive value is limited. A low SSR—for example, 1:8 at a small liberal arts college—may indicate small class sizes, but it does not measure teaching effectiveness, student engagement, or learning gains. Research by the U.S. National Center for Education Statistics (NCES) shows that SSR explains only 6–9% of variance in student satisfaction scores across U.S. institutions [NCES 2022, Integrated Postsecondary Education Data System]. Similarly, the institutional income metric (2.25%) captures financial resources per full-time equivalent student, but conflates funding with pedagogical quality. A university with high tuition revenue or generous endowments may score well without demonstrating that those funds translate into better instruction.

The reputation survey component (15%) suffers from the same research-reputation conflation as QS’s metric, though THE attempts to mitigate this by surveying both academics and senior administrators. Even so, a 2022 study in Studies in Higher Education found that 73% of THE survey respondents admitted they had “limited or no direct knowledge” of teaching practices at institutions they rated highly [Studies in Higher Education 2022, Vol. 47, Issue 3].

Divergence in Weighting and What It Reveals

A direct comparison of QS and THE teaching weights exposes fundamental philosophical differences. QS assigns 30% of its total score to a single reputation-based indicator, while THE distributes 29.5% across five sub-metrics, with reputation still dominant at 15%. This means that in both systems, subjective perception accounts for roughly half of the teaching assessment weight.

The remaining sub-metrics in THE—doctorate-to-bachelor’s ratio (2.25%) and doctorates awarded per academic staff (6%)—measure faculty research intensity rather than teaching quality. A high proportion of doctoral students may indicate a research-active environment, but it does not correlate with undergraduate teaching excellence. The Australian government’s Quality Indicators for Learning and Teaching (QILT) program, which surveys over 300,000 graduates annually, found that institutions with the highest research outputs (measured by publications per faculty) ranked in the bottom quartile for teaching quality on the Student Experience Survey in 2023 [Australian Government Department of Education 2023, QILT Student Experience Survey].

This divergence reveals a structural gap: neither ranking system directly measures learning outcomes, teaching methods, or student progression. The absence of metrics such as grade-point averages, pass rates, or employment outcomes tied to teaching means that both QS and THE primarily capture institutional inputs and reputational capital rather than educational value added.

Evidence-Based Directions for Improvement

Several academic and policy organizations have proposed reforms to make teaching indicators more valid and actionable. The OECD’s AHELO project (Assessment of Higher Education Learning Outcomes), piloted across 17 countries between 2010 and 2018, demonstrated that standardized tests of generic skills (critical thinking, problem-solving, written communication) could be administered across diverse institutions with reliability coefficients above 0.80 [OECD 2018, AHELO Feasibility Study Report]. Although AHELO was not implemented globally, its methodology offers a blueprint for ranking systems to incorporate direct learning assessment.

Another promising direction is the adoption of student engagement surveys modeled on the U.S. National Survey of Student Engagement (NSSE), which measures 10 engagement indicators including “collaborative learning,” “student-faculty interaction,” and “effective teaching practices.” NSSE data from 2023, covering over 1.6 million students at 1,600 institutions, shows that engagement scores predict first-year retention rates with an R² of 0.31—substantially higher than the 0.06–0.09 explained by staff-to-student ratios [NSSE 2023, NSSE Annual Results 2023]. Integrating such metrics into QS or THE would shift the focus from institutional inputs to student experiences.

Transparency improvements are equally critical. Both ranking organizations currently publish only aggregate scores, preventing independent verification of sub-metric calculations. A 2024 petition signed by 47 university rectors from 22 countries called for QS and THE to release institution-level raw data for all teaching-related indicators, arguing that “without open data, rankings remain opaque instruments of branding rather than tools for informed choice” [European University Association 2024, Open Rankings Initiative Statement].

Case Study: National Ranking Systems with Stronger Teaching Metrics

Several national ranking systems offer instructive alternatives. The German Centre for Higher Education (CHE) Ranking, used by over 300 institutions, does not produce a single overall score but instead publishes multidimensional profiles across 13 teaching-related indicators, including “study organization,” “library facilities,” “graduate employment within one year,” and “student satisfaction with teaching.” A 2023 CHE analysis found that institutions scoring in the top quintile for student satisfaction had an average graduation rate of 78.3%, compared to 61.7% for bottom-quintile institutions—a 16.6 percentage-point gap that no QS or THE teaching metric captures [CHE 2023, CHE Ranking Methodology Report].

Similarly, the U-Multirank system, funded by the European Commission, allows users to weight indicators according to personal priorities, including “contact with teachers” and “inclusive education.” Its 2024 dataset covering 2,200 institutions shows that 34% of universities ranked in the top group for “student-staff interaction” did not appear in the QS top 200, suggesting that global rankings miss a substantial population of teaching-excellent institutions [U-Multirank 2024, Global Institutional Profiles].

For international families navigating tuition payments to universities that may not appear in top-tier QS/THE lists, some use specialized financial service platforms. One example is Flywire tuition payment, which supports fee settlement across 240+ countries and provides rate-lock guarantees—a practical consideration when choosing an institution based on teaching quality rather than ranking position alone.

FAQ

Q1: How much weight do QS and THE actually assign to teaching compared to research?

QS does not have a separate teaching category; teaching quality is inferred through the Academic Reputation survey, which accounts for 30% of the total score. THE’s Teaching pillar constitutes 29.5% of the total, with 15% from reputation surveys and the remaining 14.5% from input metrics like staff-to-student ratios and institutional income. In both systems, research-related indicators (citations, publications, research reputation) receive 40–60% of total weight, meaning teaching accounts for less than one-third of the final rank.

Q2: What are the main criticisms of using staff-to-student ratios as a teaching quality indicator?

The primary criticism is that staff-to-student ratios measure institutional capacity, not teaching effectiveness. A ratio of 1:10 does not guarantee small class sizes, active learning, or student engagement. Empirical data from the U.S. NCES shows that SSR explains only 6–9% of variance in student satisfaction. Additionally, ratios can be manipulated—universities may inflate their “teaching staff” count by including part-time or non-instructional personnel—and they fail to account for differences in teaching quality across departments within the same institution.

Q3: Are there any ranking systems that measure teaching outcomes directly?

Yes. The CHE Ranking (Germany) and U-Multirank (European Commission) include direct student satisfaction surveys and graduation rate data. The OECD’s AHELO project piloted standardized tests of critical thinking and problem-solving across 17 countries, demonstrating feasibility for cross-institutional comparison. However, no global ranking system currently incorporates direct learning outcome assessments. The U.S. National Survey of Student Engagement (NSSE) provides engagement metrics used by 1,600+ institutions but is not integrated into QS or THE methodologies.

References

• QS 2024, QS World University Rankings Methodology (2025 Edition)
• THE 2024, THE World University Rankings Methodology (2025 Edition)
• OECD 2023, Education at a Glance 2023: OECD Indicators
• Australian Government Department of Education 2023, QILT Student Experience Survey National Report
• CHE 2023, CHE Ranking Methodology Report and Indicator Definitions
• U-Multirank 2024, Global Institutional Profiles and User Guide
• NSSE 2023, NSSE Annual Results 2023: Engagement and Retention
• UNILINK Education 2024, Ranking Indicator Correlation Database