📚 QS World University Rankings: The Five Dimensions and Their Weightings | QS 世界大学排名的五大维度与权重构成
Every autumn, millions of students, parents and counsellors open the same webpage: the QS World University Rankings. It is quoted in university brochures, in visa seminars and in family group chats. Yet very few people who use the table understand how the number was actually built. This article breaks the QS methodology down into its five scoring dimensions, shows the exact weighting attached to each, and explains what those weights mean for an A-Level student choosing where to apply.
每年秋天,数百万学生、家长和升学顾问都会打开同一个网页:QS 世界大学排名。它被写进大学宣传册、签证讲座和家庭群里。但真正使用这张榜单的人当中,极少有人了解那个数字到底是怎么算出来的。本文把 QS 的评分方法拆解为五大维度,给出每个维度的精确权重,并解释这些权重对正在选校的 A-Level 学生意味着什么。
1. What the QS Ranking Actually Measures | QS 排名究竟衡量什么
The QS World University Rankings are compiled by Quacquarelli Symonds, a UK-based higher education analytics company. Alongside the Times Higher Education (THE) ranking and the Shanghai Academic Ranking of World Universities (ARWU), QS forms the trio of global league tables most often cited by applicants. Its distinctive feature is that it deliberately mixes subjective reputational survey data with objective bibliometric data, so a university’s position reflects both what academics think of it and what its researchers actually publish.
QS 世界大学排名由英国高等教育数据分析公司 Quacquarelli Symonds 编制。它与《泰晤士高等教育》(THE)排名、上海软科世界大学学术排名(ARWU)并称三大全球大学排行榜,也是申请者引用最多的一类。QS 最鲜明的特点是:它有意把”主观的声誉调查数据”和”客观的文献计量数据”混合在一起,因此一所大学的位次既反映学术界对它的评价,也反映它的科研人员实际发表了多少成果。
The ranking is annual and global, currently covering over 1,500 institutions across roughly 100 countries and territories. Crucially, QS does not measure teaching quality directly. It uses proxies: reputation surveys stand in for prestige, staff-student ratios stand in for contact hours, and citation counts stand in for research strength.
该排名每年发布一次,覆盖全球约 100 个国家和地区的 1500 多所院校。关键的一点是:QS 并不直接测量教学质量,而是使用代理指标——声誉调查代表”声望”,师生比代表”师生接触时长”,引用量代表”科研实力”。理解”代理指标”这个概念,是读懂整个排名的前提。
2. The Five Dimensions and Their Weights | 五大维度与权重构成
Since the 2024 edition, QS has used nine individual indicators, which can be grouped into five conceptual dimensions. The grouping below is the cleanest way to read the methodology, because it shows where the real scoring power lies.
自 2024 年版起,QS 采用九个具体指标,可以归纳为五大概念维度。下面这种归组方式最能看清方法论的实质,因为它直观展示了”分数权柄”究竟握在谁手里。
| Dimension 维度 | Indicator 指标 | Weight 权重 | Dimension total 维度合计 |
|---|---|---|---|
| Research & Reputation 科研与声誉 | Academic Reputation 学术声誉 | 30% | 50% |
| Research & Reputation 科研与声誉 | Citations per Faculty 师均引用 | 20% | 50% |
| Employability 就业力 | Employer Reputation 雇主声誉 | 15% | 20% |
| Employability 就业力 | Employment Outcomes 就业成果 | 5% | 20% |
| Learning Experience 学习体验 | Faculty/Student Ratio 师生比 | 10% | 10% |
| Global Engagement 国际化 | International Faculty Ratio 国际教师比例 | 5% | 15% |
| Global Engagement 国际化 | International Students Ratio 国际学生比例 | 5% | 15% |
| Global Engagement 国际化 | International Research Network 国际研究网络 | 5% | 15% |
| Sustainability 可持续性 | Sustainability 可持续发展 | 5% | 5% |
The single most important fact to memorise is that reputation surveys together account for 45% of the total score (Academic 30% + Employer 15%). Almost half of a university’s QS position is decided by what people say about it rather than by anything the university objectively produces.
最值得记住的一个数字是:两项声誉调查合计占总分的 45%(学术声誉 30% + 雇主声誉 15%)。也就是说,一所大学的 QS 位次有近一半取决于”别人怎么评价它”,而不是它客观上产出了什么。
3. Dimension One: Academic Reputation (30%) | 维度一:学术声誉(30%)
Academic Reputation is the heaviest single indicator in the entire framework. QS distributes a survey to academics worldwide, asking them to name the institutions they consider strongest in their own field of expertise. Respondents cannot vote for their own institution, and the survey is regionally and subject-wise weighted to prevent one country or one discipline from dominating the sample.
学术声誉是整个体系中权重最高的单项指标。QS 向全球学者发放调查问卷,请他们列出自己专业领域内最顶尖的院校。受访者不能给自己所在的学校投票,而且样本会按地区和学科进行加权,以防止某一个国家或某一门学科垄断票源。
In recent cycles QS has reported response volumes in the region of 150,000 to 200,000 academics. This sounds large, but it is spread thinly across thousands of institutions and dozens of disciplines, which is why the indicator is heavily skewed toward large, old, English-language, research-intensive universities.
近几个周期中,QS 公布的学术受访者规模约为 15 万至 20 万人。这个数字看似庞大,但分摊到数千所院校和数十个学科后其实相当稀薄——这正是该指标明显偏向”规模大、建校久、以英语为教学语言、研究密集型”大学的原因。
- Strength 优点: Captures intangible prestige that no database can measure directly. 能捕捉任何数据库都无法直接量化的无形声望。
- Weakness 缺点: Reputation is sticky and lags reality by years, sometimes decades. 声誉具有粘性,往往滞后于现实数年甚至数十年。
- Practical effect 实际影响: A tiny, excellent specialist college can never outscore a historic global brand here. 一所规模极小但极其优秀的专科学院,在这个指标上永远不可能超过历史悠久的世界级品牌校。
4. Dimension Two: Citations per Faculty (20%) | 维度二:师均引用(20%)
Citations per Faculty is the main objective indicator. QS takes the total number of citations a university’s researchers received over a five-year window, divides it by the number of academic staff, and then normalises the result by the university’s subject mix, so that a medical-heavy institution is not unfairly rewarded simply because biomedical papers attract more citations.
师均引用是主要的客观指标。QS 统计一所大学的研究人员在五年窗口期内获得的总被引次数,除以该校的学术人员总数,再按学校的学科结构进行归一化处理,以免一所医学院占比极高的院校仅仅因为生物医学论文天生引用多而被不当抬高。
The underlying publication and citation data come from Scopus, the Elsevier bibliographic database. Because the divisor is faculty headcount, this indicator structurally favours small, research-intensive institutions: a university with 900 staff and world-class output can beat a university with 9,000 staff and merely good output. This is the single most common reason a small European or Asian science-focused university outperforms a much larger English-speaking one.
底层论文与被引数据来自爱思唯尔旗下的文献数据库 Scopus。由于分母是教师人数,该指标在结构上天然有利于”小而精”的研究型院校:一所只有 900 名教师但产出世界级的大学,可以击败一所拥有 9000 名教师、产出仅属良好的大学。这正是许多欧洲或亚洲的小型理工强校能够压过规模大得多的英语区高校的最常见原因。
Citations per Faculty = Total Citations (5 years) ÷ Number of Faculty, then subject-normalised
5. Dimension Three: Employer Reputation (15%) | 维度三:雇主声誉(15%)
The Employer Reputation indicator comes from the QS Employer Survey, in which recruiters and human resources professionals are asked which institutions produce the most competent, innovative and job-ready graduates. It carries 15% of the overall score, making employability the second-largest conceptual dimension after research.
雇主声誉来自 QS 雇主调查,问卷对象是企业招聘负责人和人力资源专业人士,请他们指出哪些院校培养的毕业生最有能力、最具创新性、最”即插即用”。该指标占总分 15%,使”就业力”成为仅次于”科研”的第二大维度。
This indicator matters enormously for students on professionally oriented pathways such as Law, Business, Accounting, Engineering and Computer Science. Oxbridge, Imperial, LSE and a handful of global brands score near the ceiling here, which is a large part of why they hold their positions despite middling faculty-student ratios.
对法律、商科、会计、工程和计算机等职业导向路线的学生而言,这个指标极为重要。牛剑、帝国理工、伦敦政经以及少数全球品牌校在此指标上几乎顶格,这也是它们即便师生比只属中等仍能守住高位的重要原因。
6. Dimension Four: Faculty/Student Ratio (10%) | 维度四:师生比(10%)
Faculty/Student Ratio is the only indicator in the framework that functions as a proxy for the student learning experience. It is calculated by dividing the number of full-time equivalent academic staff by the number of enrolled students; a lower ratio produces a higher normalised score.
师生比是整个体系中唯一用于代理”学生学习体验”的指标。它的算法是用全职等效学术人员数除以在册学生数,比值越低,归一化后的得分越高。
Its weight was cut from 20% to 10% in the 2024 reform. This is a much bigger deal than it sounds. Universities that had built their global reputation largely on small seminar classes and generous staff contact, particularly many UK and European institutions outside the very top, saw their scores compressed relative to research-heavy rivals.
该指标权重在 2024 年改版中由 20% 被下调至 10%。这件事的影响远比听起来大。那些主要靠”小班研讨 + 充裕师生接触”建立国际声誉的大学——尤其是牛剑之外的大量英国及欧洲院校——其得分相对于研究密集型对手被明显压缩。
- What it captures 它捕捉什么: Roughly, how much staff time is available per student. 大致反映每名学生可获得的教师时间。
- What it misses 它遗漏什么: Class size alone, contact-hour quality, supervision standards, feedback quality. 班级规模之外的一切:课时质量、导师水平、反馈质量。
- Typical winners 典型受益者: Small liberal arts colleges and specialist institutes. 小型文理学院与专科学院。
7. Dimension Five: Global Engagement (15%) | 维度五:国际化(15%)
The Global Engagement dimension is a composite of three separate 5% indicators. Together they carry 15%, which is more than the entire learning experience dimension. Understanding this helps explain why some universities with modest research output still rank surprisingly well.
国际化维度由三个各占 5% 的独立指标组成,合计 15%,比整个”学习体验”维度的权重还高。理解这一点,就能解释为什么一些科研产出平平的大学排名却出人意料地靠前。
- International Faculty Ratio (5%) 国际教师比例: The share of academic staff who are international. 学术人员中来自海外的比例。
- International Students Ratio (5%) 国际学生比例: The share of enrolled students who are international. 在册学生中来自海外的比例。
- International Research Network (5%) 国际研究网络: The breadth of a university’s international co-authorship, measured from Scopus data and subject-normalised. 依据 Scopus 数据衡量的国际合著广度,并按学科归一化。
Note the practical consequence: a university can gain 10% of its total score purely by recruiting international staff and students. This creates a real incentive for aggressive international marketing, which in turn means that a heavily internationalised institution in a small country can outperform a domestically focused one of comparable academic quality.
请注意其现实后果:一所大学仅靠招收国际教师和国际学生,就能拿到总分中的 10%。这构成了强烈的国际招生动机,也意味着一所位于小国、高度国际化的院校,可能压过学术水平相当但以本土学生为主的院校。
8. How the Scores Are Actually Calculated | 分数到底怎么算出来的
Raw indicator values are not comparable across indicators. The number of citations per faculty might run into the thousands, while the proportion of international students is a fraction between 0 and 1. QS therefore standardises each indicator before weighting it using a z-score transformation, then rescales the results onto a 0-100 scale.
各指标的原始数值彼此不可比:师均引用可能是几千,而国际学生比例是 0 到 1 之间的小数。因此 QS 在加权前先用 z 分数变换对每个指标做标准化,再把结果重新映射到 0-100 的标尺上。
z = (x − μ) ÷ σ
Here x is the institution’s raw value, μ is the mean of all ranked institutions on that indicator, and σ is the standard deviation. After transformation, the top-performing institution on each indicator is typically set to 100 and the scale adjusts accordingly. The overall score is then simply the weighted sum:
其中 x 是该校在该指标上的原始值,μ 是所有上榜院校在该指标上的均值,σ 是标准差。变换后,每个指标上表现最好的院校通常被设为 100,其余按比例缩放。总分就是加权求和:
Overall = 0.30 × AR + 0.20 × CPF + 0.15 × ER + 0.10 × FSR + 0.05 × IFR + 0.05 × ISR + 0.05 × IRN + 0.05 × EO + 0.05 × SUS
Because z-scores are sensitive to the distribution of the whole pool, adding or removing institutions, or changing the composition of the ranked set, can shift everyone’s normalised scores even when nothing about the universities themselves has changed.
由于 z 分数对整个样本的分布非常敏感,增加或移除一些院校、或改变上榜院校的构成,都会让所有学校的归一化得分发生移动——即使这些大学本身毫无变化。
9. Before and After the 2024 Reform | 2024 年改版前后对比
The 2024 methodology change was the largest in the ranking’s history. Three indicators were added (International Research Network, Employment Outcomes and Sustainability), the academic reputation weight was cut from 40% to 30%, and the faculty/student ratio weight was halved. The table below shows the shift side by side.
2024 年的方法论调整是该排名历史上幅度最大的一次:新增三个指标(国际研究网络、就业成果、可持续发展),学术声誉权重从 40% 下调至 30%,师生比权重被腰斩。下表并排展示这一变化。
| Indicator 指标 | Pre-2024 改版前 | Current 现行 | Change 变化 |
|---|---|---|---|
| Academic Reputation 学术声誉 | 40% | 30% | −10 |
| Employer Reputation 雇主声誉 | 10% | 15% | +5 |
| Faculty/Student Ratio 师生比 | 20% | 10% | −10 |
| Citations per Faculty 师均引用 | 20% | 20% | 0 |
| International Faculty Ratio 国际教师比例 | 5% | 5% | 0 |
| International Students Ratio 国际学生比例 | 5% | 5% | 0 |
| International Research Network 国际研究网络 | — | 5% | New 新增 |
| Employment Outcomes 就业成果 | — | 5% | New 新增 |
| Sustainability 可持续发展 | — | 5% | New 新增 |
The direction of travel is clear: reputation is being diluted, and employability, sustainability and international collaboration are being rewarded. Comparisons between a 2022 ranking and a 2025 ranking are therefore not like-for-like, even though media coverage rarely says so.
改革的方向很清晰:声誉的权重在被稀释,而就业力、可持续发展和国际合作正在被奖励。因此,把 2022 年的榜单和 2025 年的榜单放在一起直接比较,并不是同口径对比——尽管媒体报道很少提及这一点。
10. Known Criticisms and Structural Biases | 已知的争议与结构性偏差
QS publishes its methodology openly, which is more than can be said for every ranking, but the framework has well-documented biases that any applicant should weigh before treating a position as a verdict.
QS 公开披露自己的方法论,这一点比不少同类排名做得好;但该体系存在若干有据可查的偏差,申请者在把名次当作定论之前应当加以权衡。
- English-language bias 英语偏向: Reputation surveys circulate mainly in English, and citation databases index English-language journals most completely. 声誉调查主要以英语展开,文献数据库对英文期刊的收录也最完整。
- Age and scale bias
Published by TutorHao | Exam Prep Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导