📚 QS Rankings Data Sources: Academic Surveys, Bibliometric Data and Institution Submissions | QS 排名的数据来源构成：学术调查、文献数据与院校提交
The QS World University Rankings are published every year and read by millions of students, parents and academics. Yet relatively few readers ask the question that matters most: where do the numbers actually come from? A QS rank is not a single measurement. It is a weighted composite assembled from three very different data streams – a global reputation survey, bibliometric data extracted from a commercial citation database, and statistics that universities report about themselves. Each stream has its own method, its own error profile and its own incentives.
QS 世界大学排名每年发布，被数以百万计的学生、家长与学者阅读。然而很少有人追问那个最关键的问题：这些数字究竟从何而来？QS 排名并不是一次单一的测量，而是由三条截然不同的数据流加权合成的结果——全球声誉调查、从商业引文数据库提取的文献计量数据，以及大学自行申报的统计数据。每条数据流都有自己的方法、误差特征与激励结构。
Understanding the provenance of the data is the fastest way to judge what a rank can and cannot tell you. Two institutions sitting five places apart may differ only within the noise of a survey; two institutions that look similar on the surface may derive their position from completely different indicators.
理解数据的来源构成，是判断一个排名能说明什么、不能说明什么的最快途径。名次相差五位的两所大学，其差异可能完全落在调查误差范围之内；而表面上条件相近的两所学校，其名次也可能来自完全不同的指标来源。
1. Why Data Provenance Matters | 为什么数据来源至关重要
A ranking is a model, not a measurement. It takes several narrow observations about an institution and compresses them into a single number. Every step of that compression – what is measured, who reports it, how it is normalised, how much weight it carries – is a design decision made by the publisher, not a law of nature.
排名是一种模型，而不是一次测量。它把关于一所大学的若干片面观察压缩成一个数字。压缩过程中的每一步——测量什么、由谁报告、如何归一化、占多少权重——都是发布机构的设计决定，而非自然规律。
The three data streams in the QS methodology are fundamentally different in kind. Reputation data is subjective but global in scope: it captures what the academic community believes. Bibliometric data is objective in principle but narrow in coverage: it captures what is published in one indexed literature. Institution-submitted data is precise on paper but depends entirely on how each university defines and counts its own staff and students.
QS 方法中的三条数据流在性质上根本不同。声誉数据是主观的，但覆盖面是全球性的：它捕捉的是学术共同体的看法。文献计量数据原则上客观，但覆盖面狭窄：它只反映某一个被索引的文献体系中的产出。院校申报数据在纸面上很精确，却完全取决于每所大学如何定义和统计自己的教职工与学生。
A serious reader therefore asks three questions of any rank: which indicator drove the result, where did that indicator’s data come from, and how reliable is that source for this particular institution? The rest of this article answers those questions indicator by indicator.
因此，严肃的读者面对任何排名都会问三个问题：是哪个指标决定了这个结果？该指标的数据来自哪里？对这样一所特定院校而言，这个来源有多可靠？本文余下部分将逐项回答这些问题。
2. The Three Data Streams at a Glance | 三大数据流一览
Since the 2024 edition, the QS World University Rankings have used a nine-indicator framework. Six indicators draw on institution submissions, two on surveys, and the bibliometric indicators on Elsevier’s Scopus database. The table below sets out the standard published weights and the origin of each indicator.
自 2024 年版起，QS 世界大学排名采用九项指标框架。其中六项依赖院校申报数据，两项依赖调查，文献计量类指标则基于爱思唯尔的 Scopus 数据库。下表列出通行的公开权重与每项指标的数据来源。
| Indicator | 指标 | Weight | 权重 | Data source | 数据来源 |
| Academic Reputation | 学术声誉 | 30% | Global academic survey | 全球学术调查 |
| Employer Reputation | 雇主声誉 | 15% | Global employer survey | 全球雇主调查 |
| Faculty/Student Ratio | 师生比 | 10% | Institution submission | 院校申报 |
| Citations per Faculty | 师均引用 | 20% | Scopus (Elsevier) | Scopus 数据库 |
| International Faculty Ratio | 国际教师比例 | 5% | Institution submission | 院校申报 |
| International Students Ratio | 国际学生比例 | 5% | Institution submission | 院校申报 |
| International Research Network | 国际研究网络 | 5% | Scopus co-authorship | Scopus 合著数据 |
| Employment Outcomes | 就业成果 | 5% | Submission + alumni data | 申报 + 校友数据 |
| Sustainability | 可持续性 | 5% | Scopus-derived SDG output | 基于 Scopus 的 SDG 产出 |
Two warnings are essential. First, the weights above are the widely reported figures for the current nine-indicator framework, but QS reviews its methodology before every edition and has adjusted weights and definitions more than once. Always check the current official methodology page for the edition you are reading. Second, indicator weights are not quality weights: a 5% indicator is not unimportant to the institutions that rank on it, but it also cannot move a score by much on its own.
有两点警示至关重要。第一，上表权重是当前九指标框架下被广泛引用的数字，但 QS 在每一版之前都会审视方法论，并已不止一次调整权重与定义。阅读任何一版排名时，都应核对当版官方方法论页面。第二，指标权重不等于质量权重：占比 5% 的指标对依赖它的院校并非不重要，但它本身也很难大幅改变总分。
3. The Academic Reputation Survey: Design and Scale | 学术声誉调查：设计与规模
The academic reputation survey is the single largest component of a QS score, carrying 30% of the total. QS describes it as the largest survey of its kind, and in recent cycles it has drawn on responses from well over 100,000 academics worldwide. Exact response counts vary by edition and are published in the methodology summary for each year.
学术声誉调查是 QS 得分中占比最大的单一组成部分，占总分的 30%。QS 称其为同类调查中规模最大者，近年各轮调查收集了来自全球十万名以上学者的回复。具体回复数量因版本而异，会在每年的方法论说明中公布。
The sampling frame is the part most readers overlook. Respondents are not drawn from a random sample of all academics in the world. QS builds its invitation list from a combination of previous survey participants, nominations from institutions, and author records derived from bibliometric databases. This is a purposive, snowballing frame rather than a probability sample, which means the survey measures the opinions of a particular, research-active, internationally visible population.
最容易被读者忽略的是抽样框。受访者并非从全球所有学者中随机抽取。QS 的邀请名单由往届受访者、院校提名以及来自文献数据库的作者记录共同构成。这是一个有目的的滚雪球式抽样框，而非概率抽样，因此该调查测量的是某一特定群体——活跃于研究、具有国际可见度的学者——的意见。
Several design features are intended to limit distortion. Respondents declare their field of expertise and their region, so that results can be filtered and weighted accordingly. Self-voting is excluded: academics cannot nominate their own institution. Responses are weighted so that no single country or region dominates the global result, and a subject-level version of the survey feeds the QS World University Rankings by Subject.
若干设计特征旨在限制偏差。受访者需申报自己的专业领域与所在地区，以便对结果进行相应筛选与加权。自我投票被排除：学者不能提名自己所在的院校。回复会被加权处理，以避免任何单一国家或地区主导全球结果；此外，调查还有一个按学科划分的版本，用于 QS 世界大学学科排名。
The residual weaknesses are well documented. Reputation is sticky, so historic prestige persists long after underlying performance changes. Well-known institutions are easier to recall than excellent but obscure ones, producing a halo effect. Response rates differ sharply between regions and between disciplines, and there is no way for an outsider to verify the raw responses because QS does not publish them.
其残余弱点早已被充分记录。声誉具有粘性，历史声望会在实际表现改变之后长期延续。知名院校比优秀但低调的院校更容易被想起，从而产生光环效应。不同地区、不同学科的回复率差异显著；而由于 QS 不公开原始回复，外部观察者无法核验其数据本身。
4. The Employer Reputation Survey | 雇主声誉调查
The employer reputation survey carries 15% of the total score and asks a different question: which institutions produce the best graduates? QS reports responses in the tens of thousands from employers across a wide range of sectors and countries. Respondents indicate the sectors in which they recruit and the institutions from which they prefer to hire.
雇主声誉调查占总分的 15%，问的是一个不同的问题：哪些院校培养出最优秀的毕业生？QS 表示每年收到来自众多行业与国家的数万份雇主回复。受访者需说明自己招聘的行业领域，以及偏好录用哪些院校的毕业生。
This indicator is genuinely useful for applicants because it speaks to graduate employability rather than research output. It is also the indicator most exposed to brand effects. Large multinational recruiters dominate the sample, and their hiring patterns are concentrated in a small number of well-known institutions. Sectors such as consulting, finance and technology are over-represented relative to the overall graduate labour market.
这一指标对申请者确有价值，因为它指向毕业生就业力，而非科研产出。但它也是最容易受品牌效应影响的指标。样本由大型跨国招聘方主导，而其招聘行为集中在少数知名院校。相较整体毕业生就业市场，咨询、金融与科技等行业在样本中占比过高。
For a student aiming at a regional or public-sector career, an employer reputation score built largely from multinational recruitment patterns may be a poor guide. The same caution applies to professional fields where licensing, not institutional brand, drives employment.
对于志在区域就业或公共部门的学生而言，一个主要由跨国招聘模式构成的雇主声誉得分可能并不是好的参考。对于就业由执业资格而非院校品牌决定的专业领域，同样应保持这份警惕。
5. Bibliometric Data: Scopus and Citations per Faculty | 文献计量数据：Scopus 与师均引用
The bibliometric stream is drawn from Scopus, the citation database operated by Elsevier. QS does not collect publication data itself; it licenses and processes records from that database. This matters, because the coverage policy of Scopus becomes a coverage policy of the ranking.
文献计量数据流来自爱思唯尔运营的引文数据库 Scopus。QS 并不自行收集出版物数据，而是从该数据库获取并处理记录。这一点很重要，因为 Scopus 的收录政策在事实上变成了排名的收录政策。
The headline indicator is citations per faculty, weighted at 20%. In simplified form it is calculated as follows.
核心指标是师均引用，权重 20%。其简化计算方式如下。
Citations per Faculty = Total Citations (5-year window) ÷ FTE Academic Staff
Both halves of that fraction are contested. The numerator counts citations to papers attributed to the institution within a rolling five
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