📚 The Academic Reputation Survey: How QS Has Measured Prestige Since 2004 | 学术声誉调查：QS 自 2004 年以来使用的声誉评价机制
Every summer, millions of students, parents and university administrators wait for one document: the QS World University Rankings. Behind that single number sits a mechanism that has been running since 2004 – a global opinion poll of academics known as the Academic Reputation Survey. This article explains how that survey was born, how it is designed, how nominations are turned into scores, how its weight has changed, and why it remains the most debated instrument in higher education.
每年夏天，数百万学生、家长和大学管理者都在等待同一份文件：QS 世界大学排名。而在那个单一数字的背后，是一套自 2004 年起持续运转的机制——一项面向全球学者的意见调查，即”学术声誉调查”。本文将说明这项调查如何诞生、如何设计、提名如何转化为分数、其权重如何变化，以及它为何始终是高等教育领域争议最大的测量工具。
1. What the Academic Reputation Survey Actually Is | 学术声誉调查究竟是什么
The Academic Reputation Survey is a large-scale, invitation-based questionnaire in which academics around the world are asked to name the universities they regard as excellent for research in their own field. It is a perception measure, not an output measure: it records what scholars think, not what institutions have published.
学术声誉调查是一项大规模、凭邀请参与的问卷调查，邀请全球学者提名他们心目中在本领域科研表现卓越的大学。它是一种”感知测量”，而非”产出测量”：它记录的是学者的看法，而不是机构的实际发表成果。
Because the survey sits inside a ranking system, its results are quantified. Nomination frequencies are converted into a score on a 0-100 scale, and that score is then weighted and added to bibliometric and structural indicators to produce an institution’s overall position.
由于该调查处在一个排名体系内部，其结果会被量化：提名频次被转换为 0-100 的分数，再经过加权，与文献计量指标和结构性指标相加，最终得出院校的总分与名次。
2. Origins: 2004 and the THE-QS Partnership | 起源：2004 年与 THE-QS 的合作
The first edition appeared in 2004 as the “THE-QS World University Rankings”, produced jointly by Quacquarelli Symonds (QS) and Times Higher Education (THE). It covered roughly 200 institutions and was published in the Times Higher Education Supplement.
第一版于 2004 年问世，名为 “THE-QS 世界大学排名”，由 Quacquarelli Symonds（QS）与《泰晤士高等教育》（THE）联合制作，覆盖约 200 所院校，刊登于《泰晤士高等教育增刊》。
The partnership ended in 2009. From 2010, QS published the rankings under its own name, while THE developed a separate methodology. Crucially, the peer-review survey has been the anchor of the QS method from the very beginning: academic reputation carried a 40% weight in 2004 and remained at 40% for almost two decades.
这一合作于 2009 年结束。自 2010 年起，QS 以自己的名义发布排名，THE 则另行开发了一套方法。关键在于，同行评议调查从一开始就是 QS 方法的核心支柱：2004 年学术声誉权重即为 40%，并在此后近二十年保持不变。
3. Survey Design and Distribution | 调查的设计与发放
Respondents are asked to identify institutions they consider excellent in the field or fields in which they have expertise, with separate treatment of domestic and international nominations and a cap on the number of institutions that may be named. In practice this means a scholar in materials science cannot cast nominations in law, and cannot name an unlimited number of universities.
受访者被要求在其专长领域内提名他们认为卓越的院校，国内提名与国际提名分开处理，且可提名的院校数量设有上限。这意味着一位材料科学的学者不能对法学领域进行提名，也不能无限量地列出大学名单。
The questionnaire is administered online and in English, distributed by email invitation. QS builds its contact list from academic publication databases, publisher and conference records, and from lists of senior staff submitted by institutions themselves. This last channel is one reason critics worry about organised campaigning.
问卷通过线上、以英文形式发放，采用电子邮件邀请的方式。QS 的联系人名单来自学术出版数据库、出版商与会议记录，以及各院校自行提交的高级教职员名单。最后这一渠道正是批评者担忧”有组织拉票”的原因之一。
4. Sampling: Who Counts as an Academic? | 抽样：谁算”学术人”
The sample has grown enormously. The early editions rested on a few thousand responses; recent cycles report well over one hundred thousand academic responses, collected over a rolling multi-year window so that each edition draws on several years of data rather than a single snapshot.
样本规模增长巨大。早期版本仅依赖数千份回复；近年周期报告的有效学术回复已远超十万份，且采用滚动多年窗口采集，因此每一版排名依据的是数年数据，而非单一时点的快照。
QS also applies weighting to correct for geographical over-representation, because responses are not evenly distributed across countries, disciplines or career stages. Without such corrections, the survey would simply mirror the publishing geography of the English-speaking world.
QS 还会施加加权以纠正地域上的过度代表问题，因为回复在国家、学科与职业阶段上的分布并不均匀。若无此类校正，调查结果将只是英语世界科研版图的镜像。
- Frame bias: the sampling frame is publication-based, so teaching-focused staff are under-represented. | 抽样框偏差：抽样框基于发表记录，因此以教学为主的教师代表性不足。
- Non-response bias: busy or less internationally connected academics are less likely to reply. | 无应答偏差：事务繁忙或国际联系较少的学者更不可能回复。
- Language bias: an English-language instrument disadvantages non-anglophone systems. | 语言偏差：英文问卷对非英语体系不利。
5. From Nominations to Scores: The Statistics | 从提名到得分：统计处理
Each institution’s raw reputation figure is essentially its share of nominations within the relevant subject or region. Because raw shares are skewed, QS normalises them before use. A standard approach is the z-score, which expresses how far an institution sits from the mean in units of standard deviation:
每所院校的原始声誉数值，本质上是其在相应学科或地区中所获提名的份额。由于原始份额高度偏态，QS 在使用前会进行标准化处理。常用方法是 z 分数，它以标准差为单位衡量某院校偏离均值多远：
z = (x − μ) ÷ σ
The overall score is then a weighted sum of standardised indicator scores. Under the current framework, the academic reputation indicator enters at 30%:
随后，总分是各标准化指标得分的加权和。在现行框架下，学术声誉指标的权重为 30%：
Overall Score = 0.30 × AR + 0.15 × ER + 0.10 × FSR + 0.20 × CpF + 0.05 × IFR + 0.05 × ISR + 0.05 × IRN + 0.05 × EO + 0.05 × SUS
Where AR = academic reputation, ER = employer reputation, FSR = faculty/student ratio, CpF = citations per faculty, IFR = international faculty ratio, ISR = international student ratio, IRN = international research network, EO = employment outcomes and SUS = sustainability.
其中 AR 为学术声誉，ER 为雇主声誉，FSR 为师生比，CpF 为单位教员引用数，IFR 为国际教员比例，ISR 为国际学生比例，IRN 为国际研究网络，EO 为就业成果，SUS 为可持续性。
6. How Much Does Reputation Count? | 声誉占多大权重
Reputation has always been the single largest block in the QS model, and the balance between academic and employer opinion has shifted over time. The table below summarises the change.
声誉始终是 QS 模型中权重最大的板块，而学术声誉与雇主声誉之间的配比随时间发生了变化。下表汇总了这一变化。
| Indicator | 指标 | 2004-2023 | 2024 onwards | 2024 年起 |
| Academic Reputation 学术声誉 | 40% | 30% |
| Employer Reputation 雇主声誉 | 10% | 15% |
| Faculty/Student Ratio 师生比 | 20% | 10% |
| Citations per Faculty 单位教员引用 | 20% | 20% |
| International Faculty / Students 国际教员与学生 | 5% + 5% | 5% + 5% |
| International Research Network 国际研究网络 | – | 5% |
| Employment Outcomes 就业成果 | – | 5% |
| Sustainability 可持续性 | – | 5% |
Note that the two opinion surveys together still account for 45% of the total, so subjective judgement continues to outweigh every objective indicator in the model.
值得注意的是，两项意见调查合计仍占总分的 45%，也就是说主观判断在模型中依然压过任何一项客观指标。
7. Subject and Regional Variants | 学科与地区变体
QS runs the survey at several levels. The flagship world ranking aggregates reputation across all fields, but the Rankings by Subject group institutions into five broad faculty areas – Arts and Humanities, Engineering and Technology, Life Sciences and Medicine, Natural Sciences, and Social Sciences and Management – and weight the reputation results differently in each.
QS 在多个层面运行这一调查。旗舰性的世界大学排名汇总各领域的声誉结果，而学科排名则将院校归入五大领域——艺术与人文、工程与技术、生命科学与医学、自然科学、社会科学与管理——并在各领域中给予声誉结果不同的权重。
Regional rankings, including QS Asia, the Arab Region and Latin America, use their own survey weightings and regional correction factors. This is why a university can perform very differently in the world table and in a regional table in the same year.
包括 QS 亚洲、阿拉伯地区与拉丁美洲在内的地区排名，使用各自的调查权重与地区校正系数。这就是同一所大学在同一年度的世界排名与地区排名中表现差异明显的原因。
8. The 2024 Reform | 2024 年的方法改革
The 2024 edition introduced the most substantial methodological change in the ranking’s history. Academic reputation was cut from 40% to 30%, the employer survey was raised to 15%, the faculty/student ratio was halved to 10%, and three new indicators joined the model: international research network, employment outcomes and sustainability.
2024 年版引入了该排名历史上最重大的方法变革：学术声誉由 40% 下调至 30%，雇主调查升至 15%，师生比减半至 10%，同时新增三项指标：国际研究网络、就业成果与可持续性。
The reform was a response to sustained criticism that a single perception indicator dominated the outcome, and to competition from other ranking systems that had already diversified their metrics. Its practical effect was a reshuffling of positions, with institutions that score strongly on sustainability and international collaboration gaining ground, and some historically prestigious names slipping.
此次改革回应了长期以来的批评——即单一感知指标主导了结果，同时也回应了其他排名体系早已多元化其指标所带来的竞争。其实际效果是名次重排：在可持续性和国际合作方面表现强劲的院校上升，而一些历史声望极高的院校下滑。
9. Criticisms and Known Biases | 批评与已知偏差
The central criticism is circularity. Reputation rankings reward institutions that are already famous, and fame is partly produced by the rankings themselves. The indicator therefore tends to be self-reinforcing rather than dynamic, and it responds slowly to genuine changes in research quality.
最核心的批评是”循环论证”。声誉排名奖励的是本已著名的院校，而”著名”在某种程度上正是由排名本身制造出来的。因此该指标往往自我强化而非动态变化，对科研质量的真实变化反应迟缓。
Other documented concerns include uneven response rates between countries and disciplines, the advantage held by large institutions with many potential respondents, the role of brand recognition and media presence, and the risk that institutions actively mobilise staff to complete the questionnaire.
其他已被记录的关切包括：不同国家与学科之间回复率不均；拥有大量潜在受访者的大型院校占据优势；品牌知名度与媒体曝光度的作用；以及院校主动动员员工填写问卷所带来的风险。
- Halo effect: a strong overall brand lifts subject-level scores. | 光环效应：强势的整体品牌会拉高学科得分。
- Size effect: bigger faculties generate more nominations. | 规模效应：教员规模越大，产生的提名越多。
- Lag effect: perception trails reality by years. | 滞后效应：感知落后于现实数年。
- Campaign risk: coordinated voting is hard to detect. | 拉票风险：有组织的投票难以识别。
10. QS Compared with THE and ARWU | QS 与 THE、ARWU 的比较
Different systems treat reputation very differently, and understanding this is essential when reading any league table.
不同体系对声誉的处理方式差别很大，理解这一点对阅读任何排名表都至关重要。
| System | 体系 | Reputation Weight | 声誉权重 | Data Source | 数据来源 |
| QS | 30% academic + 15% employer | Own global surveys; Scopus citations |
| THE | Roughly one third of the total | Invited reputation survey; Scopus |
| ARWU (Shanghai) | Zero – no opinion survey | Nobel prizes, Nature/Science papers, highly cited researchers |
A university that ranks highly in QS because of reputation may rank very differently in ARWU, which ignores opinion entirely and measures only hard outputs. Neither number is “wrong”; they answer different questions.
一所因声誉而在 QS 中名列前茅的大学，在完全不看主观意见、只测量硬性产出的 ARWU 中可能名次迥异。两个数字都不”错”，它们回答的是不同的问题。
11. How Students Should Read the Results | 学生应如何解读这些结果
Do not read the headline number alone. Open the indicator breakdown and ask which components produced the position: if reputation dominates and citations are weak, the institution’s standing is driven by perception rather than measurable research output.
不要只看总名次。打开指标明细，问清是哪些分
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