大学排名权重 QS与THE指标 就业数据
每年放榜季，家长群里最常被转发的一张图就是世界大学排名。把学校排成一条从一到几百的队列，看起来一目了然，数字越小就越像正确答案。这个直觉很省事，代价是把一套有特定用途的评分表，当成了给个人做选择的标尺。这篇把排名里每一分从哪里来拆开看，再对照英国官方公布的学生数据，最后给出一套可以照着走的筛选顺序。
Every results season the same league table image travels through family group chats, and a single column of numbers looks decisive. That shortcut has a cost, because it treats a scoring sheet built for one purpose as a personal decision tool. The sections below take the score apart, then put the official UK student data next to it.
排名衡量的到底是什么：QS 五组权重与 THE 五大支柱
QS 官网在方法说明页把评分分成五组权重：研究与发现占 50%，就业与成果占 20%，全球参与占 15%，学习体验占 10%，可持续性占 5%。THE 官网的切法不同，它把 18 项指标归入教学 29.5%、研究环境 29%、研究质量 30%、国际化视野 7.5%、产业 4%。两套体系都把一半以上的分数压在科研产出与学术名声上。换成一句直白的话，排名首先回答的是这所学校在学术圈有多活跃。它不回答本科生一周能见到导师几次，也不回答大一宿舍能不能申请到。看完权重表再读名次，很多家长的第一反应会变，原来孩子最在意的那部分，在总分里可能只占个位数的百分比。
QS sets out five weighted lenses on its methodology page: Research and Discovery 50%, Employability and Outcomes 20%, Global Engagement 15%, Learning Experience 10% and Sustainability 5%. THE groups eighteen indicators into Teaching 29.5%, Research Environment 29%, Research Quality 30%, International Outlook 7.5% and Industry 4%. Both systems therefore place more than half of the score on research output and academic reputation, which is a question about a university's scholarly activity rather than about your weekly contact hours or your chances of getting a hall room.
学术声誉 30% 来自问卷提名，不是课堂质量
QS 官网的指标页写明，学术声誉这一项在总分里占 30%，是单项权重最高的指标。它的收集方式是向全球学者发问卷，请他们按学科与地区提名自己认为水平最高的院校，学校不能自己报名、也不能自己打分。THE 的研究声誉占 18%，来源同样是一年一度的学术声誉调查。把两份问卷加起来，将近一半的总分来自同行的主观判断。很多学生以为排名的分数都是硬数据，其实问卷类指标在 QS 里合计 45%，在 THE 里合计 33%。这类指标反映的是研究圈里的知名度，名气大的院系天然占优。常见的错误是把知名度当成教学投入来读，两者并不是同一件事。
Academic Reputation carries 30% of the QS score, the single largest indicator, and the figures come from a questionnaire sent to scholars rather than from anything a university reports about its own classrooms. Research Reputation is worth 18% in THE and is collected through the same kind of annual survey, so roughly a third to a half of a headline position rests on peer opinion. A large research department with an international name gathers nominations easily, even when its undergraduate seminars are taught by doctoral students who have never been trained to teach.
引用率 20% 对医学理工有利，人文社科被系统性压低
QS 把每位教师的引用量给到 20%，THE 的研究质量合计 30%，其中引用影响一项占 15%。引用是学科差异极大的指标，实验科学一篇论文常有多位作者与几十次引用，人文学科的重要成果可能写成专著，几年内只有个位数引用。两家机构都做了学科归一化处理，QS 官网与 THE 官网都说明会按学科调整之后再比较。归一化只能减缓偏差，并不能消除偏差。一所工科强、人文弱的学校，与一所学科结构相反的学校放进同一张表，引用项的比较仍然不干净。很多学生以为低名次就等于教学质量差，其实人文类院校常在引用项上被系统性压低。
Citation measures reach 20% for citations per faculty in QS and 15% for citation impact inside THE's research quality pillar, and citations behave very differently across disciplines. A laboratory paper may carry a dozen authors and dozens of citations within two years, while a landmark monograph in history might be cited a handful of times in a decade. Both organisations normalise by subject before comparing, yet normalisation softens the distortion rather than removing it, so a specialist institution can look weak on a number that says little about how well it teaches.
就业相关指标在 QS 里加起来只有 20%
QS 的就业与成果这一组占 20%，其中雇主声誉 15%、就业成果 5%。雇主声誉同样是问卷，对象是招聘方，就业成果才用到实际的毕业生去向数据。THE 没有单独的就业支柱，只在教学组里通过机构收入等项间接体现。把两组权重摆在一起看，真正与毕业生就业直接挂钩的分数在 QS 里只有 5%。这解释了一个常见的困惑，一所毕业生起薪普通的学校可以排进前五十，而一所就业表现稳定但科研体量小的学校排在一百名之外。家长在招生展上最想问的往往是毕业去向，而这个问题在排名体系里的权重，比大多数人想象的低得多。
Employability and Outcomes accounts for 20% of the QS score, split into Employer Reputation at 15% and Employment Outcomes at 5%, and only the smaller half draws on where graduates actually went. THE has no separate employability pillar at all, so a family reading a top-fifty table may be looking at a set of numbers in which direct employment evidence carries a weight of one twentieth. That gap explains why a university with unremarkable starting salaries can sit above one with steady employment records but a smaller research base.
国际化 15% 里有一个指标权重为 0%
QS 的全球参与组占 15%，由国际教师比例、国际研究网络、国际学生比例、国际学生多样性四项构成，前三项各 5%。最后一项国际学生多样性目前的权重是 0%，也就是说这条规则被保留着，但不影响今年的名次。THE 的国际学生、国际教师、国际合作各占 2.5%，另有一项海外学习经历，权重同样是 0%。两家机构给 0% 的理由相似，数据质量与国际可比性还在验证，暂时不计入。对留学生来说这件事有点讽刺，因为国际学生比例高的学校会在这一组拿到满分，而学生构成的多样性反而不计分。看到国际化分数高就推断同学来自很多国家，是把两件事混在了一起。
Global Engagement in QS is 15%, shared between international faculty, an international research network, the proportion of international students and international student diversity, with the first three at 5% each. The diversity indicator currently sits at 0%, so it is written into the rules but carries no weight in this edition, and THE likewise lists a study-abroad metric at 0%. Both organisations say the data is not yet comparable enough to score, which leaves an odd position for an applicant, since a campus can score full marks for attracting overseas students while the mix of nationalities behind that number stays unmeasured.
THE 的教学支柱占 29.5% 其中一半是声誉问卷
THE 官网把 18 项指标归入五大支柱，教学支柱占 29.5%，比研究环境与研究质量两项都小。教学支柱内部，教学声誉问卷占 15%，师生比 4.5%，博士学位授予比例 2%，博士教师比例 5.5%，机构收入 2.5%。英国教育部官网与 OfS 官网在看待本科教学质量时用的是另一套东西，教学卓越框架会给出一个总体评级，并分别给出学生体验与学业成果两项评级。两套评价的结论有时并不一致，一所综合名次靠前的学校在教学卓越框架里可能只拿到中等评级。很多学生以为排名已经把教学质量算得很细，其实专门指向本科教学的分数不到总分的一半，而且其中一半还是问卷。
THE distributes eighteen indicators across five pillars, and Teaching at 29.5% is smaller than both Research Environment and Research Quality. Inside that pillar the teaching reputation survey takes 15 points, student-staff ratio 4.5, doctorate-to-bachelor ratio 2, doctorate-to-staff ratio 5.5 and institutional income 2.5. UK quality assurance runs on a separate track, the Teaching Excellence Framework overseen by the Office for Students, which issues an overall rating plus separate ratings for student experience and student outcomes, and the two pictures do not always agree.
排名测不到的部分：学生支持、课程结构与总花费
排名表里找不到三样本科生每天要面对的东西。第一样是学生支持，包括个人导师是否定期见面、心理健康服务要等多久、写作中心能不能预约到位置。第二样是课程结构，同样是经济学学位，有的学校第一年就安排计量与数学训练，有的更偏论文与阅读。第三样是总花费，学费只是一部分，住宿、通勤与日常开支往往占更大比重。英国教育部官网公布的生活费贷款上限可以当作成本参照，2026 至 2027 学年住家是 9,118 英镑，伦敦以外是 10,830 英镑，伦敦是 14,135 英镑。这些数字与名次高低没有关系，却直接决定一年要准备多少钱。
Three things a first-year student meets every week are absent from any league table: how quickly academic support answers, how a particular degree is structured, and what the year will cost in total. Two economics degrees can look identical on a ranking page while one front-loads statistics and the other leans on essays, and the money side is usually decided by rent rather than by tuition. Student finance figures published on gov.uk put the 2026 to 2027 living cost loan ceilings at 9,118 pounds at home, 10,830 pounds outside London and 14,135 pounds in London, which gives a practical anchor for that part of the budget.
官方数据入口：Discover Uni 官网的 NSS 与毕业生去向调查
在商业排名之外，英国还有一套官方的课程级数据。Discover Uni 官网由英格兰的 OfS、北爱尔兰经济部、苏格兰拨款委员会与威尔士的 Medr 共同拥有，数据来自全国学生调查、毕业生去向调查与 LEO 收入记录。官网写明，毕业生去向调查问的是毕业十五个月后的状态，87% 的毕业生在工作或继续学习，7% 处于失业状态，85% 认为当时做的事情有意义，76% 说与自己的未来计划相符。这些数字按课程发布，比综合名次更贴近个人选择。很多学生以为官方数据只有一份总表，其实可以按课程逐个查，还能看到专业认证信息、继续就读比例与往年录取学生的成绩分布。
Alongside the commercial tables there is an official course-level dataset. Discover Uni is owned by the Office for Students, the Northern Ireland Department for the Economy, the Scottish Funding Council and Medr in Wales, and it draws on the National Student Survey, the Graduate Outcomes survey and the LEO tax records. Its published figures describe graduates fifteen months on, with 87% in work or further study and 7% unemployed, while 85% described their activity as meaningful and 76% said it matched their plans.
LEO 收入数据的三个边界：学科口径、15 人门槛、不含自雇
LEO 用税务记录计算毕业生收入，是英国官方最硬的一类就业数据。Discover Uni 官网列出它的三条边界，数据只覆盖英格兰、苏格兰与威尔士的课程，北爱尔兰没有；公开的是毕业三年与五年后的收入，而且只统计由雇主代扣税的受雇收入，自雇者不在其中；同一门课程要累积到至少十五名毕业生，才有数据发布。还有一个容易忽略的口径问题，课程页上的收入数字对应的是学校该学科领域的全体毕业生，并不等于某一门具体课程的毕业生。拿这个数字去比较两门课程，会得出偏得离谱的结论。说实话，看收入数据之前先看清楚它统计的是哪一群人，比看数字本身更重要。
LEO links tax records to student data, which makes it the hardest official evidence available on graduate earnings, and Discover Uni sets out the limits clearly. It covers courses in England, Scotland and Wales but not Northern Ireland, reports earnings three and five years after graduation, counts only taxed employment income so self-employed graduates are missing, and publishes a subject figure only where at least fifteen graduates are in scope. The number on a course page therefore describes everyone in that subject area at the institution, not the specific degree you are considering.
学科之间的差距大于学校之间的差距
英国教育部官网的 LEO 发布页给出了同类比较的基准，在性别相同、GCSE 成绩相近的人群里，毕业生的收入中位数普遍更高，理工科女性毕业生十年后高出约 8,700 到 12,800 英镑，法律经济与管理类女性毕业生高出约 6,200 到 13,200 英镑。这个规律并非人人适用，非理工科与法律类的男性毕业生在部分成绩段反而比没有高等学历的同龄人低约 700 到 2,200 英镑。学科之间的落差同样明显，毕业五年后收入中位数最高的是医学与牙医，最低的是表演艺术。经济学的四分位距达到 32,900 英镑，说明同一学科内部的分化也非常大。选对学科带来的差距，常常大于在两所名次相差三十位的学校之间反复纠结。
Government analysis in the same dataset compares people with similar GCSE attainment and finds that graduates usually earn more than those without a higher level qualification, with female STEM graduates earning between 8,700 and 12,800 pounds more ten years after GCSEs. The pattern is not universal, because male graduates in other subjects fall between 700 and 2,200 pounds below comparable non-graduates in some attainment bands. Subject choice shifts the picture far more than a ranking gap does, and medicine and dentistry sit at the top of the median earnings table five years after graduation with performing arts at the bottom.
代际流动与排名的相关性很低：IFS 与 Sutton Trust 报告
名次高是不是等于改变命运的机会大，这个问题有数据可以回答。Sutton Trust 官网在 2021 年 11 月 24 日发布的 IFS 合作报告用 LEO 数据首次为英格兰各校算出流动率，方法是把低收入学生的录取比例与这批人后来成为高收入者的比例相乘。报告的两个结论值得记住，收入回报最高的学校与最有利于代际流动的学校几乎不重合，约 58% 实现向上流动的毕业生来自 1992 年之后升格的大学。报告也说明，最挑剔的那批学校把校友之间的差距压得最小，但它们录取的低收入学生本来就少。很多学生以为只有挤进前二十才算有价值，这份报告给出的答案要复杂一些，也更值得拿去跟家长讨论。
A 2021 study by the Institute for Fiscal Studies with the Sutton Trust used LEO records to calculate a mobility rate for English universities for the first time, multiplying the share of disadvantaged students admitted by the share of those who later became high earners. Two findings are worth remembering: the institutions with the strongest earnings returns and those best for intergenerational mobility barely overlap, and about 58% of socially mobile graduates came from universities that gained status after 1992. The most selective institutions do narrow the gap between their own graduates, but they admit relatively few disadvantaged students in the first place.
把排名当筛选入口的三步做法
把上面这些拼成一套可以执行的做法。先用学科排名而不是综合排名把范围缩到十五所左右，因为同一所学校不同院系的水平可以差很远。接着到 Discover Uni 官网按课程核对四件事，毕业生去向、收入中位数、学生调查里关于教学与支持的分数、是否有你需要的专业认证，尤其注意收入数字是学科口径。最后把成本与地点放进来，2026 至 2027 学年伦敦的生活费贷款上限比伦敦以外高约三成，国际学生还要另算学费与保险。走完这三步，名次的作用就回到它合适的位置，扮演一个初筛入口，而不是最终答案。常见的错误是拿着一份总排名表按顺序申请，花了四年才发现课程结构并不合适。
A workable order of operations starts with a subject table rather than an overall one, because faculties inside the same university can sit at very different levels. The next step is to check four items on each course page at Discover Uni: destinations, median earnings, student survey scores on teaching and support, and any professional accreditation that matters for your intended career. Cost and location come last, and in 2026 to 2027 the London living cost loan ceiling is roughly a third higher than the figure outside London, before international tuition and insurance are counted.
牛津高阶英汉双解词典 第10版 商务印书馆¥168📱 去京东购买京东精选需要一对一辅导？Need One-on-One Tutoring?
A-Level / GCSE / IB / KS2-KS3 全科辅导
16621398022 同微信
Visit aleveler.com
或截图后用「扫一扫 → 相册」选择本图