A rank is a compressed answer to a particular question. It is not a course review, and it cannot tell you how a specific course will be taught.
What each ranking is actually measuring
THE describes its World University Rankings as judging research-intensive universities across teaching, research, knowledge transfer and international outlook. Its subject rankings use 18 performance indicators grouped into five pillars, with adjustments for 11 subject areas.
ARWU's indicators are research-performance measures: alumni and staff who have won Nobel Prizes or Fields Medals, Highly Cited Researchers, papers in Nature and Science, papers indexed in major citation indices, and per-capita academic performance.
National datasets answer different questions. The UK Office for Students publishes National Student Survey statistics and student outcome measures for continuation, attainment, completion and progression. The US College Scorecard publishes Department of Education data on costs, student debt, graduation rates, admissions test scores, acceptance rates, student-body diversity and post-college earnings.
None of those tables is a substitute for the others.
Why two tables can put the same university in different places
THE's 2027 methodology says its reputation survey ran from November 2025 to February 2026. The 2026 data were combined with the 2025 survey, giving more than 69,000 responses. THE weighted responses where disciplines or countries were over- or under-represented.
In 2024, THE added a measure of how many institutions have academics voting for a particular university.
THE's key statistics are supplied by universities and usually represent the 2024 academic year.
ARWU's candidate pool is built from institutions with Nobel Laureates, Fields Medalists, Highly Cited Researchers, Nature or Science papers, or significant SCIE and SSCI papers. The 2026 methodology ranks more than 2,500 universities. The 2020 methodology ranked more than 2,000 and published the top 1,000.
ARWU also says an institution's rank reflects the number of institutions above it. A rank is positional, not a fixed score interval.
Data vintage differs too. The OfS outcome files in this source set update continuation and completion to 2022-23 student data, progression to 2021-22 graduates, and student experience to final-year undergraduates in 2022-23 and 2023-24. College Scorecard's field-of-study files pool award years from 2014-15 and 2015-16 through 2018-19 and 2019-20. Its institution-level files cover 1996-97 through 2025-26, and the data were last updated on 2 November 2026.
Those dates matter. A table published in one year can contain data from several earlier years, so a rank change is not automatically a change in teaching.
What THE is useful for, and what it is not
THE's subject rankings require an institution to appear in the overall World University Rankings first. They also require at least 1,000 relevant publications over the five years examined, with at least 100 in each year.
That makes the table relevant to research-intensive universities and to subject-level research profiles. The methodology summary in this source set describes research and reputation indicators, not course contact hours or assessment mix. Those are separate course-level facts.
What ARWU is useful for, and what it is not
ARWU's indicators are prizes, highly cited researchers, Nature and Science papers, citation-indexed papers and per-capita performance. Its candidate rules also favour institutions with those outputs or significant indexed papers.
That is why ARWU is a research-output table. It can help you compare research intensity. The methodology does not include course-level teaching or assessment measures.
How to use UK official data without misreading it
The OfS publishes NSS statistics for 2026, 2025 and 2024, including student characteristics data that compares views across student groups.
Continuation is the proportion of students continuing in a higher education qualification, or who have gained one, one year and 15 days after starting; for part-time students the point is two years and 15 days after starting. Completion is the proportion who gained a qualification, or were continuing, four years and 15 days after starting; for part-time students, six years and 15 days.
Those definitions matter because continuation and completion are not the same measure.
The OfS cautions that archived datafiles reflect its position before investigations opened for the 2022-23 assessment cycle and do not reflect judgments. It also says a provider below a numerical threshold does not necessarily mean positive student outcomes are not being delivered.
Read a below-threshold flag as a reason to look at the measure and context, not as a verdict on the course.
How to use US College Scorecard data
College Scorecard lets users compare up to 10 colleges at once. Department of Education data on costs, student debt, graduation rates, admissions test scores, acceptance rates, student-body diversity and post-college earnings.
Field-of-study data cover salaries and debt of graduates. The field-of-study files pool award years from 2014-15 and 2015-16 through 2018-19 and 2019-20, and report cumulative debt at graduation and earnings one year after graduation.
Use the field-of-study view when you are comparing a major, not just a college brand. Check the award years before treating an earnings figure as current. One-year earnings are an early outcome, not a lifetime projection.
The course-level checklist rankings cannot replace
The ranking indicators described here do not report module structure, contact hours, assessment methods or class sizes by course. Those are the dimensions to compare directly for a shortlisted course.
- Module structure: which modules are compulsory, which are optional, and how the course builds over the years.
- Contact hours: scheduled lectures, seminars, labs and tutorials, separated by year.
- Assessment mix: exams, coursework, practical work and group projects, with weightings where published.
- Class sizes in the department that teaches the course, not the whole university.
- Tutoring and support: academic advising, feedback turnaround and access to staff.
- Professional recognition: the accrediting body named for the course and the jurisdiction where you intend to practise.
Accreditation is not one of the indicators described in THE, ARWU or College Scorecard. Identify the exact body and the exact course; a university-wide rank does not settle it.
How to read student reviews next to rankings
The OfS NSS data include comparisons across student characteristics. That makes the survey more structured than a handful of open reviews.
Still, a few reviews are anecdotes, not a survey. Use them to generate specific questions about timetabling, lab access, feedback or class size. Do not use them to estimate a satisfaction rate.
Pair student voices with outcome definitions. The OfS says continuation and completion are observed at different points after start. A course with strong comments but weak continuation data needs a closer look at the measure and its date.
A decision sequence that keeps rankings in their place
- Shortlist by course fit. Compare modules, contact hours, assessment mix and class sizes for the specific course.
- Match each ranking to the question it can answer. Use THE and ARWU for research-profile questions. Use OfS for UK continuation, completion and NSS data. Use College Scorecard for US costs, debt, graduation and earnings data.
- Note the data year for every table you use. THE's key statistics usually refer to the 2024 academic year. OfS outcomes run to 2022-23 for continuation and completion, and 2021-22 for progression. College Scorecard field data pool award years through 2019-20.
- Read student voices as leads, not vote counts.
- Compare cost and outcomes last, after course fit is clear.
Questions applicants ask
Do ranking gaps mean one course is taught better?
Not by themselves. A gap can reflect what is measured, how it is weighted, or when the data were collected. It is not a direct measure of teaching on a specific course.
What does THE measure?
THE says it judges research-intensive universities across teaching, research, knowledge transfer and international outlook. Its subject rankings use 18 indicators in five pillars and adjust for 11 subject areas.
What does ARWU measure?
ARWU uses Nobel and Fields awards, Highly Cited Researchers, Nature and Science papers, citation-indexed papers and per-capita academic performance. It ranks more than 2,500 universities under its 2026 methodology.
Which UK data should applicants read?
Use the OfS NSS statistics for 2026, 2025 and 2024 for student views, and the student outcomes data for continuation, attainment, completion and progression. Check the definitions: continuation is measured one year and 15 days after start, and completion four years and 15 days after start for full-time students.
Which US data should applicants read?
College Scorecard publishes costs, debt, graduation rates, acceptance rates, earnings and field-of-study debt and salary data. Its field-of-study files pool award years through 2019-20 and report earnings one year after graduation.
What should applicants compare for a specific course?
The indicators described here do not include module structure, contact hours, assessment methods or class sizes. Compare those course-level facts directly, along with tutoring, support and professional recognition.
Source notes
- ShanghaiRanking, 'Academic Ranking of World Universities Methodology 2026' and 'Methodology 2020', undated pages.
- Office for Students, 'Published statistics' and 'Student outcomes data', undated pages.
- U.S. Department of Education, College Scorecard, 'Home' and 'Data Home'; Data Home last updated 2 November 2026.