Russell Group Universities: List, Membership Rules, and How It Compares to Your Data
What the Russell Group is, its full official membership list and how membership is described, and how member institutions compare with non-members on the six dimensions unirank.uk publishes, including where the gaps are widest and smallest.
Who the Russell Group is
The Russell Group is an association of UK universities, not an admissions category or standalone ranking system. Its “Who we are” page, published 2026-06-25, says: “The Russell Group represents 24 leading universities across the UK. We are defined by excellence, impact, opportunity and scale.” It adds that the group works “through collaboration and advocacy” to advance the economic, societal and cultural contributions of its universities. These statements are the association’s own description of itself, rather than independent findings about every member.
Official member list
The Russell Group’s official “Our universities” page, checked 2026-10-02, identifies 24 members. In the order published, the complete list is:
University of Birmingham; University of Bristol; University of Cambridge; Cardiff University; Durham University; University of Edinburgh; University of Exeter; University of Glasgow; Imperial College London; King’s College London; University of Leeds; University of Liverpool; London School of Economics; University of Manchester; Newcastle University; University of Nottingham; University of Oxford; Queen Mary University of London; Queen’s University Belfast; University of Sheffield; University of Southampton; University College London; University of Warwick; and University of York.
This is a membership list, not an order of performance. Inclusion does not constitute a recommendation, and the ordering should not be read as a ranking.
Membership rules, and what they are not
The Russell Group’s published pages contain no membership admission criteria, application process for joining, or exit rules. “Membership rules” therefore means only the basis on which the association describes its own composition. “Excellence, impact, opportunity and scale” are self-descriptive principles, not a published eligibility test that an applicant can inspect or use to predict an offer.
On its page published 2026-06-25, the group states that all 24 members are in the QS World Rankings 2027 global top 200, described as the top 2%; 16 are in the top 100, described as the top 1%; and four are in the top 10. The same page attributes 80% of public research funding to Russell Group universities and says 91% were rated “world-leading” or “internationally excellent” in REF 2021. It also describes all members as high-tariff institutions, representing the top third when ranked by average entry tariff achieved, and says around ¾ of a million students, around 25% of all UK students, study at members, citing HESA HE student data. These are the group’s own published claims, not independent verification by unirank.
Official REF 2021 guidance explains the wider research submission process: 157 UK higher education institutions submitted into 34 discipline-based units of assessment, producing 1,878 submissions containing 76,132 full-time-equivalent staff. REF 2021 weighted outputs at 60%, impact at 25% and environment at 15%. That guidance explains how research assessment was produced; it does not supply missing membership criteria.
How the comparison with unirank’s data works
The comparison uses unirank’s indexed dataset, computed 2026-10-02 from 141 institutions and last verified 2026-08-11. It divides the institutions into 24 Russell Group members and 117 non-members. The six dimensions are Teaching Quality, Student Satisfaction, Graduate Employment, Graduate Earnings, Global Standing and Research Quality. Their site scores use a 0–100 percentile or mean scale, so a score such as 87.1 is a relative site score, not a percentage of graduates or a prediction for an individual.
Teaching Quality and Student Satisfaction draw on the site’s National Student Survey inputs; unirank runs no student reputation survey of its own. The comparison record does not identify a common survey edition for these scores, so they should not be presented as results from one survey year.
Graduate Employment and Graduate Earnings use the Department for Education’s Longitudinal Education Outcomes dataset for 2022/23, covering first-degree graduates in Great Britain five years after graduation. Postgraduate earnings are excluded. LEO 2022/23 reports median provider earnings of £39,800 for Russell Group members and £29,600 for non-members. The corresponding means of provider median earnings are £41,622 and £30,583. Mean sustained employment or further study is 89.8% for members and 88.5% for non-members. These pound figures are provider medians or averages of provider medians, not the percentile scores used in the ranking and not individual salary guarantees.
Global Standing is a percentile average across four published global ranking tables, not the QS position itself. The 2026-10-02 unirank dataset reports Russell Group REF GPA values from 3.27 to 3.629, with a mean of 3.42; 138 institutions in the indexed dataset have REF data. REF results, LEO results, global ranking editions and the site’s verification date are different data vintages, not one simultaneous institutional assessment.
Differences across the six dimensions
In the 2026-10-02 unirank computation, valid record counts are shown as member/non-member. Teaching Quality averages 87.9 for members and 86.5 for non-members, a 1.4-point difference, with 24/114 records. Student Satisfaction is 82.2 against 81.9, a 0.3-point difference, also with 24/114 records. These are small differences.
Graduate Employment is 89.8 against 88.5, a 1.3-point difference, with 23/111 records. Graduate Earnings is 87.1 against 42.3, a 44.8-point difference, with 23/111 records. Global Standing is 91.6 against 56.4, a 35.2-point difference, with 24/69 records. Research Quality is 85.5 against 72.0, a 13.5-point difference, with 24/114 records.
Overall, the 2026-10-02 site scores average 87.3 for members and 72.3 for non-members, a 15.0-point difference across 24/117 records. Member and non-member overall medians are 86.7 and 71.0 respectively; their ranges are 82.3–91.7 and 61.3–89.5. Mean rank positions are 18.3 and 81.8 across the dataset’s positions 1–141.
The pattern is therefore mixed: teaching, satisfaction and employment show small average differences, while earnings, global standing and research show much larger ones. Because the 24-member group is small and self-selected, and the comparison divides an uneven dataset into 24 members and 117 non-members, these figures are descriptive associations rather than a controlled estimate of an effect caused by membership.
Where membership and performance do not align
Among non-members with all six dimensions, the 2026-10-02 dataset includes University of Bath at 87.4, University of St Andrews at 87.0, University of Surrey at 85.4, University of Aberdeen at 84.1 and Loughborough University at 84.0. The lowest-scoring Russell Group members are Queen’s University Belfast at 82.3 with four dimensions, University of York at 83.7, Cardiff University at 84.4, and the University of Glasgow and University of Liverpool at 85.0.
Bath, St Andrews and Surrey therefore outscore every member in that lower-member group. The label is not a complete ordering of institutional results.
However, no member sits below the non-member overall median in this snapshot. The non-member median is 71.0, while the lowest member score is 82.3. An example of a member below the non-member median would therefore be inaccurate; the evidence supports overlap at the upper end, not a median crossover.
Coverage tiers and the limits of the overall score
According to the 2026-10-02 unirank dataset, 23 of the 24 members have all six dimensions, while Queen’s University Belfast has four. Non-members have between one and six valid dimensions; 43 have five and 66 have six. The Institute of Cancer Research has the highest overall figure in the dataset at 89.5, but it is based on only one dimension and therefore occupies the lowest coverage tier.
That is why the coverage-tier rule matters: a 0–100 overall score with one valid dimension is not equivalent evidence of breadth to a six-dimension score. Queen’s Belfast’s four-dimension result and the Institute of Cancer Research’s one-dimension result must retain their coverage labels when compared with fully covered records.
The data can describe broad group differences and overlaps at institution level. It cannot establish a course-level result, identify the best option for a particular applicant or predict an individual’s earnings and employment outcome. The membership list and indexed data can also change, so these findings remain a dated snapshot rather than a permanent ranking.
Source-access note
The Russell Group pages were checked 2026-10-02. A rendering extraction of “Our universities” returned HTTP 200-equivalent, while a rendering extraction of “Who we are” also succeeded. Plain command-line requests from this host returned HTTP 403 for both. The official pages remain live but block that form of automated access.