Pathways · University destinations

Which UK Schools Actually Send Pupils to Russell Group Universities?

“90% of leavers go to Russell Group universities” is the most repeated single statistic in UK independent schools marketing — and one of the least useful. This guide unpicks what that headline actually means, and surfaces the schools whose published per-pupil data lets you answer the more interesting question.

Published Updated 12 min read← All guides

Walk into any Sixth Form open evening and you will hear the Headmaster cite a Russell Group percentage — typically between 80% and 95% — within 20 minutes. The statistic is real. It is also almost entirely uninformative.

The Russell Group is 24 universities. They span Imperial College London (engineering tariff above 250 UCAS points) and Queen Mary (entries possible at well below half that). The difference between a school sending its leavers to Cambridge and a school sending its leavers to QM, both flagged as “Russell Group”, is enormous in life-outcome terms. Treating Russell Group as a single bucket is the marketing equivalent of saying “over 80% of our leavers walked into the building on results day”.

What we look at instead

On School Path AI we extract per-pupil university destinations from each school's own published documents — annual destinations PDFs, sixth-form results pages, ISI inspection reports. We then aggregate to compute P(university | school) for each route, with rate normalisation across multi-year vs single-year data.

The single hardest part of this work is data quality. Many schools publish only Russell Group totals or only top-3 named destinations. We classify every row at one of four confidence levels:

  • verified — extracted from a per-year published source.
  • verified-1yr — verified, but only one year of data (noisier).
  • verified-list — school publishes the list but not the counts.
  • training-knowledge — best-faith estimate. We do not display probabilities for routes built on training-knowledge data, even if it would make our charts look more impressive.

The result is that some famous schools (Eton, Harrow, Radley) sit in our database with training-knowledge data only — they don't publish per-pupil counts, and we won't pretend we have data we don't. We list those schools as connected nodes in the pathway view but withhold a probability number.

Where the data is strongest

Of the senior schools we've verified at the “high” data quality tier (≥3 verified rows AND ≥25 captured pupils per year), several stand out for the cleanness of their published data:

The Perse School (Cambridge, co-ed)

Perse's 2025 leavers' destinations cover 180 pupils across 36 destinations. The top of the list: Durham 23, UCL 17, Warwick 13, Cambridge 12, Oxford 11, Edinburgh 10, Imperial 10, York 10. This isn't a school that funnels to Oxbridge — it produces a balanced spread of Russell Group + selective non-RG destinations, with the bottom tail thinning out across smaller single-pupil destinations.

Magdalen College School Oxford (boys → co-ed)

MCS's 2020 published destinations cover 156 leavers. It is — as Oxford insiders will tell you — overwhelmingly a Cambridge feeder, not an Oxford one (despite the name). The verified data bears this out: 34 to Cambridge, only 8 to Oxford. Bristol 16 and Durham 17 round out the top four.

Westminster School (boys, central London)

Westminster's 2025 cohort had 72 Oxbridge offers — the most relevant single number we found. The school's published breakdown is 48 to Oxford and 24 to Cambridge, an unusual Oxford-skew that reflects a long historical tie. This skew is the opposite of MCS Oxford's, suggesting Oxbridge institutional identity matters more than geography in determining feeder patterns.

Tonbridge / Charterhouse / Marlborough (4-year multi-year totals)

These three schools have 4 years of verified per-pupil data (2021–2024) in our database, which makes them excellent reference points. Tonbridge's 4-year cumulative shows Durham as a dominant destination, with Edinburgh, Exeter, and Bristol close behind. Charterhouse has a similar pattern but with stronger Oxbridge representation. Marlborough's 2023 data uniquely shows Exeter (45 pupils) outranking everything else — the strongest single non-Oxbridge feeder relationship in our dataset.

Cardiff Sixth Form College (Cambridge campus)

CSFC publishes 10 years of per-year destinations — uniquely in our dataset. It is overwhelmingly an international-student London Russell Group pipeline: UCL 244 leavers cumulative, Imperial 160, LSE 130. If you've ever wondered “where do the children of internationally-mobile families end up”, the answer is largely: at UCL.

Three counter-intuitive findings

1 · “Top schools” rarely send 50%+ to Oxbridge

Even the most academically selective schools in the country — Westminster, Eton, Winchester, St Paul's — typically place 20–35% of pupils at Oxford or Cambridge in any given year. The rest go to Imperial, LSE, UCL, Durham, Edinburgh, KCL, top US universities. If a school's marketing strongly implies an Oxbridge guarantee, the data almost always disagrees.

2 · The interesting pipeline is the second tier

The largest verified flows in our data are not to Oxbridge but to Durham, Edinburgh, Bristol, Exeter, UCL, and Warwick. These are the universities where a top independent school's edge over the average shows up most clearly — Oxbridge is a near-binary pre-selection filter, while the next tier's admissions are responsive to school-quality signals across the application.

3 · Coverage is the silent variable

Many schools make headline claims (“90% Russell Group”) from data that, on inspection, only covers a fraction of leavers — scholarship recipients, Year-13 pupils with completed UCAS, or top achievers featured in marketing literature. Our coverage badge on each school's pathway view tells you what % of the cohort our verified data actually accounts for. When that's low, treat any headline percentage with scepticism — including ours.

How to read a school's destinations data

The scariest signal in school marketing is a single percentage with no denominator. Useful destinations data has all of:

  1. A clear cohort denominator (“180 leavers in 2025”) rather than just “90% Russell Group”.
  2. A multi-year window or, if a single year, an explicit statement that it is a single year.
  3. Per-university counts, not just “Russell Group / non-RG”.
  4. Some discussion of pupils who didn't apply to UK uni (apprenticeships, US, gap years, art college).
  5. A linkable PDF or page URL that doesn't change every year, so you can reference the exact source you're reading.

Schools meeting all five criteria in our database include the Perse, Magdalen College School Oxford, Trinity Croydon, The Leys, Cardiff Sixth Form College, Surbiton High, Wycombe Abbey, Abingdon, Tonbridge, Charterhouse, and Marlborough. Schools missing one or more — sometimes very famous schools — make it harder to ask the right questions.

The bigger picture

The Russell Group bucket is a marketing tool that aggregates 24 very different universities into one comfortable headline. The question worth asking is not “does this school send to the Russell Group” (almost all selective independents do) but “which Russell Group universities does it actually send to, and at what concentration?”. Our pathway probability data is built specifically to answer that — including by computing P(university | prep school) end-to-end where the data supports it.

Read more about the methodology on our Methodology page, or jump into the Pathway view to focus on a specific senior school. Both are free.