Changelog

Corrections, comparisons and a new home

Two Eton corrections, more years to compare and a much faster site

Two of Eton’s A-level figures now follow the school’s own signed tables, a fix to the comparison rules brings back years that should never have been left out, and the record is ready to move to its own, much faster address. No other figure in the record has changed.

Eton, A levels 2024. The A* share now reads 38.7% (362 of 935 grades), up from 38.6%. Eton re-signed its results table on 22 November 2024 after one Music Technology grade was raised from A to A*. The earlier table, signed on 16 October 2024, stays on the record marked as superseded, and correction C30 replaces C11, which had followed that earlier table.

Eton, A levels 2023. The 44.8% A*, 80.2% A*–A and 95.4% A*–B we showed as A-level results count Eton’s A-level and Pre-U grades together (930 grades): they reproduce exactly when the school’s two signed tables are added up. They are now labelled as combined, and comparisons use Eton’s A-level table on its own: 43.7%, 80.1% and 95.3% of 870 grades (correction C31).

More years you can compare. A publication that prints nothing for a grade band no longer knocks that year out of comparisons for the band. Nothing is estimated and no figure changed; years that were wrongly left out are simply back. Eton’s A-level A* now has 6 comparable years (it had 2), A*–A 4 (2) and A*–B 6 (3); St Paul’s A*–A has 14 (12) and A*–B 15 (12); King’s College School’s A*–A and A*–B have 9 each (8); Wycombe Abbey’s GCSE grade 9 has 2 (1); and Eton’s letter-grade GCSE A*–A and A*–B have 6 each (5). No school lost a year, and school profiles no longer count a year as both comparable and ‘sources differ’.

Verdicts no stronger than the evidence. The claim checker and the compare page now grade every verdict by one published rule: at least three matched years, a typical gap of at least 1 percentage point, and at least 60% of the tested periods pointing the same way. Anything weaker is called marginal, and with fewer than three years you see the values without a verdict. St Paul’s against Westminster on the A* share, 2010–2025, now reads ‘Marginally higher (0.7 pp); the direction depends on the years chosen’. Gaps are written one way throughout, to one decimal place, and each school keeps its own chart colour rather than green for ‘ahead’.

Tell us when a figure looks wrong. Every figure’s own page has ‘Report a problem with this figure’, which opens an email to our corrections address with the record already filled in. The corrections page now dates every entry and says which correction replaced which. About says who publishes the record and how quickly we reply, the privacy notice is rewritten for the new address, and a new accessibility statement lists every problem we know about and what we are doing about it.

A faster home of its own. The record is moving from its earlier address to theschoolsrecord.co.uk, served from Cloudflare’s network. On a slow phone connection a page now appears in about a second, where it could take more than 20 seconds before. Every page has one permanent address, a shared link shows a preview card, and links to the earlier address will lead straight to the same page.

Easier to use. Buttons and chart focus rings stay clearly visible in the dark theme; form fields are at least 16 px, so phones no longer zoom in; no text is smaller than 12 px; after ‘Check the claim’ or ‘Update comparison’ the page takes you to the answer; charts are announced as described groups with their source links intact; the phone menu sits above the compare editor and closes with Escape or a tap outside it; and numbers are set in lining figures, so columns line up.

The opening emblem. It now appears only when a visit starts on the home page, never when a link takes you straight to a school, a record or a tool, and never if your device is set to reduce motion. Any tap, click, scroll or key press dismisses it at once, and the tap that dismisses it no longer follows a link hidden beneath it.

Data fingerprint 7be200b8cae2 · frozen record unchanged · research-resolutions-2026-09 (resolution-policy-v1, resolutions c51e747cb78e) · comparison model 0f1cca6c8660 · two statistical changes, corrections C30 and C31.

Research questions

Seventeen Research questions

Research questions is a new section of short readings. Each starts with its answer and links every figure it uses to the record behind it. The seventeen published so far ask who is counted in a school percentage, what a top-grade share leaves out, why published Oxford totals differ and what should happen when publications disagree, with readings on Eton, King’s College School, St Paul’s, Westminster and Winchester.

Readings that draw on school histories, inspections or alumni publications say so, and keep those sources separate from the published results.

Live publication 5d41308362bf, checked on 22 September 2026 · data baseline 818c7c296fed · frozen record unchanged.

Research bar

Your shortlist follows you from page to page

A research bar now carries up to five schools, the measure, the years and an optional question across the site, so a comparison started on one school’s profile continues on the next page. It is kept in this browser only. A link you share carries the schools, measure, years and question, but never the notes in a saved comparison.

Publication ed46c1f1e975 · data baseline 818c7c296fed · no statistical change.

Theme

Choose a light or dark theme

A dark theme, added on 17 September and following your device setting, now has a switch in the header. Your choice is remembered in this browser and applied before the page appears. The chart colours were rebuilt for both themes and checked for colour-blind readers.

The university directory is now a compact index of all 130 universities, still searchable and still saying how many records each has.

Publications 83de15a7143c (17 September) and 6eccc2d5a13f (18 September) · data baseline 818c7c296fed · no statistical change.

Universities

A page for every university in the record

All 130 universities named in the schools’ publications now have their own page, where before only twelve did. A page without a written profile says so, and shows no location or link it does not hold.

Each school’s university records are now one set, with Oxford and Cambridge, and US and overseas, as views of it, so no record sits at an address you would have to know. Schools and Universities now lead the navigation.

Publications 86ef6e8ac3af and f8410023cf7e · data baseline 818c7c296fed · no statistical change.

Research and service preparation

Nineteen school profiles and twelve university evidence profiles

Added the eight named schools, individual university pages, credited institution images, searchable directories and clearer Home and return controls. A new St Paul’s/KCS example separates checked publication versions and explains the limits of a current shortlist claim.

The proposed adviser review has a bounded scope. Orders remain closed while seller, tax, receiver, payment and reviewer details are confirmed. Historical source recovery is continuing; the profiles do not imply complete annual coverage.

Original statistics and identifiers remain unchanged. New observations, source assessments and publication differences are recorded separately.

Extension corpus

Four schools added

Added Radley College, Harrow School, Latymer Upper School and Godolphin and Latymer School as an additive extension corpus: A level, GCSE and IGCSE, Pre-U and IB Diploma ledgers, Cambridge apply-centre cycles 2009–2025, leavers’ destinations, offers and itemised destination lists, each row retaining its source references; those references do not all resolve to a checked public publication. Extension records are tagged throughout and kept apart from the frozen figures. Oxford apply-centre cycles for the four schools remain an open gap.

Data baseline: 818c7c296fed · frozen record unchanged · extension-2026-09-four-schools added.

Accessible application restoration

The complete record, natively

Restored everything the final public product could do that the native rebuild had lost: an evidence centre that searches all 2,277 records and traces any displayed figure to its public sources; Oxford and Cambridge and US-university records for every school and corpus-wide; school-entry evidence; the corrections and published-differences ledgers; ledgers that show every published denominator; approved public source links; form acknowledgement handling, whose live retrieval was not established; page titles and shareable metadata.

Data baseline: 818c7c296fed · no statistical change.

Charts

Results by year

Added a like-for-like series panel to every row of the school index on one shared measure, a results-by-year section to each school record with one panel per qualification scale, and the comparison chart to the sample dossier. Lines join consecutive published years only; 2020–21 and the 2010 introduction of A* stay marked.

Data baseline: 818c7c296fed · no statistical change.

Historical collection update

Canonical public application

Introduced permanent server-rendered routes, responsive assets, typed schemas, privacy separation, parity gates, comparison URLs, professional offer and expanded quality assurance.

Data baseline: 818c7c296fed · no statistical change.

Reviewed public experience

Language and presentation pass

Improved homepage, navigation, comparison terminology, evidence language and responsive presentation in the earlier implementation.