GCSE results in Guernsey, 2008–2026.

On the published evidence: ending selection redistributed attainment decisively and did not raise it. Three cohorts cannot carry a stronger claim than that, and the island has since stopped publishing two of the measures that would have told us. What it cost to rebuild the school estate around the decision is a fourth thing the record does not settle, and that one is not a measurement problem: the States keeps its books in a way that makes the answer unrecoverable.

1. 01

   The gap between Guernsey’s States schools was made by the eleven-plus, not by the schools.

   In 2023, the last selective cohort, Attainment 8 across the mainstream States secondaries ran from 30.6 to 63.4. In 2024, the first non-selective cohort, it ran from 35.6 to 42.7. The spread collapsed from 32.8 points to 7.1 in a single year. No school changed its teaching overnight and no child moved house. What changed is that the top quarter stopped being removed.

   Spread of mainstream States-school Attainment 8 · 2023 32.8 pts → 2024 7.1 pts

   Evidence that selection, not school quality, produced most of the old gap.

2. 02

   Island-wide, attainment has fallen since the cutover, and faster than England’s.

   Ending selection changes which school a child attends, not who sits the exam, so the island total is the only fair test of whether it raised attainment. States-maintained Attainment 8 has fallen every year since 2023, from 39.6 to 37.4. The Bailiwick figure has fallen from 47.3 to 44.9. Over the same years England moved from 46.2 to 46.0. Guernsey’s decline is roughly ten times England’s.

   2023 → 2026 · States-maintained −2.2 · Bailiwick −2.4 · England −0.2 (to 2025)

   Evidence against, but heavily confounded. See the caveats below.

3. 03

   The gap that survived is the one bought with money, and it has never been published.

   The distance between the Bailiwick average and the States-maintained average was 7.2 points in 2019 and 7.5 points in 2026. It has not closed, because the selection that remains is selection by fee: 30% of Guernsey’s Year 11 is at the grant-aided colleges against about 7% in England. Guernsey ended selection at 11 for roughly 70% of its children and left the other selection untouched. What it did to the route into that sector is set out below.

   Bailiwick minus States-maintained · 2019 7.2 pts → 2026 7.5 pts

   The reform did not touch the largest remaining sorting mechanism.

1. 2016–17The States vote to end selection at 11, confirming it in a second vote the following March.
2. 2018The last 11-plus is sat.
3. 2019The mainstream secondary schools take their first all-ability Year 7 intakes.
4. 2023The last selectively-admitted pupils leave. Guernsey Grammar School & Sixth Form Centre is renamed Les Varendes High School, and every States 11-16 school is now non-selective. That makes 2023 the last GCSE year whose cohort included grammar-selected pupils, and 2024 the first fully non-selective one.

Read this carefully. The figures above are island-wide and cover the same pupils either way — ending selection changes which school a child attends, not who sits the exams. So the fall in Attainment 8 after 2023 is not mechanically explained by the reform, and equally cannot be attributed to it without evidence this page does not have. It is shown as context for the timeline, not as cause and effect.

The former Grammar School, renamed Les Varendes, fell 27.8 points because its intake changed. Every school that had always been non-selective rose, by between 5.0 and 8.9 points, because theirs did too. Both movements are the same event seen from opposite sides, and neither is evidence that any school taught better or worse.

Bracketed figures are the value as first published, before a later release restated it without a correction notice. gov.gg’s 2026 release publishes no school-level breakdown at all, so this panel stops at 2025.

2012 from the Education Department’s 2011-12 annual report; 2013-2016 from the 2016 release’s own comparison table; 2018-2022 from the 2022 release’s. 2017 is the only missing year. Across this whole run the Grammar School sat between 44 and 73 percentage points above the lowest high school.

States-maintained schools, excluding Les Voies

Down 9.6 percentage points in five years, and 4.9 of that came between 2023 and 2024, the cutover. gov.gg’s own caveat applies to the shaded years: 2020 and 2021 had no external exams and 2022 was graded leniently, so those three should not be compared directly with the rest. The fall from 2022 onward is not affected by that caveat. The 2025 and 2026 releases report English and Maths separately as grade distributions instead, so this student-level measure has no continuation.

“the Bailiwick figures include results from the 3 grant-aided Colleges but the Education Department does not report the Colleges’ individual results”

Three bases appear in these tables and they are not interchangeable. Percentage of grades at a band is comparable with gov.gg’s own cumulative distributions; percentage of students reaching five or more is not. Each cell carries only what its source actually states, and blanks are left blank rather than estimated. Only one of the three colleges publishes a full grade table; the other two reach the record through results-day press coverage, which is why two of these rows disagree with a second outlet reporting the same cohort.

Guernsey abolished selection in the one place it was publicly funded and blind to household income, and left it standing in the place it is priced. The colleges’ share of the cohort did not move. The route into it did.

1. 01

   Fifty-two funded places a year were abolished. The places themselves were not.

   Before 2019 the States paid college fees in full for a group it defined as the 52 highest achieving students in each year group of the 25% of students who, as a result of their 11+ examination results, would have been offered a place at the then Grammar School, but who expressed a preference to attend one of the Colleges. That was 52 pupils a year, about 352 in the system at once. It was academic selection, but it was publicly funded and it did not ask what a household earned. The scheme closed to new entrants in September 2019 and the last holder left in 2024/2025. ESC records what became of those places: “collectively, the Colleges have replaced all the Special Place Holders with fee-paying students”.

   States-funded places at the colleges · 52 a year until September 2018 → nil from September 2019

   The only route into the colleges that did not depend on household income was removed, and nothing replaced it.

2. 02

   The number of college places barely moved. Who pays for them changed completely.

   Total places rose 9%, from 1,150 in 2019/20 to 1,256 in 2024/25, which is why ESC can say the colleges’ share of secondary-aged pupils “notwithstanding the removal of the Special Place Holder scheme, the Colleges “market share” of secondary aged students has remained stable at c30%”. Inside that near-flat total, the payer switched. States-funded places fell from 286 to 39, a loss of 247, while fee-paying places rose from 864 to 1,217, a gain of 353. The two cancel to the 106 extra places. So the 41% rise in fee-paying numbers is not 41% more children at private schools. It is the same children, with the States no longer paying for 247 of them.

   Total places 2019/20 1,150 → 2024/25 1,256 (+9%) · of which States-funded 286 → 39, fee-paying 864 → 1,217

   No exodus: the island did not buy a third more private places. What it did was stop funding the places it had been funding.

3. 03

   The colleges still test at 11, and the children who pass are the same children.

   Each college screens applicants by cognitive assessment, which ESC says “that an assessment takes place prior to admission challenges any suggestion from the Colleges that they are fully “comprehensive” in nature”. Of the 269 pupils who moved from a mainstream States primary to a college between 2022 to 2024, 4 were assessed below the age-related average and 121, or 45%, were at least above average. Standard age scores above 126 are categorised “very high”, the highest achieving 4% of students across England; “above average” is the next 19%.

   States-primary movers to a college, 2022 to 2024 · 4 of 269 below average (1.5%)

   Academic selection at the point of secondary transfer continues. It is administered by the colleges rather than the States.

4. 04

   The States quadrupled what it pays towards the route it kept.

   General grant-aid to the three colleges, which is separate from the Special Place Holder fees, rose from £732,356 in 2019/2020 to £2,535,629 in 2024/2025, an increase of 246%. On RPIX alone it would have reached £915,445. Per pupil that is £637 to £2,019. The escalator was set by an amendment carried in the same 2017 debate that settled the post-selection arrangements. In March 2025 ESC asked the States to reduce it to nil by 2031/32 and was refused; the Assembly agreed instead to taper it by 4% a year in real terms to 75.1% of its current value by 2032/33, capped at the equivalent of 1,256 pupils.

   General grant-aid · 2019/2020 £732,356 → 2024/2025 £2,535,629 · RPIX-only counterfactual £915,445

   The merit route closed and the subsidy to the fee route grew, over the same six years.

One arithmetic note, in keeping with the restatements listed at the foot of this page. The policy letter gives the roll as rising from 1,147 to 1,256 and describes that as “an increase of 106 students”. It is 109. The figure of 106 is used again in section 10, measured from a different start year.

Equivalised median household income, children in Years 7 to 11, 2022

There is no overlap and nothing in between. The gap between the college with the lowest household incomes and the States school with the highest is £32,502. ESC’s own summary is that the system “is characterised by a high degree of social separation”.

Two cautions. These are equivalised medians, adjusted for household size and composition, so they are not interchangeable with the island’s gross median household income used elsewhere on this page as an England comparator; the resemblance between the two is arithmetic coincidence, not a match. And only the schools ESC names in its text appear here. St Sampson’s High and the smaller schools are plotted in the policy letter’s Figure 13, which is a flat image with no data behind it, so their values are left out rather than read off a chart.

Source: States of Guernsey, Committee for Education, Sport & Culture, P.2025/26 "Blanchelande College, Elizabeth College & The Ladies' College: Future Funding Arrangements"; States of Guernsey, "Debate on Future Funding for the Grant-Aided Colleges"; States of Guernsey, "The Role of the Grant-Aided Colleges and their Future Funding Arrangements" (September 2017).

Source: States of Guernsey, Committee for Education, Sport & Culture, P.2025/26 "Blanchelande College, Elizabeth College & The Ladies' College: Future Funding Arrangements"; States of Guernsey, "Debate on Future Funding for the Grant-Aided Colleges"; States of Guernsey, "The Role of the Grant-Aided Colleges and their Future Funding Arrangements" (September 2017).

Guernsey has not published what this reform has cost. Not because the figures are secret, but because of how the States kept its books while the money was being spent: until 2021 capital spending was written off in the year it happened, so the money spent designing schools that were never built was expensed at the time and left nothing to impair once the States began carrying assets on its balance sheet. Impairments do appear in the accounts now, but as one aggregate line with no project named.

Source: States of Guernsey, P.2021/69 "Secondary and Post 16 Education Reorganisation"; States of Guernsey, "Committee for ESC responds to amendments and concerns" (2021); States of Guernsey, "ESC - All 2 school combinations cost more than 3 School" (1 March 2018); States of Guernsey, "ESC details the costs of the new model of education in the Bailiwick" (2 July 2019); States of Guernsey Accounts 2021, accounting policy xvii (capital expenditure; superseded from the 2022 Accounts); States of Guernsey Accounts 2025, Note 11 (Property, Plant & Equipment Impairments); Scrutiny Management Committee, Letter of Comment on the States of Guernsey Accounts 2025 (19 June 2026). Figures tagged “press report only” reached the public record through dated reporting on statements the Committee made to journalists, because no States publication carries them.

Almost nobody promised better results.

The case made in 2016 rests on fairness, on the inaccuracy of sorting children at ten, and on the harm of labelling them. Deputy Sillars came closest with "full potential", and that is framed as opportunity rather than outcome. This cuts both ways. It is unfair to mark the reform against a promise its advocates largely did not make. It also means that the question St Pier asked as he moved it was never answered by anyone, so there is no agreed standard against which it could be marked at all.

Source: States of Guernsey, Official Report, Tuesday 8th March 2016 (amendments A1 and A2 laid); States of Guernsey, Official Report, Wednesday 9th March 2016 (debate and the vote); States of Guernsey, "Education publishes proposals for future structure of secondary and post-16 education" (2 February 2016); BBC News, "Guernsey 11-plus vote: End of selection confirmed by States vote" (2 December 2016); Bailiwick Express, "Education: How we got here, in the President’s words" (16 May 2023). The Official Report is published as a PDF at a stable URL per sitting; references are to the sitting date rather than a page number, because the document repaginates between printings.

At grade 4 and above the Bailiwick clears England by 4.2 points and the States sector falls short of it by 7.3. At the top of the ladder the divergence is starker still: 22.5% of Bailiwick grades are at 9-7 against 10.0% of States-maintained grades. These are percentages of all qualification entries, not of students.

Reports are published at gov.gg, not in Ofsted’s own public database — searching reports.ofsted.gov.uk for a Guernsey school returns nothing usable. Before 2022 Guernsey ran roughly two decades of validated self-evaluation against Education Scotland’s framework; those reviews used a different method and are not shown here. Coverage gap: Notre Dame du Rosaire was also inspected in March 2023, but that report is not on gov.gg's archive page and could not be retrieved.

Guernsey’s 2022 household income survey put median gross household income at £55,116 (mean £74,568) — a household, not per-head, figure, so it isn’t directly interchangeable with the series below. The nearest granular England equivalent is ONS’s gross disposable household income (GDHI) per head, by local authority: on that measure, Windsor & Maidenhead (£38,211), Wokingham (£32,259) and Richmond upon Thames (£51,388) — all 2024 — sit well above the roughly £25,000–26,000 England/UK average. All three are also fully comprehensive systems, like post-reform Guernsey: no grammar schools, so — unlike Buckinghamshire or nearby Kingston and Sutton, similarly affluent but selective — income and selection aren’t tangled together in this comparison.

Guernsey’s figure is 2026; the England figures are 2023/24, the latest year local-authority-level Attainment 8 has been published as a downloadable dataset — a year’s gap, not a strictly like-for-like single year, though Attainment 8 does not typically move sharply year to year.

1. 01

   Publish the three grant-aided colleges' results.

   Blocking

   They are 30% of the Year 11 cohort and are already inside the headline Bailiwick figure. Reporting a total while withholding a third of its components means no reader can check it, and no comparison with England is on like terms.

   Status: Exists. The colleges hold it, and one of the three already publishes a full grade table voluntarily.

2. 02

   Publish a progress measure, not only an attainment one.

   Blocking

   Attainment 8 is raw attainment. England pairs it with Progress 8, which measures what a school added to the pupils it received. Without it, no published Guernsey figure can separate a school’s contribution from its intake - which is precisely the question selection raises.

   Status: Requires Key Stage 2 baselines to be linked to Key Stage 4 outcomes. Guernsey holds both.

3. 03

   Publish the sector shares, year by year, back to 2015.

   Material

   This is the control variable for every claim made about the reform. ESC’s 2025 funding policy letter states the colleges’ share “has remained stable at c30%”, which is the answer, and it is the reason the island-wide fall since 2023 cannot be written off as compositional. But a rounded sentence in a funding paper is not a series: it cannot be checked, and the year the share moved, if it moved, cannot be located.

   Status: Exists. ESC holds it and cites it. The roll appears in that letter’s Figure 1 as an image; the share itself appears only in prose.

4. 04

   Restore the grade 4+ in both English and Maths measure.

   Material

   It ran unbroken from 2019 to 2024, spanning both the pandemic and the cutover on one basis, and was dropped the year after it reached its lowest published value. Its absence is the largest gap in the post-reform record.

   Status: Published as recently as the 2024 release. Discontinued without explanation.

5. 05

   Publish school-level Attainment 8 for 2026.

   Material

   The 2026 release gives only the three-row island table, so the first year in which the reform could show a settled effect has no school-level breakdown at all.

   Status: Promised. The 2026 release says a fuller breakdown will appear in the 2026 Education Annual Report.

6. 06

   Publish the 2017 results release.

   Gap

   It is the only missing year in an otherwise complete 2012-2022 school-level run on the old measure.

   Status: Exists. Not currently retrievable from the gov.gg media release archive.

7. 07

   Attach correction notices to restated figures.

   Gap

   At least fifteen published figures have been silently revised between releases, including every school’s 2023 Attainment 8 score. A reader who cited the first version has no way to learn it changed.

   Status: A publishing practice, not a data problem.

8. 08

   Publish cohort sizes alongside every school figure.

   Material

   Guernsey’s whole Year 11 is 624 pupils and a single school’s cohort can be under five. Without denominators, a swing of a few children is indistinguishable from a trend.

   Status: Exists. Island totals are published; school-level denominators are not.

Two further figures on this page are held back pending a check against the original tables: the 2023 Bailiwick share of all grades at 4+, which appears as both 73.9% and 76.4%. It is not used in any chart here.