Draft worked example · St Paul’s and King’s College School, Wimbledon

Does the published difference support the claim?

Does the recent published A-level A* difference support a claim that St Paul's consistently outperforms King's by about ten percentage points?

A-level entries awarded A* · 2022–2025; 2026 checked separately · version TSR-SPS-KCS-20260906-v1 · report date 2026-09-06 · draft · final founder review and approval outstanding

The answer

St Paul’s A* share is higher in all four years checked — but the gap ranges from 0.7 to 10.3 points, not a steady ten.

No. The checked publication versions show a positive St Paul's-minus-King's difference in all four years, but its size varies substantially. The current-page comparison gives 1.9 and 0.7 percentage points in 2022–23, then 10.30 and 8.55 in 2024–25. Publication disagreements mean the exact series is provisional. Four years are insufficient for the engine's five-year robustness assessment.

Insufficient evidence for a robustness verdict. Four paired years from selected publication versions do not meet the five-year threshold, and the calculation below does not resolve their disagreements. The free comparison excludes unresolved competing values, so it can show fewer eligible matched years than this version-specific worked example; that difference is explained, not hidden.

The paid increment, on this question

What the free tools give you, and what the review added.

Free claim checker · same question, same record

There is not enough comparable history.

There is 1 matched year; this check requires at least 3. Missing evidence is not a tie or a zero.

Matched years 2022–2025: 1
Years it cannot use: 3

Held back by the evidence gate: 2022 (competing published values); 2023 (competing published values); 2024 (competing published values). The free tools never choose between competing published values.

Run this question yourself

Commissioned review · the same question, taken to the publications

Four years compared on named publication versions; median 5.2 percentage points.

The review retrieved 7 original publications, transcribed each year’s A* share from a named version, recorded the 4 places where publications disagree rather than choosing between them, and drafted the wording, limitations and school questions below. A reviewer’s approval of it is still outstanding, and the page says so.

Matched years 2022–2025: 4
Publication conflicts recorded: 4

Robustness: Insufficient evidence. Four years are still fewer than the five the period check needs; the review says so rather than upgrading the verdict.

The free result is not wrong: it refuses years the record cannot settle. The review is the work of settling what can be settled from the publications, and saying plainly what cannot.

What you could say to the client

The recent published A-level A* figures are higher at St Paul's in the four years examined, with a much larger reported difference in 2024–25. Differences in publication versions remain unresolved. These school-level grade shares do not establish individual prospects or how much a school adds to attainment.

What a commissioned review adds beyond the free comparison

An agreed question, original-publication checks, explicit version conflicts, limits on the proposed client claim, questions for the schools, a reproducible calculation and reviewed handover. This example demonstrates research beyond the automatic free comparison.

The checked publication versions

A-level A* percentage of subject entries; St Paul’s minus King’s, to one decimal place (the calculation CSV keeps each exact difference)
Year	St Paul’s	King’s	Difference	Publication sources
2022	70%	68.1%	1.9 pp	St Paul’s source King’s sourceEarlier stored observations: St Paul’s record · King’s record
2023	51%	50.3%	0.7 pp	St Paul’s source King’s sourceEarlier stored observations: St Paul’s record · King’s record
2024	61%	50.70%	10.3 pp	St Paul’s source King’s sourceEarlier stored observations: St Paul’s record · King’s record
2025	59.3%	50.75%	8.6 pp	St Paul’s source King’s sourceEarlier stored observations: St Paul’s record · King’s record

Median of these four annual differences: 5.2 percentage points. The unrounded arithmetic is 5.225; the published figures do not justify that many measurement digits.

Each difference is St Paul’s percentage minus King’s percentage in the same year, on one definition throughout (A-level A* share of subject entries). The median is the average of the middle two ordered differences. It is not a pooled entry rate or pupil outcome.

Why this differs from some stored records

The original stored figures are preserved. For 2022, 2023 and 2024, St Paul’s current summary prints 2022: 70%; 2023: 51%; 2024: 61%, while the stored series contains 2022: 70.9%; 2023: 50.6%; 2024: 60.2%. King’s 2025 retrospective also differs from earlier versions for 2023. The reasons remain unresolved; this report names the versions it used and does not choose a definitive corrected history. Using the stored St Paul’s values instead gives a median of 5.675 pp.

Where the publications disagree, and what was done about it

Each disagreement is recorded with every competing value and its source. None is silently resolved; each carries the action a reviewer would take next.

Conflicts investigated in this review
School · year	Competing published values	Cause	Status and next action
St Paul’s · 2022	70.9% (fig:st_pauls_alevel:26); 70% (original checked)	Unresolved; no dated amendment or matching original precision recovered in this bounded check.	Unresolved. Locate original detailed 2022 table; do not assume ordinary rounding or overwrite the frozen value.
St Paul’s · 2024	60.2% (fig:st_pauls_alevel:28); 61% (original checked)	Unresolved; current summary differs from frozen precise figure.	Unresolved. Locate dated detailed 2024 table and any revision statement.
King’s · 2023	50.1% (original checked); 50.1% (original checked); 50.3% (original checked)	Unresolved. 2023 detailed table gives 288 A* of 575 entries; 2025 retrospective gives 50.3 without amended counts.	Unresolved. Retain both; ask school about revised counts only after owner approval if required.
St Paul’s · 2025	57% (original checked, 2025-08-14); 59.3% (original checked)	Results-day news and current detailed result page differ; a revision explanation was not established.	Unresolved. Use the explicitly named current 2025 results-page version for this descriptive example; do not assert a proven re-mark cause.

What about 2026?

St Paul's 13 August 2026 publication reports 60% A*. King's 13 August 2026 publication says only that more than half of A-level grades are A*. Its 56% figure includes IB equivalents. The current King's results hub links detailed tables only through 2025. Therefore no exact 2026 difference is calculated; this is a dated retrieval limit, not a claim of non-publication.

Material limitations

The named publication versions are analysed descriptively; none is declared the definitive revised series.
King's has a separate IB population. Its 2025 A-level PDF identifies 182 A-level-route candidates plus one IB pupil taking an A level. The 666 A-level entries are not 666 pupils, and the whole sixth-form cohort is not described.
Whole-percent St Paul's summary values and more precise King's percentages have different reporting precision. 5.225 is arithmetic on printed values, not that level of measurement accuracy.
No pupil intake, subject mix, individual grades, teaching effect or probability of admission is estimated. 2020–21 assessed results are excluded; 2022 is a transitional award year and is compared within the same calendar year.
No earlier historical conclusion is offered: older St Paul's observations were not originally verified in this bounded sample pass. This avoids presenting internally reproducible historical rows as source-checked research.

Questions that move the advice forward

Which detailed final A-level table and issue date should control 2022 and 2024 at St Paul's? Were entry populations or reviewed grades different?
Why does King's 2025 retrospective give 50.3% A* for 2023, while the 2023 and 2024 PDFs give 50.1%? Are amended counts available?
Can both schools supply the final 2026 A-level-only grade distributions and subject-entry counts?
For the client's intended subjects, what are the actual entry counts and grade distributions, and which sixth-form route do they concern?

Take the evidence with you

These are the deliverables a commissioned review hands over: the report as PDF, the reproducible calculation as CSV, and the versioned inputs so a later change on a school’s website cannot silently alter what was issued.

Download the sample PDF Calculation CSV Versioned calculation inputs

Draft worked example · source-checked · final founder review and approval outstanding. This is not a signed client report.

Permitted use of a commissioned report: An adviser may share the complete commissioned report and relevant attributed extracts/CSV for the agreed named client brief, retaining qualifications and version/date. No bulk redistribution of source publications or general dataset licence is included.

Report date: 2026-09-06. Later source changes do not silently change this issued version; a correction would be issued as a new version with its reason and effect.