Data Sources & Disclaimer

What you should never do with our data

• Do not treat any probability or score as a prediction for an individual child. Every figure describes historical cohorts.
• Do not use our rankings as the only input into a major schooling decision. Visit the schools, talk to admissions staff, get advice tailored to the child.
• Do not assume any school listed has endorsed, partnered with, or paid for placement on our service. None of them have.

Confidence labels — what each one means

• verified — the count comes from a per-year published source: the school's annual destinations PDF, an ISI inspection report, an explicit table on the school's sixth-form page. We have a URL or PDF reference on file.
• verified-1yr — verified, but only one academic year of data is available so the figure is noisier than a multi-year average.
• verified-list — the school has published a list of destination schools or universities but not per-row counts. We use a placeholder weight so the route appears in the chart; we do not claim it as a precise rate.
• training-knowledge — a best-faith estimate generated from public reputation signals when no published source could be located. This is the weakest tier. Probabilities are not displayed for routes that go through training-knowledge seniors.

Data quality gating on the Pathways page

When you focus on a single prep school, the Top end-to-end pathways panel only shows percentages backed by senior schools meeting both:

• ≥ 3 verified rows in our destinations table
• ≥ 25 captured pupils per year on average

Universities reachable via lower-quality routes are listed at the bottom of the panel as text chips, deliberately without numbers. If we showed you 12% on a route built from 5 estimated rows for a 150-pupil cohort, we would be misleading you — so we don't.

The reliable coverage badge tells you what fraction of the prep's leavers flow into seniors that meet the gate. When this number is low, take the headline percentages as best-of-the-known, not the full picture.

Pathway probability methodology

We model the journey as a 3-state Markov chain (Primary → Senior → University) and combine the two legs by the law of total probability:

P(uni = u | primary = p)
   = Σ_seniors  P(senior = s | primary = p) × P(uni = u | senior = s)

Each row is rate-normalised by the number of academic years it covers, so a 5-year cumulative source and a 1-year source compare fairly. See Methodology for the full explanation.

We do not predict whether your child will go to university at all — apprenticeships, gap years, and direct work are not in our data. The probabilities are conditional on going to a UK university.

Where the underlying data comes from

• DfE GIAS / Performance Tables — the spine of our school directory and exam performance metrics.
• ISI inspection reports — quality ratings for independent schools, scraped from isi.net/reports.
• Each school's own destinations page or PDF — we follow the URL printed on the “Where do leavers go?” page and store it as the evidence record. You can click through from any link tooltip on the Sankey to see the source.
• Curated tier lists — a small set of hand-built reference files in our repo, mainly to anchor the senior school prestige_score.

What we do NOT use

• We do not buy paywalled datasets like the Times Schools Guide or Tatler ratings.
• We do not accept payment from schools to influence rankings.
• We do not use any personal data from our users to compute scores. The Recommend feature uses only the attributes you tell it (entry stage, region, budget) and never sees a name or address.

Reporting an error

Found a number that looks wrong? Click the “Contest this data” link on any school detail page (coming soon — until then email data@schoolpath.ai). Include the URL and what you believe the correct figure is. We will acknowledge within 5 business days and either correct, downgrade the confidence label, or explain our source.

Bottom line

Use School Path AI to generate hypotheses, narrow your shortlist, and prepare smart questions for the schools you visit. The numbers are honest descriptions of the past — they are not promises about the future.