Methodology & sourcing
What this site is, where the numbers come from, and what they don't tell you.
What this is
England's Department for Education and the Ministry of Justice both publish statistics about SEND (Special Educational Needs and Disabilities) tribunal appeals — but only at national level. Neither publishes a breakdown by individual local authority of how many appeals a council won, lost, or had conceded before a hearing, or what kind of decision was most often overturned. That breakdown exists — HM Courts & Tribunals Service holds it centrally and will disclose it on request — it just isn't proactively published as open data.
A number of parents, campaigners and journalists have already asked individual councils this exact question under the Freedom of Information Act, and published the responses on WhatDoTheyKnow, the UK's public FOI request archive run by the charity mySociety. This site is an aggregation of those already-public responses into one place, organised by council — nothing more. It is not a new disclosure, not an official statistic, and not affiliated with any council, the DfE, the MoJ, or HMCTS.
What "data available" actually means
Every council on this site is marked with one of these statuses:
| Status | What it means |
|---|---|
| Data available | At least one FOI response gives real, transcribed figures on appeals, outcomes, or decision types for this council. |
| Answered — document pending | A council has confirmed answering this exact question, but supplied the figures as a PDF, Word or Excel attachment that automated extraction could not read. The source is linked — the answer exists, nobody has typed it up yet. |
| No record found | No FOI request covering this council's SEND tribunal record was found. A genuine gap. |
Coverage is uneven by design, not by omission: it reflects whatever has actually been asked, by whoever happened to ask it, whenever they asked it. Some councils have been asked by five different people across five different years and give a rich picture; some have never been asked at all.
Why there's no single "win rate"
The original framing behind this site was a simple win/loss percentage per council. The underlying responses don't support that as cleanly as hoped. Several councils have told FOI requesters directly that they don't categorise appeals as simply "won" or "lost", because most SEND appeals are resolved by a consent order — an agreement reached between the council and the family — rather than a contested tribunal ruling with a declared winner. Where a council's response distinguishes conceded-before-hearing from a contested decision, that distinction is preserved on this site rather than collapsed into one number. Where a council's own figures are ambiguous or don't reconcile with another source for the same year, that's noted rather than silently resolved one way.
Known data-quality caveats
These are specific, named issues surfaced while assembling this site — kept visible rather than quietly corrected:
- Some councils' own figures don't internally reconcile (e.g. a "conceded" count exceeding the "registered" count for the same year in one council's response).
- Where two independent FOI responses to the same council disagree for an overlapping year, both are shown rather than one being preferred.
- Terminology varies: some councils' responses appear to conflate "appeals" with "mediations", or don't clarify whether "upheld" means in the parent's or the council's favour. Where that ambiguity exists in the source, it's carried through rather than resolved by assumption.
- Local government reorganisation has changed some council boundaries since a response was given (for example, Surrey split into two new unitary authorities in 2026) — older data is attributed to the authority as it existed at the time.
Licensing and re-use
FOI responses come with mixed re-use terms. Some councils release information under the Open Government Licence; others state it may only be used for non-commercial research or news reporting, with commercial re-use requiring the council's separate permission. Where a response stated its terms, they're shown alongside it. Where nothing was stated, no licence should be assumed.
How the England-wide figures and per-council estimates are built
The England analysis pages and any council marked "Estimated, not reported" use a deliberately simple, disclosed method — not a black-box model. Two separate things are combined:
- Real appeal volume, for every council. The DfE publishes SEND tribunal appeal counts by local authority for 2024 and 2025 (SEND tribunal appealable decisions rate 2024–2025, sourced from HMCTS's own case system). This is an actual government figure for all 153 councils, not an estimate — no volume number on this site is invented.
- An outcome-rate profile, pooled from real FOI responses. Of the 115 councils with usable FOI figures, only a minority gave a response detailed enough to state all four outcome buckets (resolved before hearing, contested-and-lost, contested-and-won, partially split) — as of the last build, 34 councils, covering roughly 3,800 pooled appeal outcomes. Those real counts are pooled to compute a rate: what proportion of appeals, across that real evidence base, fall into each bucket.
For a council with no usable outcome breakdown of its own, that pooled rate is applied to that council's own real DfE appeal volume to produce an estimate. Regional pooling is used only where a region has at least six real qualifying records — as of this build, that's the North West only. Every other region, and the England-wide total, uses the national pooled rate instead, because the regional sample would otherwise be two to five councils dressed up as a trend.
What this deliberately doesn't do: it doesn't model council-specific factors (deprivation, SEND team capacity, legal spend, anything else that might make one council's real rate differ from the pool). It assumes the pooled rate is a reasonable stand-in absent better information — which is a real limitation, not a hidden one. The 34 councils behind the pooled rate are not a random sample of England; they're whichever councils happened to be asked a detailed-enough question and answered it clearly. Treat every estimated figure as an informed approximation, not a disclosure — and check the real source data behind the pooled rate, linked from the analysis pages, before relying on it for anything that matters.
Corrections and removal
If a figure here is wrong, missing important context, or you are a council officer who can supply a fuller or more current answer, this is a preview build without a live contact address yet — that will be added before this site is public.