What actually gets overturned
Not all SEND appeals are the same fight. One category — a council refusing to even assess a child — gets overturned four times out of five.
A SEND tribunal appeal is never just "an appeal" — it's always about one specific kind of decision. A council can refuse to assess a child at all; refuse to issue a plan after assessing; get the description of a child's needs wrong; get the provision wrong; or name the wrong school. Half of the 115 councils with usable Freedom of Information data — 50 of them — broke their figures down this way, and pooling what they actually said produces a clear, lopsided picture.
Where the appeals actually are
"Other / bundled" isn't a real category — it's where a council reported several sections together without separating them, and is left out of the rest of this piece for that reason.
Where councils actually lose
Volume is one question. What happens when each type of decision is actually challenged is a different one — and the two don't move together.
Based on real, council-stated figures only — categories with fewer than 15 reported outcomes are left out as too small to trust.
A council that refuses to even assess a child is overturned on that refusal about 4 times out of 5.
Refusal to assess is both the single most common reason families end up at tribunal and the decision type councils are worst at defending — overturned in roughly 81% of the cases where an outcome was reported, a real figure built from 687 reported outcomes across dozens of councils, not a handful. That's a genuinely large, well-evidenced gap between how often councils make this call and how often it turns out to be the wrong one.
Compare that with refusal to issue an EHC plan — the next stage, once an assessment has actually happened — where councils are overturned only around 16% of the time. Once a council has assessed a child and decided a plan still isn't warranted, that decision holds up far more often than the earlier refusal to assess in the first place. Section I appeals — disputes over which school is named — sit in between, overturned roughly half the time.
Why this matters more than a single win rate
A council publishing "we win most of our tribunal cases" and a council publishing "we lose most of our tribunal cases" could both be telling the truth, depending on which decisions they're mostly being challenged on. A council whose disputes are mostly about assessment refusals should expect to lose more often than one whose disputes are mostly about which school is named — not because one council is worse at SEND than the other, but because tribunals treat those two questions very differently. Comparing councils on a single blended win rate, without knowing the mix of what's actually being appealed, compares two different things as if they were one.
A note on where these numbers come from
These figures are pooled directly from what councils themselves stated in Freedom of Information responses — not modelled or estimated. The overturn-rate figures only include categories with at least 15 reported outcomes, to keep a single small council from swinging the picture. One council's figures were excluded from the overturn-rate calculation specifically: its own reported numbers showed a suspicious pattern already flagged elsewhere on this site (identical appeals-received and appeals-overturned counts in every year given), which reads as a likely reporting or data-conflation issue rather than a genuine result. See the methodology page for the full picture, and every council's own source data for the underlying figures.