Someone counted every Ofsted comment on science. The complaints are short and familiar.

Ofsted does not grade subjects. An inspector writes a sentence about science when there is something worth saying, that sentence goes into one school’s report, and nobody ever reads it next to the other twenty thousand. SchoolDash has now done exactly that: it processed 40,641 Ofsted reports for secondary schools, kept the 31,748 that carried the most science, and turned twenty-seven years of inspectors’ sentences into something you can count.

The method is worth a line, because it is the reason this could not have been done five years ago. Sentences mentioning science, biology, chemistry or physics were pulled out of every report. One language model read them and built a taxonomy of the ten themes inspectors keep returning to. A second model then read each report in full and decided which of those themes it discussed, and whether what it said was positive, neutral or negative. A random sample of the results was checked by hand.

Ten themes, and the three that never go away

Teaching, staffing and curriculum are mentioned in every era of inspection. Practicals, laboratories, disciplinary literacy and safety were common until the mid-2000s and have largely gone quiet since. Attainment and enrichment had both fallen away, and both have risen again under the inspection framework that took effect in November 2025.

Mentioned in every era

Teaching · Staffing · Curriculum

Perennial, give or take a dip during the pandemic.

Quiet since the mid-2000s

Practicals · Laboratories · Disciplinary literacy · Safety

Common until the mid-2000s. Rarely mentioned now.

A gentler decline

Assessment

Down, but not off a cliff.

Rising again since November 2025

Attainment · Enrichment

Both had fallen. Both tick up under the current framework.

That middle group is a fact about inspectors, not about schools. Practical work did not stop in 2006; inspectors stopped writing about it. It is still worth a department’s attention, because the same word comes back in the criticisms below — this time as something missing from lessons.

What inspectors praise, and what they don’t

Two themes were read in detail: teaching and curriculum. In both, praise outnumbers criticism by a distance — which is itself a finding, since a report that dwells on the positive will flatter every school in the sample equally.

Teaching 10,229 positive · 3,288 negative

76% 24%

Praised

• Specialist subject knowledge, and explanations that make a hard idea clear
• Strong progress and conceptual understanding
• Practical work and investigation that earns its time
• Questioning that finds the misconception rather than the right answer
• Literacy, numeracy and data handling taught inside science

Criticised

• Staffing instability, and cover by non-specialists
• Quality that varies by teacher and by key stage, with low expectations
• Passive lessons: copying, worksheets, teacher talk
• Marking that does not change what happens next
• Too little independent investigation or data analysis

Curriculum 12,034 positive · 2,344 negative

84% 16%

Praised

• Broad pathways at Key Stage 4 and post-16
• Provision that stretches high attainers, including triple science
• Sequencing that builds across key stages, including from primary
• Enrichment: clubs, partnerships, trips, speakers
• Specialist status and funding, where a school had it

Criticised

• Fragmented planning and sequencing, especially where classes rotate between the three sciences
• Too little teaching time, and timetables that split a class across teachers
• Pathways that narrow: triple withheld from pupils who could take it, or pushed on pupils who cannot
• Underdeveloped practical, investigative and data work
• Shortages of laboratories, equipment, specialists (physics especially) and technician time

Read the two criticised columns next to each other and much of it is one problem wearing two hats. A class rotates between the three sciences, so a topic is met, dropped for a term and picked up by a different teacher. The sequence is fragmented at curriculum level, and at lesson level it shows up as low expectations, a worksheet and marking that cannot fix what the pupil has already forgotten. Sequencing is not a planning document. It is what the pupil actually experiences, in order.

The one science that cannot be staffed

The first criticism in that column — staffing instability, and cover by non-specialists — is the one a head of science can do least about, and the national figures say which subject it is. In the Department for Education’s teacher training census for 2025/26, biology recruited half as many trainees again as it needed and chemistry comfortably beat its target. Physics recruited 1,095 against a target of 1,410.

Biology

153% 1,506 of 985

Chemistry

120% 878 of 730

Physics

78% 1,095 of 1,410

That 78 per cent was, in the Department’s own words, the highest number of physics entrants since comparable figures began in 2014/15 — the shortfall is long-standing, not a bad year. In a department it turns into something ordinary: physics taught, at least in part, by colleagues who trained in one of the other two sciences, which is precisely the situation in which what is written down for a unit stops being paperwork and starts being the lesson.

The gap that runs through all of it

Sorting the most recent reports by the schools that received them produces the finding that matters most. Schools serving poorer and lower-attaining communities get consistently less positive comments about science. The gap is largest on attainment, which follows the gap already visible in national results, but it is also there in teaching, staffing and enrichment. Schools in high-deprivation communities are also less likely to have any of these themes mentioned at all.

The same divide shows up in what pupils are entered for. Triple science entry is much higher in the south and in schools serving more affluent communities, while double science predominates in the north and midlands. Double science entry peaks in the middle of the deprivation range rather than at the top of it, and SchoolDash gives the reason plainly: in the highest-deprivation schools, many pupils sit no science GCSE at all.

Combined science

951,510 +1.8%

Biology

169,915 −1.0%

Chemistry

161,400 −1.0%

Physics

161,305 −1.0%

Nationally, then, the separate sciences are the minority route by a factor of about six, and this summer all three of them fell by one per cent while combined science rose. The decision SchoolDash found varying by region and by deprivation is, in most schools, the decision about which of those two lines a pupil joins.

What this is not

One sentence in SchoolDash’s post decides how the rest of it should be read. They point out that schools serving poorer communities may not be performing any worse, but being judged more harshly, and that its data “cannot distinguish between those two possibilities” — adding that either would be a serious cause for concern. This article takes it the same way. What has been counted here is what inspectors wrote about science, which is not the same thing as how science is taught.

Two more limits are worth stating plainly. The archive runs across four inspection frameworks, so a comment from 2003 and a comment from 2026 were written to different rules about what an inspector should look at and say — the themes that fade are partly a record of that, not only of schools. And an inspector writes about science when something prompts it, which means silence in a report is not a finding. The counts here measure attention as much as quality.

The part a department actually controls

Plenty of what inspectors criticise sits above a head of science: a physics vacancy nobody applies for, a laboratory that was built in 1974, a technician post that was cut, minutes on a whole-school timetable. Reading the criticised columns again, though, three things are decided inside the department and written down by it — what a unit assumes has been taught before it, what marking is for, and who is offered which pathway.

None of that is new, which is rather the point: the same handful of criticisms have been written about science departments for twenty-seven years, across four inspection frameworks. They are the reason the science handbook is organised around curriculum, knowledge, enquiry, practical work and consistency rather than around the three subjects — a department that teaches three hierarchies in parallel has more scope for inconsistency than any other in the school.

Each of those decisions has to end up written down somewhere, which is what the rest of the science shelf is for: schemes of work for the sequence itself, knowledge organisers for what a class must still know when a topic comes back a term later, and question books for the applied marks that a pupil loses while knowing the content. None of it recruits a physicist. All of it decides what happens in the room when the physicist is not there.

Twenty-seven years, five complaints

An analysis like this cannot tell a head of science whether their department is good. It can tell them what inspectors have found worth writing down about science for a quarter of a century — and that the list is short enough to work through, one meeting at a time.