The teacher workload survey nobody reads past the headline — and what it actually says

Once a year, the Department for Education publishes its survey of teachers' working lives, a handful of outlets run a single line about average hours, and the conversation moves on within a day. That's a shame, because the actual report is one of the more honest pictures available of what the job costs — and this year's edition has more nuance in it than "teachers are tired" headlines usually capture.

The headline number, and the one underneath it

The topline is the one everyone quotes: average working hours have eased back slightly compared to the year before. Primary classroom teachers are reporting a little over 51 hours a week, secondary teachers a little under 50, and school leaders well above both at over 56. Those are small movements — a matter of an hour or so in each case — but they're movements in the right direction, and that's worth saying plainly rather than burying in caveats.

The number that matters more, though, sits a little further down the page: the proportion of staff who describe their workload as "not acceptable." That figure has dropped from around half to a little under half. Again — real movement, not nothing. But "a little under half" is still a description of roughly two members of staff in every five walking into a building each morning already feeling that the basic shape of the job is wrong. A modest improvement on a difficult baseline is still a difficult baseline.

Where the hours actually go

The part of these surveys that rarely makes it into a headline is the breakdown of where the time goes — and it's rarely where the public conversation assumes. Teaching itself is, relatively speaking, the well-defended part of the week. The hours that pile up around it are the ones that are hardest to plan for and least visible from outside a school: marking and feedback, data entry and tracking, behaviour follow-up that doesn't end when the lesson does, communication with parents, and the kind of admin that exists to prove the teaching happened rather than to make it better.

None of that is a surprise to anyone who's lived it. But it's worth naming precisely, because "workload" as a word is vague enough to mean almost anything — and vague problems are much harder to act on than specific ones. "Too much marking" is a problem you can target. "Workload" is a problem you can only sympathise with.

What moved, and what that suggests

It's worth asking why the numbers moved this year, even slightly. The honest answer is that nobody can point to a single lever — there's no one national policy that explains a one-hour shift in average reported hours. What's more likely is an accumulation of smaller things: marking policies that asked for less, schools experimenting with how meetings are run, some early signs of AI tools taking a slice off planning and resourcing (a trend covered in its own right here), and — in some cases — simply a year without a major new initiative landing on top of an already full one.

That's a useful thing to notice, because it suggests the lever that moves these numbers isn't usually a single dramatic change. It's the absence of new weight, plus a series of small reductions in the existing weight, compounding quietly over a year. Which is a less satisfying story than a single bold reform — but probably a truer one.

Why the headline undersells the detail

A one-line summary like "teacher hours fall slightly" tells you that something moved. It doesn't tell you what moved, why it might have moved, or which part of the job is still the heaviest once you look past the average. Those are the questions worth sitting with for longer than the news cycle gives them — because they're the questions that point towards what's actually worth changing next, in your school and in the system around it.

The full survey is publicly available, runs to many more pages than any summary of it, and is worth ten minutes of anyone's time who wants the real shape of the picture rather than the rounded-off version of it.