The things schools ask most often about attendance, answered properly. These are the DfE definitions with the official sources linked, plus the boundary cases that are easy to get wrong by hand. Nothing here is held back for a sales call.
What counts as persistent absence?90%
A pupil is persistently absent if they miss 10% or more of their possible sessions, which is attendance at or below 90%. A session is half a day, so a school year of roughly 380 sessions puts the line at about 38 missed sessions, or 19 days. It is counted per enrolment rather than per child, so a pupil who moves school partway through the year is counted for the sessions they were on each roll. The trap worth knowing is the boundary: exactly 90.0% is persistently absent, not just outside it, and a system that rounds a percentage before comparing will quietly miss those pupils.
What counts as severe absence?50%
Missing 50% or more of possible sessions, so attendance at or below 50%. Severely absent pupils are also persistently absent by definition, so they appear in both counts rather than being separated out. The number is usually small and it is the one that matters most: these are the children missing half their education, and they generally need a different response from a pupil drifting at 88%.
How is school attendance percentage calculated?sessions
Present sessions divided by possible sessions, where a session is half a day and approved off-site educational activity counts as present. Overall absence is simply 100 minus that. The part that catches people out is the whole-school figure: it is session weighted, meaning you add up every session across every enrolment and divide once. It is not the average of your year group percentages. Those two produce different numbers whenever year groups differ in size, and only the first matches what the DfE publishes.
When does a school have to consider a penalty notice?10 in 10
The National Framework threshold is 10 sessions of unauthorised absence within a rolling 10 school week period. Ten sessions is five school days, and they do not have to be consecutive. Two points are easy to get wrong by hand. The window rolls, so it is not a term or a half term, and a pupil can cross the line, drop back under it and cross again. And weeks with no sessions, such as holidays, should not dilute the window. Tracking this manually across a whole school is where most of the error comes from, because it means re-counting every pupil every week.
What is the difference between authorised and unauthorised absence?Both
Authorised absence is absence the school has agreed to, such as illness or a medical appointment. Unauthorised absence is absence the school has not agreed to, including a holiday taken in term time without leave. Both count identically towards attendance and towards persistent absence, which surprises families: authorising an absence does not make it disappear from a child's attendance figure. Only unauthorised absence counts towards the penalty notice threshold.
Does authorising an absence improve a pupil's attendance figure?No
No. Authorised and unauthorised absence are both absence, and both reduce attendance by exactly the same amount. The distinction changes what enforcement is available, not what the percentage says. This is worth being clear about with families, because a parent who has been told an absence was authorised often believes it has been discounted.
How often should a school monitor attendance?Weekly
Weekly is the pattern that actually changes outcomes, because the point of monitoring is the conversation you have next week rather than the figure you report next term. A pupil drifting towards 90% is reachable; a pupil who arrived at 84% over a term usually is not, and by then the arithmetic is against you. Recovering from 88% to above 90% part way through a year can require near-perfect attendance for the rest of it, which is why noticing early matters more than measuring precisely.
What does Ofsted look at for attendance?vs national
Ofsted's framework asks whether attendance is broadly in line with national averages, or improving towards them. That phrasing has two consequences. You need your figure next to the published national figure and, for a fairer comparison, your local authority's. And you need direction over time, not a single snapshot, because a school below national but improving is answering the question differently from a school below national and flat. Inspectors also ask what you do about persistent absence specifically, so being able to name the pupils, and say what has been tried, is worth more than the percentage.
What attendance information do governors need?No names
Enough to hold the school to account, and no pupil named. Governors need the whole-school figure with its direction of travel, the persistent and severe absence counts, comparison against national and local authority figures, and any group where attendance is materially behind the school's own average. They do not need, and should not be given, a list of children. Naming a pupil to a governing body is a disclosure with no purpose behind it, since the decisions governors take are about strategy and resource rather than about individuals.
How do you tell a real attendance pattern from noise?Whole school
By checking whether it holds across the whole school rather than for one class. A single weekday standing clear of the others across every year group is almost never about families: it is usually timetable, transport, or something that happens on that day. Mornings against afternoons works the same way. A whole-school skew towards afternoon absence points at lunchtime or at the last lesson rather than at getting children out of the house. Both are worth checking before starting conversations with parents, because the fix may not involve parents at all.
Can a multi-academy trust compare attendance across its schools?Yes
Yes, and the thing to be careful about is comparing like with like. Whole-school percentages across schools of different sizes and phases tell you less than persistent absence rates do, and a primary and a secondary in the same trust are not really comparable on a raw figure at all. Comparing each school against its own local authority average, rather than against its sibling schools, usually gives a trust board a fairer picture of which school is actually underperforming its context.
What data do you need to analyse attendance properly?3 exports
Less than people expect. Session-level marks with their dates and DfE codes, and enough pupil identity to group and name them: first name, last name, unique pupil number, year group and registration group. Every MIS in common use, including SIMS, Arbor and Bromcom, exports this. You do not need dates of birth, addresses, contact details or any free text, and a system that asks for them is collecting more than the job requires.
Where these definitions come from
Government sources, reviewed on 23 August 2026. Attendance policy moves, so the date is here for you to judge whether to check the live pages rather than take our word for it.
For use from 7 September 2026. Attendance and behaviour is one evaluation area; the grade descriptors are from page 41.
Doing all of this from a spreadsheet?
Attendance Insight works the answers above out of the exports your MIS already produces: who is drifting below your target, who crosses the penalty notice threshold and in which ten weeks, which patterns are real, and the reports for governors and inspectors already written. Made by iCT4 Limited with attendance leads.