What persistent absence tells you that attendance rates miss

A trust can post a respectable overall attendance figure while a growing group of pupils quietly crosses the persistent absence line. The headline percentage is an average, and persistent absence is a count of children, so the two can move in different directions for a full year before anyone notices.

For trust leaders, headteachers and inclusion leads, the useful question is how to read persistent absence as more than a single percentage on a dashboard. That means being clear on what the measure counts, what it hides, and which experience-side signals arrive before it moves. It also means having a realistic plan for this term, whatever capacity you have.

How persistent absence is defined and why it carries weight

As of the most recent DfE guidance and statistics, a pupil is classed as persistently absent when they miss 10% or more of their possible sessions. A session is a morning or an afternoon, and the figure covers all absence, authorised and unauthorised. The threshold and the wording have been revised in the past, so it is worth checking the current definition in the DfE's published statistics before you put a number in a board paper.

The detail that matters most is that persistent absence is a count of pupils, not an average. A school's overall absence rate is the total sessions missed divided by total possible sessions. A handful of pupils with very high absence add relatively little to that total, so the overall rate barely changes. Many pupils sitting just over the 10% line can change the persistent absence figure sharply while adding only a modest amount to the overall rate.

That is why the measure has such standing with trustees and inspectors. The DfE's statutory guidance, Working together to improve school attendance, expects schools and trusts to identify pupils at risk and to track patterns by group, and the Department's published Pupil absence in schools in England statistics report persistent absence alongside overall and unauthorised absence. It is also the figure that tells you how many children are missing a meaningful share of their education, which the average cannot.

Because it is a pupil-level measure, it points towards action in a way that a rate does not. A rate invites a general response. A list of pupils, grouped by year and characteristic, invites a specific one. The next section looks at how easily a healthy rate can sit on top of a persistent absence problem.

Why a healthy overall attendance rate can hide a persistent absence problem

A simple illustration shows the arithmetic. Suppose two academies each have 600 pupils and each reports 94% overall attendance. In the first, absence is spread thinly, and 45 pupils sit below 90%. In the second, a large middle group has drifted to between 88% and 92%, and 110 pupils sit below 90%. The headline figure is identical. The second academy has more than twice as many persistently absent pupils, and a bigger group still hovering near the line.

Averages hide concentration as well as volume. Persistent absence often clusters in particular year groups and in particular cohorts, such as pupils with SEND, those eligible for Pupil Premium or free school meals, pupils with EAL, and children in care. The DfE publishes persistent absence by pupil characteristic each year, and the gaps are usually visible. The exact figures change with each release, so it is sensible to take them from the latest Pupil absence in schools in England publication rather than from memory or an older board paper.

At local level, the same pattern can sit inside a single year group. A Year 9 cohort may look unremarkable overall while the disadvantaged pupils within it carry most of the persistent absence. Leaders who break results down by cohort tend to find that a whole-school figure was describing two or three very different experiences.

One habit is worth examining gently. Many schools review authorised and unauthorised totals each half-term, which is useful for spotting reasons and coding patterns. Those totals say little about how many pupils are drifting towards the threshold. A pupil at 91% is not yet persistently absent, but a few more missed weeks will change that. Looking at the band between roughly 90% and 94% as its own group gives leaders a window to act before the label applies. That raises the question of why attendance data reaches that window so late.

What sits behind persistent absence, and why attendance data comes late

Attendance data is a lagging indicator. It records an outcome, the sessions a pupil has already missed. By the time a pupil appears on a persistent absence list, the pattern has often been running for weeks or terms, and the habits around it have set. Even a well-run weekly review catches the pupil after the drift has started, not before.

The reasons behind that drift are varied. Illness and family circumstances account for a large share, and some absence reflects needs that have not been met, such as a pupil with SEND for whom the school day is not working. Others are about experience inside school: feeling disconnected from peers, feeling unsafe in particular spaces such as the lunch hall or the corridor between lessons, or sensing that staff do not know or understand them. Persistent absence is therefore not only a matter of individual family choices, a misconception that tends to push responses towards sanctions and letters rather than conversations about what school feels like.

This is where belonging comes in as an experience-side signal. The working hypothesis in many schools is that pupils who do not feel they belong are more likely to withdraw, first in engagement and later in attendance. Academic literature on school belonging, including work published in peer-reviewed journals, generally treats belonging as associated with engagement and attendance. The strength of that link varies between studies and settings, and the evidence is stronger for association than for cause. For a trust, the more practical stance is to treat it as a hypothesis worth testing locally: does low belonging in a cohort tend to precede rising absence in your own schools?

That test needs data collected on the experience side, on a regular rhythm, with questions that work the same way in every academy. The next section looks at what that data looks like and how it sits alongside attendance figures.

How belonging data gives an earlier signal than attendance figures

A termly belonging survey asks pupils, staff and families a short set of structured questions. For pupils, these are agree or disagree statements and rating scales, such as whether they feel they have somewhere to go at lunch, whether they feel known by at least one adult, or whether they feel safe in different parts of the school. There are no free-text boxes. Staff and parent surveys can include open comments, but pupil voice is captured as scores and patterns, which keeps the survey short and the results easy to compare.

Identical questions and identical timing matter more than they first appear. If every academy asks the same questions at the same point in the term, a trust board can compare results directly. Headteacher summaries, however well intentioned, vary in what they emphasise and how they are framed. A common instrument replaces that variation with data collected straight from the people involved.

The difference from attendance data is timing. Attendance records what has happened. Belonging responses describe how pupils feel now, so a drop in a cohort's scores can appear while their attendance still looks acceptable. That is the sense in which belonging data is a leading indicator and attendance a lagging one.

The strongest reading comes from putting the two side by side. A cohort with low belonging scores and attendance beginning to slip is a stronger case for action than either signal alone. A cohort with low belonging and steady attendance is one to watch. A cohort with falling attendance but healthy belonging suggests that the cause may lie elsewhere, such as health or family circumstances, and calls for a different conversation.

One caution applies throughout. Belonging scores indicate risk, not cause. A low score tells you where to look and whom to ask, and it does not prove why a group is withdrawing. Used that way, as a prompt for inquiry rather than a verdict, it fits comfortably alongside the data a trust already holds. The next section shows how that might work across a trust.

A worked example: a trust comparing Year 7 across four academies

The following is a hypothetical illustration. Imagine a four-academy trust in which every academy reports overall attendance within a point of the others, around 94%. On the headline figures, nothing separates them, and the trust board has no particular reason to ask questions.

In the autumn term the trust runs the same belonging survey in all four academies. Looking at Year 7, three academies show similar scores. In the fourth, the Pupil Premium cohort scores markedly lower than its peers on questions about feeling known by staff and having someone to turn to. The same academy also has more Year 7 pupils sitting between 90% and 94% attendance than the others, though few have yet crossed into persistent absence.

Reading this, the trust leader works through a sequence. First, segment by cohort, so that the Pupil Premium group is visible as its own picture instead of being averaged away. Second, compare that cohort against the trust picture, to see whether the gap is local or common. Third, look at where scores and attendance diverge: here, both point the same way, which strengthens the case. Fourth, agree a named owner and an intervention. The academy might, for example, ask its head of year to lead a mentoring and transition-follow-up plan for that cohort, with a clear owner and a date for review.

What the board sees is a comparable view across the four academies, built from the same questions and timing, in place of four narrative summaries. The purpose is planning. It shows the trust where to direct support and attention, and it gives the fourth academy's leaders early sight of a pattern they can act on, which is a better position than discovering it in next year's persistent absence figures.

Next term the trust re-surveys. If belonging scores for that cohort have moved, and the number of pupils in the 90% to 94% band has fallen, the intervention has some evidence behind it. If not, the trust has learned something early enough to adjust. That cycle can be built into routine practice, as the next section sets out.

What a trust or school can put in place this term

The approach that works best is a repeating cycle, not a one-off survey. Many trusts structure it in five stages:

1. Measure belonging with the same short survey, at the same point each term.
2. Segment results by year group and cohort, including SEND, Pupil Premium and EAL.
3. Connect the results to attendance and persistent absence data for the same cohorts.
4. Respond with a named owner and a specific action for each low-scoring cohort.
5. Assure by reporting to trustees alongside attendance data, then repeat the following term.

A few practical details tend to make the difference. Running the first survey at a consistent point in the term, such as the same fortnight each time, keeps results comparable from one term to the next. Pupil surveys that take five to ten minutes and use structured questions only get better completion, particularly from the groups leaders most need to hear from. Sharing results with staff soon afterwards, with a named owner against each action, keeps the data tied to decisions. Trustees usually find it easier to engage when belonging results appear on the same page as attendance and persistent absence, not in a separate pastoral report.

Where capacity is tight, it is reasonable to start small. Choosing one year group and two cohorts, for instance Year 7 and its Pupil Premium and SEND groups, gives a manageable baseline. Setting that baseline now, and then reviewing it next term against the same cohorts' attendance and persistent absence numbers, produces a first real comparison without asking staff to take on a major new workload.

How Satchel Pulse can help

Perceptions, from Satchel Pulse, runs termly belonging surveys for pupils, staff and families, using a validated question set that takes five to ten minutes. Results can be broken down by cohort through the Cohort Explorer, which builds year group, SEND, Pupil Premium and EAL groups from the MIS, and a trust-wide dashboard puts every academy on one comparable view. The platform is UK hosted and compliant with UK GDPR.

Satchel Pulse's own analysis indicates that belonging data can flag cohorts at risk of rising absence up to 18 months before it shows in attendance or behaviour figures. That is the company's finding, based on its own survey and attendance data, and it is best treated as a reason to test the link in your own schools instead of a guarantee.

One baseline survey, set against this term's persistent absence list

The most useful step this term is a small one. Run a baseline belonging survey in a single year group, line the results up against your current persistent absence data for the same pupils, and look at where the two agree and where they do not. Then agree one cohort and one named owner, and put the review date in the diary for next term.

That gives you a first answer to the question this article started with: whether belonging in your own schools moves ahead of attendance. If you would like to see how Perceptions could support that baseline, Learn more about our services.