Build an attendance cohort view by fixing the question, population and period before applying filters. Show counts, coverage and missing data beside every rate. Use one characteristic at a time, then return to authorised pupil-level evidence before deciding what support or review follows.
A cohort can reveal where to look. It cannot explain why a pupil is absent, diagnose a need, assign blame or choose an automatic referral, sanction, penalty notice or legal route.
This guide is for attendance leaders, school leaders, pastoral teams, DSLs, family-support staff, data leads and SENCos in England. It is current to 12 August 2026.
Start with one question the school can act on
The July 2026 edition of the Department for Education's statutory attendance guidance asks schools to analyse weekly patterns at pupil, cohort and year-group level, then use deeper half-termly, termly and full-year review.
The question should name a period, population and possible school response. “Which enrolled Year 9 pupils show a recurring afternoon pattern that needs a conversation about access?” is usable. “Why is Year 9 attendance poor?” asks the chart to write a biography.
Decide who owns the analysis and when it will be reviewed. Agree who can see pupil-level detail. A colourful dashboard without ownership is still an unattended spreadsheet, only with better lighting.
Ask five questions in order
- Who is included? Define enrolled pupils, leavers, compulsory-school-age scope, admission dates and any planned exclusions from possible attendance sessions.
- Which period and measure? Use comparable dates and state whether the view shows attendance, absence, unauthorised absence, persistent absence or another measure.
- How complete is the source? Show cohort size, possible sessions, missing marks, unknown characteristics, late feeds and changes to pupil status.
- Which single lens helps? Start with year, class, SEND, free school meals or another relevant characteristic. Add intersections only when the group remains safe and interpretable.
- What pupil-level review follows? Open the underlying sessions and relevant context, hear pupil and family views, then record an action, owner and review question.
Keep that sequence on the analysis card. It makes the denominator and the human follow-up part of the result, instead of notes added after a meeting has already adopted the headline.
Freeze the denominator before comparing groups
Write the inclusion rule beside the result. The DfE Monitor your school attendance tool can include leavers and pupils outside compulsory school age until filters are applied. Entry-year comparisons can also lack data from a previous school.
Its filters use the most recent pupil characteristics from the MIS. A pupil who received SEN support last year but does not now will be grouped by current status. That is a valid view for some questions and a misleading one for others.
Compare like with like: the same dates, possible-session rules, pupil scope and measure. The DfE notes that school reports, daily-data tools and published statistics may differ because their coverage, timing and calculations differ.
Keep rapid, revisable daily-MIS evidence separate from accredited census-based statistics. Neither becomes interchangeable because both arrived in a spreadsheet.
Layer characteristics slowly and keep missingness visible
The statutory guidance says schools choose relevant cohorts from their context. It names groups such as year, SEND, free school meals, young carers and pupils with a social worker. This is a prompt for proportionate analysis, not a request to combine every field.
Begin with one lens. If the pattern persists, test a second characteristic because it could change the operational question. Stop when the group becomes too small, unstable or identifiable, or when missing fields make the comparison unreliable.
Counts matter beside percentages. So do unknowns. Record pupils excluded by the period, missing sessions, late joiners, leavers and characteristics that are unavailable or current only. A blank cell is evidence about the analysis, even when it is not evidence about a child.
Detailed cohort views remain personal data when staff can identify pupils directly or by combining fields. Restrict access, follow the school's data-protection procedures and involve the DPO when the purpose, audience or sharing changes.
Walk a made-up cohort back to the children
This synthetic example combines invented circumstances to explain the method. It is not a school result, customer story or claim about any real pupil.
An attendance lead sees a recurring afternoon pattern among enrolled Year 9 pupils. A SEND lens shows that several pupils receiving SEN support sit within it. The pattern is a reason to inspect the evidence, not an explanation.
The lead checks possible sessions and finds that the selected period includes late joiners. Current-status filtering also hides changes in SEN support. Some assessment records are incomplete, so no attainment comparison is made for those pupils.
Authorised pupil-level review shows different stories. One pupil missed teaching during a phased return. Another had approved off-site education. A third describes a difficult transition after lunch, while provision records show that agreed support was not consistently available.
The cohort finding therefore becomes several human questions. Staff check coding and access, hear each pupil and family, review adjustments or support, and use the DSL route immediately if separate safeguarding evidence creates an urgent concern.
The aggregate did not prove a common cause. It helped the team find where a careful review was needed and prevented one group label from becoming a substitute for individual context.
Turn the finding into a review question
End the analysis with one sentence: “We will check whether the identified in-school barrier is present, whether agreed support is accessible, and what pupils and families say; the attendance lead will review this on the agreed date.”
Name what would change the decision. Corrected codes, fuller source coverage, pupil or family evidence, delivered adjustments, curriculum access, health advice or a different time pattern may all alter the interpretation.
Review all absence bands, not only pupils who have crossed a persistent or severe absence threshold. The DfE absence-bandings guide also directs leaders to consider underlying reasons and then inspect individual pupil data.
What the cohort view cannot prove
A group difference cannot prove motive, parenting, illness, unmet SEND, exploitation, safeguarding harm or the effect of school provision. Correlation across attendance, behaviour and attainment cannot establish which factor caused another.
A threshold does not decide support or enforcement. Statutory guidance places monitoring, listening, understanding and support before formal or legal routes, with decisions based on the individual circumstances and local procedures.
Urgent safeguarding concerns still go to the DSL and established local or emergency route. Attendance analysis supports that professional system; it does not replace it.
Where Student Radar fits
Student Radar's available Attendance Analytics can review headline, period, class, session and cohort patterns. The available Heatmap can help authorised staff move from a group signal to the pupil context behind it.
Where the Assess package is in use, Assessment Reports can add school, year, class and pupil context with small-group caveats. Staff still validate the sources and decide what the evidence means.
Student Radar does not diagnose a barrier, predict future absence, guarantee an outcome or choose a referral, sanction, penalty notice or legal route. The source MIS remains the attendance record, and professional review remains with the school.
A ten-minute setup for September
- Write one attendance question and the school response it could inform.
- Record the population, period, measure and possible-session rule.
- Choose one characteristic lens and list the missing fields beside it.
- Set a minimum safe group rule with the data lead or DPO.
- Name who may open pupil-level evidence and who will hear pupil or family views.
- Give the finding an owner, review date and a question that can be revised.
For the wider evidence route, read what connected pupil data means in practice. For register accuracy, use the pre-term attendance-code check.
To review how this workflow could work with your authorised school data, request an attendance-focused walkthrough.
Sources and further reading
- Working together to improve school attendance, Department for Education, published 6 May 2022; updated 9 July 2026.
- Monitor your school attendance: user guide, Department for Education, published 27 March 2024; updated 22 July 2026.
- Monitor your school attendance: how to use the absence bandings report, Department for Education, updated 22 July 2026.
- Pupil attendance in schools: methodology, Department for Education, published and updated 25 September 2025.
- RISE attendance improvement, Department for Education, updated 9 June 2026.
- Data protection in schools, Department for Education, published 3 February 2023; updated 9 July 2026.
- Introduction to anonymisation, Information Commissioner's Office, live guidance accessed 12 August 2026. The ICO says the guidance is under review following the Data (Use and Access) Act.