How to evidence SEN support properly in a UK school
Most schools do more SEN support than their paperwork shows.
That creates a problem later. When evidence is patchy, it is harder to show what has been tried, what has worked, and what should happen next.
The aim is not more forms. It is one clear record that can be reused.
Start with the graduated approach
In England, SEN support is usually discussed through the assess, plan, do, review cycle.
That sounds neat, but the value comes from what is actually recorded at each stage.
What good evidence looks like
Good SEN evidence usually has four features:
1. It is specific 2. It is dated 3. It shows action, not just concern 4. It links support to some kind of outcome
A note that says “concerns about reading” is not enough on its own.
A note that says “15 minutes of guided reading three times a week started on 8 January, with weekly fluency checks, and the pupil moved from 58 to 66 words per minute by 12 February” is much more useful.
What schools should capture
You do not need a giant system to do this well. You need consistency.
1. The concern
Record what the issue is, in plain language.
Examples:
- reading fluency remains slow
- transitions between tasks are difficult
- pupil struggles to retain multi-step instructions
- anxiety rises before tests
2. The support plan
Record what the school is going to do.
Examples:
- small-group phonics booster
- visual timetable in class
- daily check-in with trusted adult
- shorter instructions with visual prompts
3. The delivery
Record whether the support actually happened.
This is the bit many schools miss.
If the intervention was planned but never delivered regularly, the evidence is weak. If staffing changed halfway through, note that too.
4. The impact
Record what changed.
That does not always mean a huge jump in attainment. Sometimes the evidence is:
- fewer incidents of shutdown
- better attendance at intervention
- more confidence in class
- improved independence
- small but steady academic progress
What to avoid
There are a few common traps.
Vague language
Avoid notes like:
- “support given”
- “worked well”
- “needs more help”
These phrases do not tell anyone what actually happened.
Copy and paste records
If every child sounds the same, the evidence will not help much.
A school can use templates, but the detail still needs to be real.
Hidden information
If the only copy of a note sits in one person’s inbox, the system is brittle.
Duplicate entry
This is a huge one.
Schools often write the same thing in the intervention sheet, the pupil file, the meeting summary, the email chain, and the annual review pack. That creates extra work and more chances for errors.
A simple structure that works
For each SEN support cycle, try this structure.
Assess
- what is the concern?
- what evidence do we have?
- what do the adults and the pupil say?
Plan
- what support are we putting in place?
- who owns it?
- how often will it happen?
- what outcome are we looking for?
Do
- was the support delivered?
- were there any changes?
- what else happened in school during that period?
Review
- what changed?
- what should continue?
- what should stop?
- what needs to be escalated?
Good evidence is helpful beyond compliance
This is not just about being inspection-ready.
Good SEN evidence also helps schools:
- explain decisions to parents
- hand over between staff
- spot patterns earlier
- avoid repeating work
- build a more accurate picture of what support costs in time and effort
The real standard
The question is not whether the paperwork looks impressive.
The question is whether another member of staff could pick up the record and understand, quickly and accurately, what support the child has had and whether it helped.
If the answer is yes, the evidence is probably doing its job.
In practice
A system like MeritDocs can help schools capture support once and reuse it across reviews, parent meetings, and reporting.
Final thought
Good SEN support records are not about perfect prose. They are about clarity, continuity, and making sure the next decision is better than the last one.