How do we deploy teaching assistants effectively for SEND?

It is the last week of the summer term and you are looking at the high-needs budget next to the September staffing plan. Most of the top-up money has, almost by reflex, turned into hours of one-to-one adult support, and most of those hours sit with teaching assistants attached to individual children. Money is tight, an Area SEND inspection could land any year, and the evidence you keep being sent says that this exact model can hold back the very children it is meant to help. So the real question is not how many hours you can fund. It is what good looks like instead.

The finding that should change how you staff

The largest UK study of teaching assistants found that the more support a pupil got from a TA, the less academic progress they tended to make. That is the uncomfortable headline most "how to deploy your TAs" articles skip past, and it is the right place to start.

The study was the Deployment and Impact of Support Staff project, run out of the Institute of Education between 2004 and 2009. It tracked around 8,200 pupils. Of 21 measured results, 16 pointed in a negative direction: heavier TA support, slower progress (Blatchford, Russell, Webster et al., DISS, 2009). For the children with the highest level of need, the gap was widest.

Read the next sentence carefully, because the whole article turns on it. This was not the fault of the TAs. The researchers were clear that the problem sat with how schools deployed and prepared them, decisions made above the TA's pay grade and outside their control. TAs were routinely handed the lowest-attaining groups, given little or no time with the teacher to plan, and left to cover subject knowledge they had never been trained in. The model failed. The people did not.

The specific pattern that did the damage is the one you will recognise instantly: the "Velcro TA". A single adult fixed to a single child all day. What it produced was a quiet inversion of who teaches whom. The pupils with the greatest need spent the least time with the qualified teacher and the most time with the least-qualified adult in the room. And it bred what the team called learned dependence: children who waited for the adult before attempting anything, because an adult was always there to wait for.

What the 2025 EEF guidance actually recommends

The current guidance is the Education Endowment Foundation's "Deployment of Teaching Assistants", published on 26 March 2025, with five recommendations. If your TA policy still cites "Making Best Use of Teaching Assistants" and its seven points, it is out of date.

That 2015 report has been superseded. The new five split cleanly into two halves: what effective practice looks like, and how you actually make it happen (EEF, Deployment of Teaching Assistants, March 2025).

1. Deploy TAs so that every pupil can access high-quality teaching from the teacher.
2. Deploy TAs to scaffold learning and build pupil independence, not to do the work for the child.
3. Deploy TAs to deliver evidence-based, structured interventions.
4. Prepare and train your staff around how TAs are deployed.
5. Get every member of staff behind making it happen.

Recommendations one to three are the practice. Four and five are the implementation, and they are the half schools usually drop. The single principle that matters most for SEND sits underneath all five: a TA should add to and extend the teacher's work, never replace it. The EEF puts it plainly: the pupils who struggle most should spend at least as much time with the teacher as their peers, if not more.

The numbers back the model. In the EEF's Teaching and Learning Toolkit, structured TA-led interventions are rated at around +4 months' additional progress: moderate impact, moderate cost, moderate evidence, with an effect size of roughly 0.35 (EEF Toolkit, Teaching Assistant Interventions, 2025). The same Toolkit is just as clear about the other side: general, undirected, in-class TA support typically shows no positive benefit. The adult hours are not the lever. What the adult does with them is.

Scaffolding, not rescuing: the least-support-first ladder

The difference between a TA who builds independence and one who quietly erodes it comes down to a single habit: giving the least amount of help first, and letting the child have a real go before stepping in.

The EEF sets this out as a five-level scaffolding ladder for how a TA talks to a pupil, running from most adult input to least (EEF scaffolding framework, 2025):

• Correcting: giving the answer. The most support, and the level to leave fastest.
• Modelling: showing how, then asking the pupil to copy the approach.
• Clueing: offering a hint that narrows things down.
• Prompting: a nudge towards a strategy the pupil already knows.
• Self-scaffolding: the pupil works it through alone. The top of the ladder, and the goal.

The operating rule is simple. Start low on the ladder. Wait. Only climb if the pupil is genuinely stuck, and step back down as soon as they get traction. A TA who opens every interaction at "correcting" is not helping a child learn. They are doing the learning instead.

Here is the nuance that stops this becoming "pull all the support away". For pupils with SEND, a scaffold may need to stay in place longer, sometimes a good deal longer, to build competence and confidence that can survive once it is taken off. The target is graduated withdrawal, not a cliff edge. The technical word is fading: planning the slow removal of help so it disappears as ability grows, rather than vanishing the moment a budget gets tight.

What fading looks like in practice

The fade is built from ordinary tools, used deliberately. Planners, timers and checklists the pupil learns to reach for themselves. Reflective questions after a task: what worked, what would you try differently next time. Coaching a child towards positive self-talk so the inner voice starts to do the job the adult used to. None of this is expensive. All of it grows the metacognition and self-regulation that 1:1 hovering suppresses.

Where TAs add the most value: structured interventions

If a TA is going to move the dial on a SEND pupil's progress, the evidence points to one thing: short, regular, well-structured interventions delivered by someone trained to deliver them.

Hold the two findings side by side. Everyday in-class support, the TA drifting between tables, shows little measurable benefit. Brief, structured one-to-one or small-group sessions are where that +4 months actually comes from. Same staff member, completely different outcome, because the work is designed rather than improvised.

"Structured" is not a vibe. In practice it means a specific set of things:

• An evidence-based programme, not a worksheet someone printed off.
• Explicit training for the TA in that programme, before they run it.
• Sessions of consistent length and frequency, typically short and often: think three to five times a week across a set block of weeks.
• Fidelity to the programme, so it is delivered as designed.
• A feedback loop with the teacher, who keeps responsibility for the pupil's learning.

The last point on connection is the one schools lose. An intervention run as a silo in a corridor, disconnected from what the child is doing in class, will not transfer. The gains have to be wired back into the curriculum the pupil sits in for the rest of the week, or they evaporate.

There is a budget argument here too, and it cuts against instinct. Spending on training a TA to run a structured programme well is, more often than not, better value than buying another tranche of undirected 1:1 hours. You are paying for impact, not for an adult standing nearby.

Decoupling EHCP funding from 'hours of an adult'

The trap most schools fall into is reading an EHCP's top-up funding and its support hours as an instruction to glue one named TA to one child. The plan rarely requires that, and the model is exactly the one the DISS data warns against.

Section F of an EHCP, the part that sets out the special educational provision, has to be specific and the council has a legal duty to secure it (Children and Families Act 2014, s.42). But "secure the provision" means deliver the strategies and outcomes named in the plan. It does not, by itself, mean a fixed quantity of adult proximity. For how that duty bites, see does a school have to follow Section F of an EHCP and what is Section F and why it matters.

So reframe how you describe provision. Write it as outcomes and strategies: access arrangements, named structured interventions, scaffolds, environmental adjustments, and protected teacher-led time. Not as a headcount of hours someone hovers. That is a more honest reading of most plans, and a more effective one.

On the ground that means rostering TAs flexibly across a class to keep quality-first teaching moving, which frees the teacher to work directly with the highest-need children, rather than ring-fencing one adult to one pupil from nine until three. The graduated approach in the SEND Code of Practice, assess, plan, do, review, is built for exactly this kind of responsive deployment (SEND Code of Practice 2015, 6.36 to 6.39). See the graduated approach explained.

This is also your answer when someone says "but the plan says 25 hours of TA". Your duties under the Equality Act 2010, the reasonable adjustments owed to disabled pupils (Equality Act 2010, s.20), are met by effective provision and outcomes. They are not met by adult hours for their own sake. A child who is learning to work independently with a well-designed scaffold is better served than one shadowed into dependence, whatever the hour count on paper. For pupils without a plan at all, the same logic runs through what a school must do for a child with SEN but no EHCP.

Leading it: training, timetabling, and the funding clock

None of this happens by writing a better policy. EEF recommendations four and five, the implementation half, put it on senior leaders to own the deployment decisions, build in TA preparation time, and fund ongoing training. A TA cannot scaffold towards independence if no one has ever shown them how.

Name the failure modes out loud in your school, because they are comfortable and they are common:

• No protected liaison time between teacher and TA to plan or hand over.
• TAs given the lowest-attaining group by default, every lesson.
• TAs left to plug subject knowledge gaps they were never trained for.
• The TA functioning as the de facto primary teacher for the SEND pupils, while the qualified teacher works with everyone else.

Now the reason this is a this-year problem and not a someday one. The money is on a clock. The Dedicated Schools Grant statutory override, which lets councils carry high-needs deficits off their main balance sheet, runs only to 1 April 2028. The government has set out plans to absorb roughly 90% of historic high-needs deficits, around £6bn, but the write-off is conditional: every local authority has to submit an approved local SEND reform plan to get it. From 2028 to 2029, high-needs spending moves inside the main departmental budget (DfE, DSG deficits explanatory note, 2026).

And inspection is watching the same thing. The Area SEND inspection framework, run jointly by Ofsted and the Care Quality Commission since January 2023 and updated in June 2025, does not judge individual TAs. It judges how effectively a setting deploys them (Ofsted and CQC, Area SEND framework, 2023, updated 2025). A school still running the Velcro model in 2026 is not just spending badly. It is carrying an inspection risk.

Where the policy is heading

The direction of travel reinforces all of this. The 2026 Schools White Paper, "Every child achieving and thriving", published on 23 February 2026, and the Education for All Bill propose Individual Support Plans for SEND pupils and a narrowing of EHCPs to the most complex needs by 2035 (House of Commons Library, 2026). Nothing changes before the late 2020s, so do not rebuild your provision around a Bill that has not passed. But the destination is a more inclusive, mainstream model resting on strong quality-first teaching. Effective TA deployment is not a detour from that. It is a head start on it.

What to do this term

One audit, before you lock the September timetable. Pull your SEND pupils with the highest top-up funding and answer one question for each: across a typical week, do they spend more time with the qualified teacher, or less, than their peers? If the answer is less, that is the thing to change first.

1. This fortnight: map who your TAs are attached to and for how long. Flag every pupil on a near-full-time 1:1 with one named adult. That is your highest-risk, highest-cost group.
2. Before September: swap your TA policy's reference to the 2015 report for the 2025 "Deployment of Teaching Assistants" guidance, and build teacher-TA liaison time into the timetable. Even ten protected minutes before a lesson changes what a TA can do in it.
3. Across the autumn term: pick one evidence-based structured intervention, train two or three TAs to deliver it properly, and run it to a fixed schedule with a feedback loop to the class teacher. Measure it.
4. At the next annual review: rewrite provision as outcomes and strategies, not adult hours. Evidence your graduated- approach work as you go, which also strengthens any future plan request (see how to evidence the impact of SEN support).

For the funding side of the same decisions, including how top-up money reaches your school in the first place, see how high needs funding works for schools.

The honest version: getting this right will not save you money in week one, and it will be more work for your teachers before it is less. But a school that spends its TA budget on building independence rather than renting proximity is the school that comes out of the next five years in better shape, on the budget and in the inspection.

Where the evidence and law come from

• EEF, "Deployment of Teaching Assistants" guidance report and scaffolding framework, 26 March 2025 (the five recommendations and the five-level ladder).
• EEF Teaching and Learning Toolkit, "Teaching Assistant Interventions", 2025 (+4 months; effect size ~0.35).
• DISS project, Blatchford, Russell, Webster et al., Institute of Education, 2009 (8,200 pupils; deployment, not TAs, drove the negative association). Led to the MITA programme.
• Children and Families Act 2014, Part 3, s.42 (duty to secure Section F provision); Equality Act 2010, s.20 (reasonable adjustments); SEND Code of Practice 2015, paras 6.36 to 6.39 (graduated approach).
• DfE DSG deficits explanatory note, 2026 (override to 1 April 2028; ~£6bn write-off conditional on local SEND reform plans).
• Ofsted and CQC Area SEND inspection framework, January 2023, updated June 2025. House of Commons Library briefing on the 2026 Schools White Paper and Education for All Bill.

This article is general information for school leaders, not legal advice on an individual pupil's plan. It has been reviewed against the SEND Code of Practice 2015 and current statutory guidance, but does not replace advice from your local authority or a solicitor on a specific case.