FSM Pupils Excluded 5x More: The Behaviour Evidence Gap

This is the education commute. Both voices on this show are synthesized. The judgment isn't. So today our mission is to look at the Education Endowment Foundation. Specifically, we are exploring their updated May 2025 toolkit strand on behavior interventions, which, you know, looks across 83 different studies to figure out what actually works. But before we get to the data, I really want to set the scene. Right. Yeah, pull it right down to the ground level. Exactly. I want you to picture this. It is 12 .41 on the playground.

You're out there. You've got 40 meters of tarmac in front of you. You're wearing a high V's vest and you are watching a football argument that started as a completely harmless joke about someone's cousin. Well, turn entirely separate. Oh, we all know that exact moment. You can see the weather changing, can't you? It hasn't turned into a physical altercation yet, but. Two boys who were laughing 30 seconds ago laughing with each other are suddenly, well, they're not laughing anymore. No, they're really not. They're laughing at each other.

The posture shifts, you know, the volume drops, but the intensity just spikes. Spot on. Now, for today, I'm claiming the pastoral leads chair. So I'm the one who chooses the intervention, holds the budget and has to find the staffing for it. But you get the midday supervisor's chair, you know both of these boys' names, you know both their moms, and you know exactly how this 12 .41 incident ends. Well, yeah. It ends at 10 past 1 in someone else's classroom. Because when that bell rings, the argument doesn't just evaporate.

It literally walks inside. And when you radio it in, the look you get from the colleague in the corridor says what it always says, which is, uh... Lunchtime problem, right? Okay. Let's unpack that because there is no such thing as a lunchtime problem is there absolutely not I mean there was one child an unresolved conflict and a clock ticking down Yeah, period five is already burning and the teacher hasn't even opened the door yet. That is the absolute reality of it Yeah, the lunchtime problem is just you know, it's a classroom problem in waiting Yeah, and what actually works for that child not the glossy poster version of a school behavior policy But the measured evaluated version it actually has a hard number attached to it There are those 83 studies behind it, and there is one core finding that, well, it completely flips what most schools would instinctively do first to try and solve this.

Okay, so what does this all mean for us today? Just to clarify our terms for everyone listening when we say behavior interventions based on the toolkit We mean anything that aims to raise attainment by reducing challenging behavior, right? So that could be a whole school ethos universal classroom programs or targeted programs for specific pupils But it's important to differentiate this from the broader guidance because I know what you're thinking. Didn't we cover behavior a fortnight ago, the first 20 days, the doorway, the ratio. But here's the critical difference.

That was the guidance report. That was. The manual, really, the step -by -step how -to. This toolkit strand, this is the receipt. The receipt, I really like that distinction. Yeah, it's the how much. What the entire mountain of 83 studies says the work is actually worth when you measure it in hard months of academic progress. It's the exact same evidence family, so they should absolutely agree, but today we're looking at the exact cost and the exact payout before we even think about how to implement it on the ground.

Well, what's fascinating here is that the headline number, you know, the payout is very clear. On average, behavior interventions yield plus three months of additional progress over a year. Plus three months, okay. Plus three. But, as always, that number comes with two honesty labels stapled to it. First, it is an average across 83 very different studies. Some schools found much more progress while some found well, absolutely nothing at all, right? The context matter exactly and second the evidence security is labeled as moderate. Hold on Let's just make sure that's crystal clear for everyone when we talk about evidence security in the toolkit We are usually talking about the padlocks aren't we we are yeah The toolkit uses a padlock system to rate how secure and reliable the evidence is We are looking at three out of five padlocks here.

Okay three out of five Yeah, so that means it is a moderate rating It is a solid, reliable average based on a good chunk of data, but it is not an absolute proven law of physics that will replicate perfectly in every single context. Here's where it gets really interesting, though. If we look at the consistency of that plus three months, it holds surprisingly steady across different contexts, doesn't it? It really does, which is actually quite unusual. It is remarkably similar for both primary and secondary phases, plus three in both.

Wow, really? Yeah, and it's similar across subjects, too. Maths and literacy are moved by about the exact same amount. Which makes it rarer than it sounds because most strands wobble somewhere depending on the age group or the subject. But I have to push back a little on this plus three months figure. Oh, go on then. Because three months for behavior work still feels, I don't know, it feels a bit underpriced to me. It feels low. I mean, a child who isn't in the room because of a suspension learns nothing at all.

That's zero months of progress. Well, if we connect this to the bigger picture. That pushback is exactly where the strand places its equity case. And frankly, this is the starkest line in the entire document. Based on Department for Education data recorded in the toolkit, pupils on free school meals, so FSM, are suspended at more than four times the rate of their peers. Over four times the rate. Just let that sit for a second. That is massive. It is. And they are permanently excluded at around five times the rate.

And what is the most common reason given on the paperwork for those removals? Persistent disruptive behavior. Right. So we need to completely reframe this. We have to look at this through an entirely different lens. Behavior support isn't just a discipline issue wearing a pastoral hat. It is disadvantage policy. The kids with the least slack at home, the ones who literally rely on the structure and safety of the school building the most, are the ones the system removes the most. We're talking about an equity emergency here.

It is a total emergency. Yeah. When a child on free school meals exhibits persistent disruptive behavior, the data shows they are disproportionately more likely to be sent home than a more advantaged peer exhibiting the exact same behavior, which is just heartbreaking. It is implicit biases within the system, a lack of resources, the way we interpret conflict. It all compounds when we get behavior interventions wrong. we're getting it wrong at the absolute bottom of the attainment gap. That is a crucial way to look at it. So with those immense stakes to find, we really have to look at what kind of intervention actually delivers those plus three months and keeps those vulnerable kids in the building.

So tell me the findings. How do we get the payout? Well, the first major finding is that schools don't actually have to pick a single philosophy. Both universal approaches, meaning whole school or classroom level strategies for everyone, and targeted programs for specific pupils have positive effects. Okay, but what about the costs? There's got to be a difference there. Oh, they range significantly. Universal approaches mostly sit at the low to moderate end of the financial spectrum, usually just costing staff time and initial training. The targeted programs are where the invoice really starts to grow.

sometimes requiring external resources or highly specialized staff. But the shape of the intervention matters way more than the brand on the box, right? Exactly. The highest impacts didn't come from a specific company's branded shrink -wrapped package. They came from approaches with an academic component, where behavior support is threaded through actual learning or approaches built fundamentally on role play and rehearsal. Rehearsal. Let's dig into that mechanism. You mean literally practicing the situation before it actually happens. Why does that work so well? Well, it works because behavior is largely a learned response to stress or social friction.

Practicing the exact moment means taking that 12 .41 playground flashpoint and walking through it on a calm Tuesday morning. Right, when everyone's adrenaline is low. Yes. What he says, what you say back, where your feet go, how your breathing feels. When you rehearse it calmly, you are building a new cognitive pathway. It frames behavior as a skill that you train rather than a moral failing that you just punish into existence. OK, let's unpack this timing aspect because the evidence is really clear on frequency, isn't it?

It can't just be a one off thing. Frequency is everything here. The session should happen several times a week and they need to be sustained. The evidence suggests that running these interventions for a year or more appears most successful, with programs running two to six months producing the more lasting results compared to shorter stints. This is so important. It's the drip, not the splash, right? Exactly. We aren't talking about getting a flashy theater company to come into the school hall for a one -off 90 -minute assembly on respect.

That might feel good on the day, but it doesn't change the wiring. The toolkit says we need frequent, sustained, boringly consistent practice. which brings us directly into conflict with a deeply held staff room orthodoxy. If frequent rehearsal is the key, we have to ask a massive logistical question. Who should be doing the rehearsing? And with how many kids at once? Oh no, go on, break my heart. I'm sitting in the pastoral chair and I know exactly what everyone in the building expects. When a kid's behavior is falling apart, the absolute reflex from the staff is that they need one -to -one support.

Right. A trusted adult off -time table, just them in a quiet room talking it through. Well, the evidence does not agree. In fact, large group and whole class approaches, meaning six children or more, came out slightly more effective than one -to -one and small group work. Wait, what? The whole class? Or a group of at least six? That actually beats the trusted adult, off -time table, giving undivided attention. Slightly more effective. And I am keeping the word slightly in that sentence because it's the toolkit's exact word.

This isn't the evidence saying one -to -one is a complete waste of time. It's saying that the expensive, precious, one adult, one child model isn't the automatic, unquestionable best answer that the staff room naturally assumes it is. But why? What is the actual mechanism behind that? Why on earth would sharing an adult with five other kids be better than having that adult all to yourself? Because behavior doesn't happen in a vacuum. It happens in a social dynamic. When you pull a child into a one -to -one setting, you are isolating them from the very environment where the friction occurs.

The room itself, the presence of peers, is an intervention. Ah, I see. In a group of six or more, you get peer modeling. They see other children successfully navigating the script, which creates a new social norm. It also removes the intense, isolating spotlight of being the bad kid sitting alone with an adult. They get to practice the social navigation with an audience, but in a safe, controlled way. Okay, I hear the logic, I see the mechanism, but I really have to channel the corridor here. I have to push back hard on this on behalf of every practitioner and parent listening.

Because from the yard, one -to -one isn't a research position. It's an emotional one. Yeah, it feels like care. Exactly. It feels like care. When a kid is in absolute crisis, giving them a dedicated one -to -one person feels like the school is finally taking that crisis seriously. Telling a parent, don't worry, his behavior is spiraling, so we're putting him in a group of eight. Well, that sounds like a massive downgrade. He does. It sounds like he's been put back in the queue and we are just trying to save money.

And let's be totally honest about who is actually doing this one to one work in schools anyway. It is rarely a qualified teacher. No, it's usually not. It's a teaching assistant or it's someone in the midday supervisor's chair. It's someone unqualified, unbriefed, holding the hardest child in the building because everyone else has a class of 30 to teach. So, if the toolkit says a group of six works slightly better, fine. But somebody has to actually sit in a small office and say that to a parent who's demanding one -to -one support.

And the toolkit offers absolutely zero help on how to sell that to an angry mom. You're right, it doesn't. But this is where we have to bridge the gap between the data and practitioner craft. Yeah. Because you can absolutely manage that conversation if you frame it around the evidence we just discussed. How so? You don't sell it to the parent as a downgrade or a budget cut. You sell it as a deliberate evidence -backed strategy. You tell them, we are putting him in a specialized peer group because children learn emotional regulation best from watching other children succeed, not just from hearing in a adult lecture them in isolation.

We were building a team around him so he could practice his responses with his peers. Oh, okay. That is a much better way to frame it. You make it sound like an exclusive leadership program rather than, you know, the bargain bin of interventions. Exactly. So, accepting that group rehearsal works better than one -on -one, we need to guide schools on how to actually build these interventions without falling into the traps. Because this show does the caveats as main segments, not small print. Good. So what are the actual pitfalls we need to avoid?

There are two big traps when it comes to implementation. The first is the time trade. The toolkit says this in absolute terms. If the behavior intervention takes up more classroom time than the disruption it displaces, then engaged learning time doesn't actually go up. Right, so the cure ends up costing more lessons than the disease. You pull them out of maths for an hour to talk about their behavior, but their behavior was only costing them 20 minutes of lost learning a day. You can intervention a timetable to death if you aren't careful.

Precisely. And the second trap is staff time. Behavior interventions are incredibly hungry for staff time. The targeted ones are the hungriest of all. They require dedicated training, reliable cover, and total consistency. That is why the Strand explicitly says schools must pick programs that have actually been evaluated and they must actually train their staff to deliver them. Yeah, buying a glossy binder in July and skimming it on inset day in September isn't running a program. It's really not, which raises an important question. How does a school choose between the universal layer and the targeted layer if we know both can work?

Well, let's use the strand zone logic on this. There are some influences on behavior that a teacher can reach perfectly well with good, solid classroom management. That's the universal layer, and frankly, it's the cheap layer. Right. But there are other behaviors that need specialists to support self -regulation, specific social and emotional skills. That's the targeted layer, and that is the layer that costs money and time. The mistake schools make is buying the expensive targeted layer to compensate for a completely missing universal layer. That's an excellent way to frame it.

It's like building the ambulance before you build the fence. You're preparing for the crisis and paying top dollar for it instead of preventing it from happening in the first place with the cheap universal barrier at the top of the cliff. Exactly. So let's slow down and look at what this actually looks like in practice. Let's walk through the three rooms. Yes, show us the application. Break it down for primary, secondary, and specialist settings. Let's start with primary. Picture that rehearsal we talked about, but built directly into the weekly rhythm of the classroom.

You have the whole class, well over our six plus threshold, practicing this specific scenario. They are role playing, what do we do when we both want the same ball? And the key is they're doing it before lunch. on a calm morning, not after an incident has already happened, and adrenaline is high. A whole class doing this isn't just the budget option. The evidence tells us peer modeling makes it the better option. And how do we translate that to secondary? Because a class of year 10s aren't going to sit in a circle and role play wanting the same ball.

No, they're not. In secondary, we focus heavily on the academic component. The behavior support lives inside the English class or the maths class, not in a separate pastoral room instead of it. What does that actually look like, though? How do you thread behavior into an English lesson? It looks like a pupil rehearsing a specific reentry script. If a teenager feels a flare -up coming, or if they've just had an outburst, they know exactly what to say to come back into the learning. And the teacher knows exactly what to say in response.

Give me a concrete example. What does a reentry script sound like in that moment? It might be as simple as the pupil saying, I'm ready to rejoin the lesson, but I need to sit at the back for two minutes to read the text before I answer any questions. And the teacher responds, thank you, take your two minutes at the back. Right. The script is agreed upon in advance. Practice when the student is calm. So period five actually has a safe, predictable door back in. It removes the standoff.

I love that. It removes the audience effect where the kid feels they have to double down to save face. So what does this all mean for a specialist setting? Well, In a specialist setting, rehearsal might be the most natural fit of all. It's the flashpoint walkthrough using symbols or a social story. It's the exact same script used every single time, several times a week. Sameness and frequency aren't limitations here. The evidence tells us they are the active ingredients that build security and predictable outcomes. Which brings us right back to 12 .

41. It does. We are back on the playground. It's the same 40 meters of tarmac. It's the same football. It's the exact same joke about the cousin curdling into a tense argument. But this time, the two boys have rehearsed this exact weather. Now, not a heat of the moment, but weeks ago. They did it in a group. They did it more than once. They saw their peers navigated. Yeah. And just as importantly, you and the Hy -Vee's vest have been briefed. The midday team got told what the script was during their morning check -in.

So you step in and you say the line they have practiced hearing. And maybe it holds. Or maybe it only half holds. Because the evidence promises a plus three months average progress over a year. It doesn't promise an infallible miracle at 12 .41 on a Tuesday. But period five starts at 10 past one, either way. The question is what mindset walks through that door with them? Do they have a fighting chance at learning? Have we given them the tools to deescalate? Okay, so we are moving from the theory back into the reality of the school day.

What are the immediate next steps? What can you do today? Well, I have a challenge for you to take to the kettle at break time today. Ask around your colleagues. who actually gets one -to -one intervention in this school? And more importantly, who decided that one -to -one is what works best? Is it based on evidence, or is it just inheritance? Or are we just doing it because it's what we've always done? And here is the data challenge. This is something you should do today before a governor or a local authority officer reviewing exclusions does it for you.

Pull your suspension data and split it by pupil premium status. Can you see that four times gap? Is it visible to you right now, or would it be absolute news to you if someone else pointed it out? Most schools have this data sitting right there on their dashboard. But not many have actually looked at it split that specific way since September. Proactive leadership means it's something you could show, not just something you promised to check. And if you are running a small group session this week, here is an immediate action.

Open your next session with just five minutes of rehearsal. Yes. Walk through a flashpoint before the academic content starts. Don't wait until after it has happened. Try this new shape on a group you already have in the room before you go out and buy a brand new program. Because shape beats brand every single time. All right, let's pull all of this together. Here are the four things off the drive today. Number one. plus three months. That is the toolkit's average impact for behavior interventions, holding a moderate security rating of three out of five padlocks.

Keep both labels on when you quote it. Number two, the free school meals gap is an absolute emergency. FSM pupils are suspended at four times the rate and permanently excluded at five times the rate for persistent disruptive behavior. This isn't just discipline. This is disadvantage policy. Number three, shape beats brand. You want interventions with academic components built on frequent rehearsal delivered little and often over two to six months. And number four, groups of six or more slightly beat one to one. Which means that trusted adult reflex we all have.

It deserves a second look, not a funeral. It's about deploying that dedicated adult in a group setting where the peers themselves become part of the intervention. It's about using the social dynamic to heal the social friction. Everything's linked in the show notes, the strand page, the padlocks, the lot. Read them. We're the trailer, not the film. Both voices on this show are synthesized. The judgment isn't. Safe trip in. See you at the gates.