+5 Months, Just 1 Padlock: The Weakest Evidence We've Quoted

This is the education commute. Both voices on this show are synthesized. The judgment isn't. Absolutely. So I want you to picture your classroom. It's, you know, two o 'clock on a Thursday afternoon. Oh, the danger zone. Right. The sky outside is this dull gray. The energy in the room is just flagging. And well, frankly, you are just exhausted. Yeah, we've all been there. Exactly. So you deploy one of the oldest, most reliable tricks in the entire teacher playbook. You just clap your hands and say, right, get into groups of four.

And within seconds, there's that familiar scraping of chairs, the dragging of desks, and a sudden burst of noise. It's a very specific sound, isn't it? It really is. And it feels like active learning. It looks dynamic. And I have to confess, sitting in the class teacher's chair today, I do this all the time. Like I must use that exact phrase four or five times a week without even a second thought. Just muscle memory at that point. Totally pure reflex. My default setting whenever I need to just shift the pacing of a lesson is to put them in groups.

But today we're looking at evidence suggesting that when I do that, I might just be giving, well, three out of those four kids a completely free ride. Yeah. And let me just reassure you right away. You are definitely not alone in that instinct. Well, that's a relief. Because, I mean. Sitting in the CPD leads chair for this conversation, really looking at our continuing professional development and how we actually train staff, I can tell you that almost every teacher shares that reflex. Right. It's universal. It is.

In fact, that exact instinct was present in practically all of the 212 studies behind this week's strand. 212? 212 studies. So when you rely on that off -the -cuff get -into -groups instruction, you're recreating the exact scenario researchers have evaluated across hundreds of classrooms. Wow. Okay. But that is precisely where our autopsy has to start today. Because as a CPD lead, I don't want us starting this autopsy by simply asking like a basic yes or no question about whether group work functions. Yeah, that's just too broad.

Way too broad. We have to start by dissecting what the word it even refers to across all those different studies. Right. That makes total sense. We need to define our terms before we actually judge them. Exactly. But before we get out the scalpels for this autopsy and dissect what it actually was, let's talk about the headline. OK, let's do it. Because, look, I know educational research often promises the moon. And if I'm already putting kids in groups of four just to survive a Thursday afternoon, what's the actual return on investment here?

Right. Because I'm naturally a bit skeptical that just rearranging the furniture yields massive academic results. What's the promise of this intervention? Well, the promise is actually incredibly substantial, which is it's why it remains so popular. We're talking about collaborative learning. And the headline figure here is plus five months of additional progress over the course of an academic year. Wait, plus five months? Plus five months. And remember, that's an aggregate. Backed by those 212 studies we just mentioned. Wow. So on paper, at least, you know, you're looking at a highly solid, incredibly cheap everyday win for your classroom.

Yeah. I mean, you don't need to buy expensive software or hire extra staff. Exactly. You just change the way pupils interact. Plus five months of additional progress just for collaborative learning. That's the headline. Honestly, if I can get nearly half a year's extra progress. Just by optimizing something I'm already doing, I will absolutely take that. It sounds like an incredible, almost unbelievable return on investment. It really does sound incredible. However, I recently ran this exact strand as an inset autopsy. Oh, right, for the in -service education and training days.

Exactly, during our inset days for staff. So I've had to sit down and really grapple with the cold reality of this data, not just that shiny headline number. Okay, I sense a but coming. A huge but. There's a massive catch. You just logically accepted the promise. But here is the vital metric you are absolutely not supposed to skip past. All right. Lay it on me. One padlock. Wait. Out of five. You're telling me the foundation trusts this data about as much as a chocolate teapot? Pretty much.

Yeah. We have this massive promise of plus five months, but a security rating of just. one padlock. That is exactly what I'm telling you. It's actually the lowest security rating this show has ever quoted for a strand that actually has a positive result attached to it. OK, I want to make sure I'm fully understanding the tension here, because that is just a wild contradiction. It is. We have a headline of plus five months of additional progress, which, you know, sounds like a miracle cure, but a security rating of one out of five.

So what the plus five months is technically real, but we basically shouldn't trust it to happen in our own rooms. That is probably the most honest read of the situation. Wow. Yeah, and to understand why it sits at that abysmal rating despite the high number, we really have to look closely at the exact reasons provided by the Education Endowment Foundation. Right, the EEF. Yes, the Education Endowment Foundation. They're the ones who evaluate these interventions, and they gave it that one padlock rating for three very specific reasons.

Okay, what's the first one? First... there are surprisingly limited recent studies a lot of the foundational research on this is just getting quite old okay fair enough second there's a high share of evaluations that were not independently checked. Ah, let me stop you there. Not independently checked. Does that basically mean researchers were just marking their own homework? Yes, exactly that. Like they designed the group work intervention, they ran it, and then they evaluated their own success without any outside verification. Precisely. It introduces a significant risk of bias, but it's actually the third reason that is the most important for us today.

Okay, what's the third? The final reason for the one padlock rating, and crucially these are their exact words, is that there is a large amount of unexplained variation between the results. Unexplained variation. Yes. Okay, let's translate that academic phrasing into what it actually feels like on the ground in a busy school. Let's do it. Does that mean that out of those 212 studies, some teachers implemented collaborative learning and their kids learned brilliantly, making huge strides, and then other teachers in different schools did what they thought was the exact same thing and it barely moved the needle at all?

That is exactly what it means on the ground. And if we connect this to the bigger picture of our CPD, it's why doing a deep autopsy matters infinitely more here than it does on, say, a high -security five -padlock strand. The average is so deceptive. Totally. The average outcome of plus five months is mathematically real -like. If you add all the studies up and divide them, you get that number. Yeah. But that average is hiding a massive chaotic amount of disagreement underneath it. So the results are just everywhere.

Incredibly scattered. Some classrooms saw transformational gains. Others saw absolutely none. And some might have even gone backward. Wow. Yet all of them were operating under the exact same broad, generic umbrella of group work. So the average is a mathematical truth, but basically a classroom lie. Yeah, that's a good way to put it. It's like saying if I have my head in the oven and my feet in the freezer, on average, my temperature is perfectly comfortable. Exactly. We have to figure out how to find the actual signal in all of that statistical noise.

Right. So if the average hides all this disagreement, let's get into the mechanics of it. What specifically... like mechanically, separates the successful studies, the ones that actually achieve the plus five months or more, from the unsuccessful ones where the kids basically just wasted an afternoon. Well, the first critical mechanical detail separating the winners from the losers is the sheer size of the group. Oh, really? Size matters that much? The huge impact. The successful studies consistently feature groups of three to five pupils. And it's not just the physical headcount itself.

It's the dynamic that size creates. Those three to five pupils must hold genuine joint responsibility for a single outcome. Interesting. Yeah. The data specifically shows that smaller pairs often lack the diversity of thought needed for true collaboration. And bigger groups, like groups of six or more, they actively encourage social loafing. Social loafing. The absolute bane in my existence. Oh, I know. But, you know, my default groups are actually usually exactly four simply because of how the desks fit together in my room. Right. So I suppose on that specific bit, the group size of three to five pupils, I've actually gotten it right by pure geographical accident.

Well, you might have the numerical headcount right by accident, but the element you almost certainly haven't got right is the next crucial piece. Oh, here we go. What is it? Structure. The toolkit's language is very direct and unyielding on this, and I want to read you the exact structure quote from the foundation. Okay, let's hear it. They state that, Structured approaches with well -designed tasks lead to the greatest learning gains. I need to heavily emphasize that point. Just pushing four desks together and saying, discuss this topic, is absolutely not what succeeded in the successful studies among those 212.

Okay, I have to seriously push back on my own practice here. Because if I'm completely honest with myself, that is exactly what most of my group work consists of. I think that's true for a lot of us. Yeah. My instructions are entirely structural in a physical sense, not an academic one. Yes. I just say, get into groups of four. They drag four desks together. I throw a broad discussion prompt on the whiteboard and I tell them they have 20 minutes to talk about it. And that's it.

that is the entire extent of the scaffolding four desks one prompt 20 minutes well if we look critically at the data your approach is almost certainly sitting right in the middle of the unexplained variation pile ouch yeah sorry it is definitely not sitting securely in the plus five months pile unstructured open -ended discussion prompts just do not reliably generate that promised academic progress they generate noise but not necessarily learning Okay. Okay. So if my 20 minute open prompt isn't earning the plus five months, let's break down the mechanics of a success.

Let's do it. What does a study that actually earned that high number look like? What is the practical step -by -step fix for all that unexplained variation? A study that genuinely earns that high number features a highly defined joint outcome. Meaning what exactly? Meaning the group is entirely jointly accountable for producing one specific thing. Furthermore, you need to assign specific roles or rigid task structure that mathematically stops one highly motivated pupil from just, you know, carrying the entire group on their back. I see. I think of this like a dinner party.

Okay, go on. My unstructured group work is basically a bad potluck where nobody coordinated the dishes. Oh, no. Right. Everyone just shows up. You end up with four bags of tortilla chips, no main course, and everyone leaves hungry but polite. Yes. But what you're describing with structured group work is a precise recipe where every single student has a specific vital ingredient to chop. And if the student with the onions doesn't do their job, the whole recipe fails. So the group inherently holds them accountable. That is a phenomenal way to visualize it.

The potluck is the unexplained variation. The precise recipe is the plus five months. Makes total sense. And crucially, to make that recipe work, there must be explicit, deliberate attention paid to your lower attaining pupils. How so? Well, the evidence is incredibly specific and quite brutal here. Without careful, intentional task design, those lower attaining pupils end up as passengers, not participants. Passengers, not participants. Wow. Yeah. I have absolutely vividly watched that happen in my own classroom more times than I really care to admit. We all have.

You look over at a table of four. You see three kids leaning in, heavily debating the topic, doing all the intellectual heavy lifting, while one kid has physically leaned back in their chair and is just. Silently, meticulously drawing an elaborate border around the edge of the poster paper. A classic border drawer. Right. They are technically in the croup, but they're entirely insulated from the actual learning. Exactly. They're in the vehicle, but they aren't touching the steering wheel. Wow. And the important thing for us to realize as educators is that this dynamic is a failure mode with an actual name now.

It's not just a bad afternoon in your specific classroom. It is a documented systemic failure mode seen across these studies. So how do we fix it? To fix it, to prevent the border -drawing passenger, there is one more vital piece that is incredibly easy to skip. It's the realization that pupils simply do not collaborate well automatically. We have to treat collaboration as a taught skill. Collaboration as a taught skill, meaning I can't just expect them to inherently know how to negotiate, compromise, and delegate effectively the very moment I push the desks together.

Precisely. We assume it's an innate human trait, but in an academic setting, it isn't. The toolkits line explicitly states that pupils need support and practice to work together. It's a discrete, taught skill, just like long division or paragraph structure. You have to explicitly teach them how to challenge a peer's idea politely, how to synthesize two different viewpoints, how to keep time. That takes work. It does. And to facilitate that professional development, that inset time we mentioned, is crucial to help teachers learn how to manage and teach those skills properly.

This completely reframes how I think about that quick Thursday afternoon instruction. I mean, I thought I was giving them a break. Right. I was actually throwing them into a highly complex social dynamic without any tools. Pretty much. So the honest autopsy result here is this. The plus five months number is mathematically real. But our confidence that it will happen by default isn't. And almost everything that makes the difference between a successful intervention and a total failure is the hard stuff I've been skipping. You know, designing joint accountability, building rigid task structure, and actually taking the time to teach pupils how to collaborate before I ever expect them to do it.

That is exactly what survived the autopsy today. The conclusion of this data isn't group work doesn't work. We aren't saying you should never push the desks together again. The conclusion is that unstructured group work is a total coin flip. The strand is here to tell you exactly which coin you need to stop flipping and how to weight the dice in your favor instead. So I can actually keep saying get into groups of four to survive the end of the week. Yes. But only if they are working toward an actual joint product where everyone's input is mandatory, not just me throwing out the word discuss.

And, crucially, with the defined role for each of them, at least until they've had enough deliberate practice with collaboration to not need those training wheels anymore. I love that. I have a question for anyone listening to take into their school tomorrow to test this out. Oh, let's hear it. Let's call it the kettle question. The kettle question. Yeah. At break time, ask whoever happens to be standing next to you by the kettle. In your last group task, could you actually name what each individual pupil produced or decided?

Oh, that's good. Or would the honest, uncomfortable answer be the group produced something between them somehow? That is the perfect diagnostic question for a staff room. It instantly separates the potluck from the recipe. Right. And I actually have a second one to add to it, which we can call the bet question. The bet question. Okay. If you had to bet your own money on this strand's promise of plus five months holding up in your own classroom, would you confidently bet on your usual off -the -cuff get -into -groups instruction?

Or would you only place that bet on a version that featured a tightly defined joint outcome and assigned roles? Honestly. If it was my own money, not the version I'm currently running, I would absolutely not bet a single penny on my current version. I would lose that bet. So let's solidify this shift in practice before we wrap up. Let's get four things off the drive today. Let's do it. One collaborative learning promises plus five months of progress drawn from 212 studies, but it is rated the lowest security the show has ever quoted.

Just one padlock. Two. The phrase unexplained variation between studies is the honest headline here. It's just as important, if not more so, than the plus five months number itself, because it proves that simply doing group work guarantees absolutely nothing. Three, if you want to succeed, keep groups of three to five pupils and ensure they have genuine joint accountability for one single outcome. That is the common thread what the better performing studies share. And four, collaboration is a taught skill. not a natural instinct that pupils magically arrive with.

Careful structure and deliberate practice are what separate this strand's massive win from its frustrating point flip. Everything I have referenced today is linked in the show notes. And there is a printable one -pager specifically for this topic available at theeducationcommute .co .uk, perfect for parents or to leave out in the staff room. Both voices on this show are synthesized. The judgment isn't. Safe trip in. See you at the gates.