Here lies the three-part lesson: remembered by many, mourned by few. It is interred alongside its many relatives: the four-, five-, six- and seven-part lessons, the diamond-shaped lesson, the funnel-shaped lesson and many, many others.

Here lie also its dependants: the “engaging” starter and the “fun” plenary, who were unable to survive without it.

Cause of death? Cognitive science. Most notably, arguments from research that teachers should move away from viewing lessons as units of learning and that they should instead focus on learning sequences across a number of lessons. Taken together with the new Ofsted-inspired focus on the curriculum, this evidence has made planning for single lessons seem less significant, if not completely irrelevant, in schools. Indeed, in some corners of the profession, it has suddenly become very unfashionable to talk about individual lessons at all.

But, as you may have begun to notice, the death of the three-part lesson presents a problem. Because even if you do not want to plan in single lessons, and even if you believe a lesson is not a sensible unit by which to measure learning, you still have to be in lessons. It’s still how we deliver teaching. As such, careful structuring of those lessons would seem sensible.

So, if we have killed the lesson structure many of us were taught to use, what is it that we should be doing instead?

Let’s look first - in eulogy - at what we had. The basic structure of the “traditional” three-part lesson, while varying between individuals and schools, usually consisted of:

• A starter activity designed to “situate” the learning in a wider context (such as what the class already knows, the curriculum as a whole, or the “real world” ) in order to generate interest and motivation to learn.
• The learning activity (or activities) to introduce new content.
• A plenary session to review the learning covered in the lesson, often via some form of competitive or communal recap activity.

Perhaps surprisingly, it is not difficult to find a research base to support the use of such a structure. Generating interest and engagement in the lesson might appear to be a sensible strategy, given that self-reported interest in the topic of study has frequently been found to be positively related to learning. Additionally, plenary activities provide opportunities for retrieval practice, which has been extensively shown to improve learning compared to other activities, such as restudying, highlighting and rereading.

Plenaries could also provide teachers with formative assessment opportunities, which seems sensible given evidence that students who receive frequent feedback show substantial learning gains.

Who killed the three-part lesson?

So, if it all makes sense, why have some been so keen to see it six feet under in the graveyard of abandoned teaching ideas?

John-Jacques Rousseau has the answer. The very first sentence of his 1762 treatise Emile, or On Education, states: “Everything is good as it leaves the hands of the author; everything degenerates in the hands of man” (we will refer to this as the Emile principle - the idea that perfectly sensible ideas can become burdensome and ineffective chores when organisations try to implement them).

The three-part lesson definitely degenerated in the hands of school leaders. Many saw it not as a flexible classroom tool but as a way of squeezing every lesson into the same mould so they could better judge learning across specialisms. The neat three-section structure was also a convenient way to judge progress across a single lesson: plenaries became little more than a confirmation for anyone watching that the class had not just been awake for the lesson but had actually remembered something they were told 20 minutes previously.

We now know - and perhaps many of us knew then - that this does not make for useful learning. For example, the evidence suggests the most effective retrieval schedules are those that test knowledge repeatedly over weeks and months, not just once, a matter of minutes after learning the content. And the three-part structure made lessons a prescriptive, and restrictive, set menu of accepted activities, hardwired into teaching practice through official school lesson plans and pro formas. Reacting to the pupils’ learning via formative assessment went largely out of the classroom window.

So, when the evidence-informed movement came calling, the three-part lesson was a long way from the research base and, thus, it landed bang in the centre of the research-savvy teacher’s crosshairs. Science pulled the trigger but misuse put it in the firing line.

That left a crucial question: what should these teachers do next? Barak Rosenshine, it seemed, had the solution.

Those teachers who knew that the prescriptive nature of the three-part lesson was ill-advised wanted to find a more pick-and-mix solution to lesson planning that was flexible, evidence informed and that would fit into the longer-term planning now advocated. In Rosenshine’s Principles of Instruction, they found a ready-made list of ingredients that fitted the bill. His report synthesised 10 key pedagogical principles from three strands of research evidence:

• Cognitive science research on how our brains acquire and use new information.
• Instructional research on the classroom practices of teachers whose students show the highest learning gains.
• Findings from “cognitive supports”, research studies that taught learning strategies to students (Rosenshine, 2010).

The 10 principles highlight a number of themes, including regularly reviewing previously covered material, using questions frequently to accurately assess understanding, carefully sequencing and breaking down new material, and allowing repeated, regular practice. How might this be used for planning individual lessons? Here’s a good example.

Jennifer Beattie, director of modern foreign languages at Unity Schools Partnership, a multi-academy trust with schools across the south-east of England, argues that Rosenshine’s principles fit neatly into an existing lesson planning aid, already used in the trust, of “I do, we do, you do”.

In a Rosenshine-inspired version of this system, “I do” would include teacher-led aspects of the principles, such as presenting appropriately chunked new material and providing models. “We do” would involve the collaborative aspects, such as worked examples, questioning and reviewing previous material (appropriately spaced over time). “You do” would cover student activities such as scaffolded practice designed to lead to a high success rate (Rosenshine recommended students be able to successfully achieve 80 per cent of a task).

There are no time limits on each section and there is no prescription of what each section should look like. It’s a flexible way of thinking about lesson planning into which Rosenshine plugs some evidence-informed options for teachers to choose from.

Great, you might think. We’ve found our solution! Except, not everywhere is using Rosenshine as Beattie is. In fact, in many instances, use of Rosenshine is going the same way as the three-part lesson.

That’s right, the Emile principle has sneaked back in. “Rosenshine lesson plans” are now widely available online, some frantically trying to crowbar all 10 principles into a single lesson. Other schools have adopted “research-informed” lesson plans, in which all subjects must adopt a uniform lesson structure based on an application of Rosenshine’s principles (for example, a quiz, followed by teacher modelling and instruction, followed by a practice phase).

If you hadn’t guessed already, it’s not supposed to be used like this, as Tom Sherrington, a former headteacher and author of a book on the principles, explains. “Where Rosenshine’s principles are useful is in providing structure at the level of a learning sequence spanning several lessons, not individual lessons,” he told Tes.

“Beyond that level of guidance, mandating the form or frequency of any of those ideas is unhealthy because the context of each class and subject will vary so much, with ebbs and flows of fluency from lesson to lesson: you will only generate robotic, unresponsive teaching practices, not confident, skilful teachers.”

This is not the only problem that has emerged. Rosenshine’s principles are not the complete package of evidence we can pick from, and yet they are collectively being used as the ingredient, not an ingredient.

Pay it forward

What’s missing? The list is potentially endless. But here’s one interesting example of research that has not yet been widely discussed in teaching circles. It concerns something known as the “forward-testing effect”. This is the premise that regular testing not only improves the learning of content covered but can also prime a pupil to better learn future content.

In one study (Szpunar, Khan and Schacter, 2013), a video lecture was split into four chunks: one group of students were given a short quiz on the content they had just watched after each of the first three chunks (a structure called interpolated testing), while the other group reread their notes in that time. Both groups were then tested on the content of the fourth chunk, and the group that had been tested three times previously remembered more.

The study replicates a number of other findings showing a “forward effect” of testing on learning. David Shanks from University College London - a leading authority on the forward effects of testing - is “99.9 per cent sure that if you are doing repeated cycles of testing, then you will also be getting forward effects on learning, as well as backward effects”.

He explains: “If students are getting used to the idea that they will be regularly tested, then this expectation in itself seems to enhance the learning and retention of new material and, of course, the test itself will help to consolidate previous material. You’re getting the forward benefits of new learning and attention on top of the backward benefits on retention.” 

There are lots of interesting areas of research like this out there that could be used to inform lesson planning; focusing purely on Rosenshine means you are unlikely to see - or use - any of it.

So, what does all this mean for Rosenshine’s principles? Should we ready a place in the soil for them alongside the three-part lesson, all the better to hurl ourselves headlong into the next “evidence-informed” mandate that arrives from the senior leadership team?

Instead, maybe we should all take a step back. The examples above of what has happened to lesson planning using Rosenshine should tell us something important: instead of any one evidence base, it is the Emile principle that reigns supreme in schools. So, rather than continuing to tackle lesson planning itself, maybe we should tackle what goes around it instead?

Here’s what we think we should do, put as simply as possible: trust teachers.

If we want learning to happen, it seems essential that teachers are trusted to apply research evidence to their own classes as best as they see fit - to be given access to the full banquet of evidence in order to build their “plate” based on what they know of their pupils, what they know of themselves and what they are doing in the classroom at that moment. It’s not a three-part plan, or a Rosenshine plan or a communicative language-teaching plan: it is their plan, tailored to their students.

This plan presents challenges, of course. How do we begin a mindset change among school leadership? How does the possession of a range of evidence-informed lesson-planning tools square with the understandable desire of many teachers for clear and consistent classroom routines (such as previously outlined in these pages)?

How do we make the expanse of relevant research visible to teachers so they have a genuine choice? And how does a call for teacher professionalisation and independence work in the case of new trainees or newly qualified teachers who may benefit from more explicit guidance?

These are not easy issues to overcome. But trying to solve them would be a much better use of everyone’s time than trying out, and ultimately killing off, yet another quick fix.

If we truly want better learning in our schools, then the legacy of the three-part lesson must be that we don’t repeat the same mistakes. We can’t simply adopt the latest fad, replacing one regimented system for another; instead, we need to change the whole way we see single-lesson planning. We need to recognise that effective learning relies on teachers structuring each lesson based on the specific requirements of the learning within it, and only then choosing activities designed to assist that learning, based on their expertise and a familiarity with a wide research base.

Essentially, we need teachers to be informed and we need them to be trusted. That is a message that we hope that schools can all get behind. Maybe even enough to put it at the top of their lesson plans.

Michael Hobbiss is a post-doctoral researcher and teacher. Jenni Kemp is a teacher at Bourne Grammar School

This article originally appeared in the 17 January 2020 issue under the headline “The untimely death of the three-part lesson”