Rosenshine is everywhere. And with good reason. As a basic primer for how to do Explicit Instruction, nothing is better. The paper is short, but jam-packed with good stuff about teaching and learning – good stuff which has been notably absent from teacher discourse and training for far too long. There are now books based on Rosenshine, CPD based on Rosenshine and entire academy chains using Rosenshine as the backbone of their teaching and learning policies.

To my mind, this is all good. Compared with the stream of unmitigated nonsense teachers have been subjected to for so long, the popularising of evidence based practices that don’t de-profressionalise teachers into glorified childrens’ party entertainers is definitely a Good Thing. Sadly, as a profession, the Ways Of The Past hold strong and there is a real and present danger of Rosenshine lethally mutating and becoming something which is at best unhelpful, and at worst actively damaging.

I was recently chatting with a colleague about one area where this could happen, and in particular relating to teacher exposition length. When I trained, I was always told not to talk for longer than five minutes or so as the students would become disengaged. An obvious evidence-free urban myth, but still surprisingly pervasive. But let’s imagine an assistant head gets up at whole staff briefing, puts a picture of a tweet by Dylan Wiliam about cognitive load theory on the board, and explains that the working memory can only process four or five things at any one time. So far so good. Because of this, it’s best to limit the amount of information you give to students at any one time. Still, so far so good. Rosenshine says present material in small steps, so far so good. Make sure to provide practice inbetween each step, so far so good. Ok, off you go have a great day. Cool.

Next lesson, I’m teaching the coal power plant. I need to explain how coal is first combusted, with the resultant energy release used to heat up some water. The water heats up and boils, turning to steam. The steam moves through the plant and pushes a turbine, increasing the energy in the turbine’s kinetic store by mechanical working, the turbine then turns the generator, and the generator then increases the overall energy in the National Grid. There are a lot of steps there, so we can imagine two classrooms:

In the first classroom, the teacher takes each step, and teaches them individually. Between each one, there is practice. First the students practise what burning coal looks like, then how it can heat up water, then how water can boil and so and so forth.

In the second classroom, the teacher takes the whole process, and tells the story of how energy from coal is eventually released into the National Grid. They work through all the stages, one after the other, then give the students a clear and detailed table summary with each stage, along with the energy stores and transfers involved.

The assistant head walks into the first classroom with his clipboard, and gives it a clean bill of health. Rosenshine: check. Cognitive load theory: check. Student practice: check. He makes a note to have this teacher present for a couple of minutes at the next all staff briefing. He then walks into the next room, where the same material is being taught, but in a very different way. Here, material is not being presented in small steps. Cognitive load does not seem to be taken into account. It’s half an hour into the lesson before the teacher-led discussion stops and the students get the chance to practise. Rosenshine has been broken. Frustrated, the assistant head leaves the room and makes a note to raise with the head of department that clearly the training from the whole staff briefing isn’t getting through to the science department.

But here’s the thing: the first classroom is fine. The students will probably learn the material and get on ok. But, in my opinion, the second classroom is better. There is a story to be told here – a narrative process that doesn’t really make sense except as a narrative, when the parts flow one into the next. And provided students can vaguely follow what you are doing as you move from one step to the next, it will make sense to them. They won’t think “what on earth is going on here?” as they struggle to mentally see the big picture. To use a now-hackneyed phrase, if all you show your students are the trees, they will never appreciate the woods. Actually that’s not even it – it’s more like “if they can’t see the woods, the individual trees don’t make as much sense.

Some content requires you to be more expansive in your exposition, to weave a tale that moves through a complex process from item to item, retaining your thread and grounding your students’ nascent understanding in a narrative sequence. There’s not much that you can do about that. Breaking it down too much runs the risk of it becoming granular and disembodied, with the learning experience feeling erratic, flighty and, ultimately, confusing – an explanation untethered from reality.

This doesn’t mean cognitive load theory is wrong. Your students’ working memory doesn’t magically expand because you are telling a story. What it means is that you need to compensate for the increased load. Cognitive load theory isn’t just about the number of things you are communicating, it’s about their difficulty, the support you provide your students and the way that new information interplays dynamically with their already existing prior knowledge. In a previous post, I put that like this:


Task quantity is how much new stuff you are giving them at once, and if it is high, then challenge increases. But you can moderate that – fight against it – with the other variables. The content demand is also fixed, but in this case, it isn’t that crazy – the abstractness of the process isn’t too difficult to actually get your head around in a fundamental sense. I can also ensure that student prior knowledge is high. If they are already super confident on combustion, energy stores, energy transfers, changes of state and the like, then that fights against the task quantity. And finally, I have external supports. Knowing that my explanation was long and multifaceted, I’ll make sure that students have a really great support to help them as they begin to practice, a clear and concise table or diagram that I will show them how to use as they begin to consolidate this new information through practice.

So you can “go long” – show students the woods and the trees at the same time, even if you know that you will have to help them in some other way to take in all that new information.

So I’m worried about our assistant head. I’m worried about our tick-box culture, and I’m worried that I may have contributed to it. I’m worried that we are going to take something cool and sensible like Rosenshine and be blinded by the surface simplicity and forget the dynamicism and complexity that underlines decisions about how specific content should be taught. We need to reject vehemently the simplicity that turns these ideas into banal soundbites which run the risk of turning the clock back on everything we’ve learnt. Indeed, to return to our hackeyed metaphor, we need to not get bogged down in the teaching and learning trees, ignoring the woods in their tangled and overgrown complexity.

Update: two hours on

I’ve seen a couple of comments online about this blog that don’t represent what I thought I was trying to say so I want to clear up a couple of things.

  1. This is not a “no best way overall” blog. I was very clear that in this case there is a best way to teach coal, and it is broadly using the framework of explicit instruction, just in my formulation, not in the “do this Rosenshine thing every time” formulation. I don’t think Rosenshine would disagree on the way that I choose to teach coal. The other way of teaching it, following Rosenshine to the letter, is good, but it could be better.
  2. As such even though I’ve said this is “breaking Rosenshine” it isn’t really, it’s just breaking a narrow and formulaic approach to Rosenshine. I’m not advocating project based learning, or group work or whatever. I am emphasising that explicit instruction as a teaching technique is highly effective, but complicated.
  3. This also isn’t a “it’s all about the kids” blog – the students are one variable in my equation – prior knowledge. There are generalisable principles that apply to all students.
  4. As such, this post has nothing to do with relationships. You could be teaching this content to a class you had never met before ever and everything written here would still apply.
  5. It’s also not about “telling stories” like you might from the history of science or whatever. This particular blog is about structuring your explanation as a flowing narrative where one piece leads to the next.
  6. The logic applies to any of Rosenshine’s principles, most of the time: follow them. But there are times where they need to be bent or broken or moved around depending on the other variables in the equation.
  7. Apologies for not being clear – given the subject matter of this blog that’s somewhat ironic.

An interesting further example is quantitative chemistry. One of the hardest branches of chemistry, there is a process students need to be able to complete, which we will call D. D is made of three sub-parts, A, B and C, so A + B + C = D. Now the thing is that A, B and C can’t really be understood individually, they are very much gears in the same engine. Equally, the engine doesn’t really make sense until you’ve laid the groundwork. In this case, the task quantity would be massive if I wanted to cover A, B, C and D at once. I could give a nice support, but even that wouldn’t help because actually the content demand is really high too. So there isn’t much I can do. Instead I level with the students and I tell them that “this won’t really make sense at all until three lessons’ time. I’m really sorry, but you’ll just have to bear with me.” But that’s the exception, it’s not the rule.