Rosenshine in action…

Putting theory into practice

CPD Books

A Great blog from Mark Esner (Heathfield School) on how he is using Rosenshine in the classroom: Full blog page is here

I love a good teaching and learning book, as anyone who has popped into my classroom or tried to find something on my desk can attest. I also really enjoy attending research conferences (I’m speaking at ResearchEd Durrington this weekend and ResearchEd national conference in London in the Autumn) and find leafing through a research paper relaxing. I’m weird like that.

One question I am sometimes asked is what difference all this reading, tweeting and writing about education actually makes in the classroom. I’d like to take one example of a research review and show how I have applied it.

Rosenshine’s Principles of Instruction (2012)

Barak Rosenshine’s article may be the most useful thing ever published on teaching and learning. It takes research from cognitive science in how people learn, the cognitive supports that make this process easier and from looking at highly effective teachers in the classroom. He suggests that excellent teaching contains ten key characteristics.

1. Begin each lesson with a short review of previous learning

Starting a lesson by recapping things the class have already learnt means we can take advantage of the testing effect, which suggests that every time we recall information we make it easier to access it again in the future. It also means we have somewhere to hook new information so that it fits with what we already know, we start to build a more complex picture of the subject.

I do this by ensuring that the start of the lesson is dedicated to recall. This will often be a short quiz (10 questions on one slide and answers on the next) or one longer question to which they need to apply what they learnt previously. I make sure that these questions link to the topic that we will be about to cover. For example, before a lesson on rainforest management I’ll include questions on low pressure weather systems, the nutrient cycle and sustainability. These are concepts they will use in the lesson and strengthening recall now will support their working memory later.

2. Present new material in short steps with students practicing after each step. 

When I started teaching we were encouraged to limit the amount of time we spent at the front of the class (I am sure that most of us will have been told the myth about pupils only remembering 10% of what they are told but 90% of what they discover for themselves) and to set long open ended projects for pupils to complete during the lesson. This would allow them to explore the task for themselves and construct their own meaning. This form of minimal instruction “discovery learning” leaves pupils swamped with information they struggle to process and lacking the guidance to make sense of it. This influential paper by Clark, Sweller & Kirschner (2012) suggests that pupils benefit from very clear and explicit instruction from an expert rather than the expert simply facilitating their discovery.

Rosenshine’s research found that the most effective teachers spoke for a total of 23 minutes in a 40 minute maths lesson compared to just 11 minutes from the least effective. Effective teachers used the extra time to explain new material very clearly, give lots of examples and asked lots of questions (see below). This time wasn’t in a block but spread over the lesson, interspersed with deliberate practice from the pupils.

This research has made me much more comfortable standing at the front and being a “sage on the stage”. I use a clear Input – Application model of teaching where I talk the class through something they need to know, give them examples, use analogies, show model answers before giving them a short task to do. They still carry out longer pieces independent work but only at the end of the phase, once I am sure the building blocks are in place.

3. Ask a large number of questions and check the responses of all students. 

Rosenshine suggests that not only do the most effective teachers ask a lot of questions, they also ask different questions; they are more likely to ask questions about the process that they have used to work out the answer. Questioning allows pupils to practice using the information they have been taught. It also allows us to receive feedback on their understanding and correct any misconceptions.

In the classroom I try to ensure that I target questions carefully (rather than using a random generator) and ask follow up questions. These might include

  • Why do you think that would be the right answer?
  • How does that link to what we know about X?
  • Can you explain that but include a reference to Y?
  • How would you know if that was the right answer?

I try to make sure that I receive feedback from as many pupils as possible by asking pupils to discuss it in a pair before sharing, using mini-white boards and by continuing to ask questions to small groups during the lesson. I try to think about the feedback I need from the class and focus my questions on common misconceptions and threshold concepts (See Meyer & Land 2003 and this piece on Threshold Concepts).

4. Provide models

Models and worked examples help to provide cognitive support to pupils so that they can focus on applying what they have learnt rather than concentrating on the form of the answer. They also allow pupils to see very clearly what your expectations are and allows you to set the bar high.

Over the last couple of years I have started using more and more models in the classroom. I try to show pupils examples of excellent answers and then unpick this answer with them so that they understand the criteria that makes this an excellent piece of geography. I also try to make sure I model things carefully where I know there are often misconceptions. Addressing the problem before it appears in their work.

Over time, it is important to remove the scaffolding that pupils get from modelling so that they can complete the task for themselves. I try to do this in a structured way. When they first try to draw a climate graph I will draw one with them and talk them through the process before giving them one to complete for themselves. The second time I remind them of the key points to remember and remind them of common mistakes before leaving them to complete it themselves whilst I circulate and support as needed. By the time they try it for a third time I expect them to be able to complete it with minimal modeling.

5. Guide student practice

As mentioned above, pupils benefit from guided instruction and having material presented in small chunks. It is important that they do something with this material as soon as possible. Using the information will mean they have to think hard about it and, as Daniel Willingham explains, memory is the residue of thought. It is also important that this practice is monitored. Some feedback, especially feedback on processes, might be best off delayed but feedback on the task, on tackling misconceptions, needs to be tackled immediately to prevent the error being embedded. Practice doesn’t make perfect, but it does make permanent. For more on different types of feedback and feedback during practice see Hattie and Timperley (2007) The Power of Feedback.

I try to apply this to my own classroom by making sure that lessons are built around answering “fertile questions” (See Benson and Knight 2013). These big questions are then broken down into the small steps that pupils need to take in order to answer them. This naturally encourages the lesson to be broken into smaller chunks with a focus on applying what they have learnt to a problem they have to think hard about.

I also make sure that I am monitoring the room as pupils are working and pick up on misconceptions as they happen. I am much more likely to stop the lesson to address a problem and reteach something than I used to be.

6. Check for student understanding

Rosenshine explains that the most effective teachers are always checking for student understanding whereas the less effective would ask “any questions?” before moving on. When we are building knowledge we are taking new information and linking it to other things that we know. This is the point where misconceptions can develop and information is mis-linked. For example, pupils learn that global temperatures are increasing and link that to what they have heard about there being a hole in the ozone layer.

In the classroom, I try to check for understanding by asking pupils to apply what they have learnt to a new situation. For example, if pupils have learnt how the nutrient cycle works in the rainforest and have learnt about the conditions in the desert, they should be able to apply one thing to the other and describe what the nutrient cycle will be like.

I am wary of plenaries at the end of a lesson to check for understanding for two reasons. Firstly, checking for understanding that close to the end of the lesson risks the chance of just getting mimicry rather than genuine learning. Secondly, the end of the lesson is too late to do anything about misconceptions you may learn about. If you check for understanding as they leave the room and discover most are walking out thinking the hole in the ozone layer is responsible for global warming, what then?

7. Obtain a high success rate

Following on from the points above, Rosenshine is clear that teachers  need to ensure that pupils really understand something before working independently on it. The idea that pupils need to be secure in a small step before moving on to the next one is sometimes called mastery learning. Most work on this seems to come from the context of maths where knowledge is structured hierarchically with a clear progression in the subject’s difficulty. Geography, and many other subjects, are organised cumulatively whereby pupils gain a greater breadth of  study over time but difficulty only tends to increase between key stages.

There are, however, areas where I have been able to apply some of this idea. As mentioned above, I am more likely now to stop a lesson to reteach something that clearly hasn’t worked and to set homework tasks designed to fill very specific gaps in what a pupil can do.

8. Provide scaffolds for difficult tasks

This point overlaps with that requiring the use of models. Scaffolding can take many forms but can be thought of as anything that lessens the cognitive burden of the task. This might include providing prompts to start the lesson, talking through the answer yourself out loud to show your own way of approaching it or providing  checklists for a task to make the expectations clear.

One way I increasingly use scaffolding is through the use of diagrams and images. As I explain an idea to the class I make notes on the white board, draw flow diagrams and pictures to illustrate the key points. This serves two functions.

  1. Pupils take information in on both a visual and auditory channel. Both spoken and written text uses the auditory channel and reading out a text as they are reading it themselves can make it more difficult to learn. However, talking over images can strengthen recall. This is the principle of dual coding (See Mayer & Anderson 1992 – The instructive animation)
  2. It also acts a reminder of what has been said. Spoken words are transient and too many ideas overwhelm the working memory and can’t then be applied to the task. By leaving a visual record of my explanation and thought process I am allowing them to use part of my schema to support their work and build their own.

9. Require and monitor independent practice

Most of the preceding 8 principles are about effective direct instruction. It is important however for pupils to have the opportunity to apply what they have learnt. Rosenshine suggests that pupils work best when the teacher is circulating the room and monitoring their work and when there is an opportunity for pupils to share their work with those around them.

As mentioned above, I tend to phrase my lesson as a “fertile question” that needs to be answered. This question is answered independently using what they have learnt. During this time I stand back more than during the knowledge building phase but look for pupils who may be struggling or who I know might be having a problem. I find that my classes work in near silence but I always remind them that they can talk but it needs to be “A whisper, to their neighbour about the work”. This allows them to check a point or receive feedback on what they have said from someone else more quickly than I am able to do so.

10. Engage students in weekly or monthly review

This principle returns us to the beginning. Pupils need the chance to review what they have learnt and to consider how it fits into a bigger picture. The better they can do this the more load is taken off their working memory as they can recall the information they need to undertake a task from their long term memory.

I try to support this regular review in a number of ways.

  • I sit with a pupil during a lesson and talk back over their work from the last few weeks. We look at the progress they have made and discuss what they have learnt.
  • We use learning checklists and knowledge organisers to help them see how what they have learnt fits into a bigger picture.
  • I make sure that the regular quizzes at the start of the lesson goes back over previous topics and not just material from the last lesson.


When re-reading Rosenshine’s principles again I am struck by just how simple it all sounds. But this shouldn’t be surprising. Teaching, at its core is simple. Recap previous information, input of new information, apply it, test and respond. However, doing these simple things well is complex and deserves some consideration.

Mark Enser


Rosenshine – practical examples!

Ten Principles of Instruction- Rosenshine, B. (2012) Explained from a classroom perspective via @ASTsupportAAli

Here are the principles in their entirety from the 2012 paper- Have them open as you read through the below…

1- Begin Each Lesson With A Short Review of Previous Learning– Start each lesson with a 5 question recall check. Give students 5 minutes to complete this. If they don’t know the answer, they write the question down and leave a space to fill in the answer later. You could do also do these 5 questions via multiple choice questions. Ensure you provide 2 plausible answers with 1 misconception. These must be high frequency and low stakes. Meaning you don’t take grades in, or ask students to share their scores out loud.

2- Present New Material In Short Steps With Students Practicing After Each Step. Explain tasks fully, ask students to repeat back the instructions given. Ask them to explain why you’re doing that task. Then work through examples of the completed task, model the answers. Work on items in front of your class under the visualiser, while they listen. Be OK with talking and explaining. Then get students to do. The check.

3- Ask A Large Number of Questions and Check the Responses of All StudentsAsk lots of probing questions. ‘What if?’ ‘How do you know?’ Vary your questioning techniques. Hands up. Hands down. Add, Build, Challenge. Gadfly questioning; socratic questions. Plan for your questions. Do not ask questions before using a random question generator. Share the fact that answers to questions are for everybody, and therefore everybody should list and be ready to add, build, or challenge given answer. When students answer a question, ask them if they are sure, how sure and how do they know.

4- Provide Models– Vary the way you present modelled answers. Students should know that it isn’t solely how much you write that earns you more marks. Ensure students see the process of a modelled answer. Annotate and break down examples of completed tasks. When giving grades back to a class, use aspirational marking and only give the marks off the next grade students are, rather than their current grade.

5- Guide Student Practice– Instead of objectives or intentions for lessons try setting Big Questions- enable steps to answer those big questions. Enable the discussion to take place that breaks the big questions down. Practice and repeat. Allow for live marking to tackle misconceptions immediately, do this collectively if you can? Whole class feedback could support this. 

6- Check for Student Understanding– Don’t just take a blanket response from the class to the questions. ‘Are we OK with this?’ ‘Any questions’ and so on. Enable enough time in your lessons for the students task to be checked for their understanding. How do you know they are good with moving onto to their next step in learning?

7- Obtain a High Success Rate– Enable students to have understood something before moving on. Think about teaching and reapplying concepts in 3 different contexts for true understanding and longer memory building. I call it ‘re-mixing‘ lessons. Students should not say, we have done this, unless, they have learnt it!

8- Provide Scaffolds for Difficult Tasks– These can be provided in a variety of ways- think about the concept of Dual Coding. Do you explain visually alongside your written and oral instructions? Do you have a consistent way of dual coding to avoid cognitive overload in lessons? Also have you considered linking abstract concepts with concrete representations. Are you able to bring something obscure into something tangible?

9- Require and Monitor Independent Practice– Circulate the room, check over tasks. Monitor as work is being completed. Don’t be afraid to ask for tasks to be completed in silence. Do not worry about getting students to edit tasks. I prefer the term edit, rather than re-do. Try Red Dot Marking?

10- Engage Students in Weekly or Monthly Review– Remind students you have not just taught them from last lesson, but from the start of the year. Remind them that every lessons learning is vital. Think about the spacing effect in your curriculum . How do you enable recall. Do you refresh on core concepts, skills and important threshold concepts throughout the year? What is your assessment cycle like? Quality first teaching is preceded by quality first planning.

Finally read this blog by on Rosenshine- and blog too- They have both allowed me to summarise my thoughts. Thank you!

Click on the tweet below to read my original thread. 


Download an amazing visual by Oliver Caviglioli here

Rosenshine – Whole School Approach

The Rosenshine Papers

A really interesting blog post form Steve Adcock on how they moved towards a whole school approach: Why Rosenshine?

In 2018 we (United Learning) adopted Rosenshine’s principles of instruction as the basis for our approach to teaching and learning across our schools. It’s the first time that we’ve taken a collective position on teaching and learning, rather than leaving this critical issue to each school. Our focus previously was on supporting each school in having an internally coherent and effective T&L strategy. With the adoption of the Rosenshine principles we were attempting to go a step further by ensuring that each school’s approach was anchored in a shared understanding of the characteristics of effective teaching.

We did this for a few reasons. Firstly, we wanted to support schools in challenging approaches to teaching that are not supported by good evidence, such as teaching which is overly driven by the exam specification, teaching that is founded on the belief that pupils learn better by discovering things for themselves, teaching that takes differentiation too far by placing different groups of pupils on different ‘tracks’ in the same lesson, and teaching that is overly focused on securing evidence of progress in each lesson, rather than gradually building a secure long-term understanding of each subject.

As a growing Trust, and a Trust that comprises primary and secondary schools in the state and independent sector, as well as an initial teacher training programme, we could see benefits in building a shared understanding of the characteristics of effective teaching. A trainee teacher could leave their summer institute and arrive at their school in September safe in the knowledge that the philosophy towards teaching and learning would be consistent; a deputy head leading on teaching and learning could share resources with counterparts in our other schools; subject advisors could produce curriculum materials confident that they would be applied in the classroom in similar ways. We would move from each school having an internally coherent approach to teaching and learning, towards a coherent approach across the whole group which would serve as a foundation for great teaching in each school and each subject.

Over time we are using the principles to develop a shared and precise language for the way we talk about teaching and learning. In my experience, the language commonly used to describe teaching and learning is anything but precise. Obvious examples would be phrases such as ‘the lesson lacked a bit of oomph’ or ‘pupils weren’t fully engaged’ or, more positively, the lesson featured ‘awe and wonder’. But even terms that seem more clear such as ‘pace’ and ‘challenge’ can lack the precision required to develop teaching practice. Take ‘pace’ – do we mean that the teacher went through things too slowly or that pupils didn’t work quickly enough, or perhaps the teacher wasn’t clear on timings, or maybe the start of the lesson drifted and time was squeezed for the challenging stuff at the end? That leads us to ‘challenge’ – was the content itself too easy, or was it the task, or are we simply saying that not enough pupils produced work at the standard required?

We chose the Rosenshine principles because they’re sensible, evidence-informed and provide the shared foundation we were seeking rather than a rigid checklist to be applied to every lesson. As an established set of principles we were able to avoid a long process of navel-gazing which would inevitably have been required if we had attempted to write our own. The fact they’ve been around for a while also enabled us to reassure our schools that we would commit to these principles for several years ahead, rather than replace them with a passing fad in twelve months’ time.

We’ve got a long way to go, but we’re seeing some early fruits of our labour.  I write this while returning from an inset day in Shoreham where all teachers from four of our schools started 2019 by gathering together to explore the principles in the context of their own subject. Meanwhile our subject advisors have written case studies on how to apply these principles in their subject. The curriculum resources we are producing contain the modelling, the question prompts and the scaffolds that Rosenshine promotes in his work.

So what might Rosenshine look like in the classroom?

As we’ve worked with schools in exploring Rosenshine’s work we’ve confronted the question of what his principles look like in the classroom. I’m in two minds here as to how usefully Rosenshine presented his research. On the one hand, I’m grateful that his principles are contained in short, concise pamphlets such as this 2012 one and this 2010 one. One of the simplest things we’ve done is simply ask schools to ensure that all teachers read all 9 pages of the 2012 paper.

But I do have a few gripes with the way Rosenshine presented his work. Firstly, the 2012 paper contains a list of 17 principles alongside the main list of 10. Rosenshine explains this decision (the list of 17 provides slightly more detail and overlaps with the list of 10) but given Rosenshine’s knowledge of the limits of working memory and cognitive load, it seems slightly curious to share two separate lists alongside each other.

We can take this overlap as a reminder that the principles do not seek to provide a checklist to be followed in order in every lesson. This becomes clear when we note his sub-heading for point 6 (check for student understanding): “checking for student understanding at each point can help students learn the material with fewer errors” (my emphasis). So – to be clear – we don’t check for understanding between point 5 (guide student practice) and point 7 (obtain a high success rate), we check for understanding throughout the whole process. Tom Sherrington has noted that this becomes clear when we read Rosenshine’s 1986 and 1982 papers which emphasise the importance of checking for understanding.

The 1982 paper also helps us understand Rosenshine’s intentions in proposing the principles:


There’s another gem lurking in his earlier papers that I think gets lost in the latter versions. In his 1986 teaching functions paper Rosenshine writes:

“Three of these functions form the instructional core: demonstration, guided practice, and independent practice. The first step is the demonstration of what is to be learned. This is followed by guided student practice in which the teacher leads the students in practice, provides prompts, checks for understanding, and provides corrections and repetition. When students are firm in their initial learning, the teacher moves them to independent practice where the students work with less guidance. The objective of the independent practice is to provide sufficient practice so that students achieve overlearning (Brophy, 1982) and demonstrate quickness and competence. A simple version of this core is used frequently in the elementary grades when a teacher says: “I’ll say it first, then you’ll say it with me, and then you’ll say it by yourself”.”

This seems like critical guidance, and helps us to understand the intention behind Rosenshine’s principles, which I think we can now summarise as:

  • Prior review
  • Instructional core (I>we>you):
    • Demonstration (explanation and modelling) of new material in small steps
    • Guided practice with prompts and scaffolds
    • Independent practice with monitoring and feedback from teacher
  • Future review

At each of these points – every single one of them – we check the understanding of all pupils by asking lots of questions and providing correction and feedback.

This model – the instructional core sandwiched between prior review and future review, with checking for understanding at each point – captures the essence of Rosenshine’s principles of instruction and provides an answer to that question of what Rosenshine looks like in the classroom.

Rosenshine’s back catalogue also helps us understand his 7th principle ‘Obtain a high success rate’.  In his 1986 Teaching Functions paper he writes: “Although there are no scientific guidelines as to exactly what the percentage of correct answers should be, a reasonable recommendation at the present time (suggested by Brophy, 1980) is an 80% success rate when practicing new material. When reviewing, the success rate should be very high, perhaps 95% and student responses should be rapid, smooth and confident.” So this idea of success rate supports teachers in deciding when to move through the instructional core, particularly when to move from guided practice (when around 80% of student responses are correct) to independent practice (when around 95% of student responses are correct).  This 7th principle seems a bit obvious and not overly helpful in the 2012 pamphlet, but it gains practical use thanks to the 1986 paper.

These principles now serve as a foundation for our support for teaching and learning across our schools. There’s a couple of things about foundations – in the sense of a building’s foundations – that I think are useful here. One is that we don’t tinker with foundations once they’re in place. They’re built to last. The second is that foundations are designed to be built on. We hope that throughout United Learning our teachers will explore these principles and bring them to life in the context of their school, their subject and their pupils. Rosenshine closes his 1982 paper with this very point:

in sum