The Biggest Problem With Lesson Visits

How to improve your English department so it is the best performing team in the school. Click here.

David Didau ran the English teams for Ormiston Academies Trust.

He more than knows his stuff.

His books have influenced thousands of English teachers in positive ways - he translates research and cognitive science into practical strategies. He ought to have a massive impact.

But he hasn’t yet. I think he was there for 3 years. Ormiston Trust’s progress 8 in 2024 was -0.33, and the progress for the English teams is almost identical.

Although I am using David’s blog, this is a useful case study. Schools get poor progress because they don’t think about scale. Research schools don’t make any extra progress, year on year, because they don’t think about scale.

You are probably not getting the progress you could, because you aren’t thinking enough about scale.

What is Scaleable?

School improvement has to begin with research, with best bets. But it succeeds on one thing, and one thing only: is it scaleable?

You can’t scale a teacher. In his latest blog Didau tells us about his experiences modelling, teaching live. You can read his blog here. However brilliant the lesson, it isn’t scaleable. It is therefore an inefficient practice.

This is the biggest problem with lesson visits. What you learn needs to be scaleable.

(Remember, it is actually the biggest problem with everything schools try to do - they don’t think hard enough about how to make good practice scaleable).

Let’s look at what David Didau has noticed in perhaps hundreds of lesson visits in the Trust, and I’ll show you what I mean by a scaleable approach. He says:

I also get to watch a lot of lessons. All too often, lessons end without teachers having done anything designed to give them a clear idea about what students think.

Here are some scaleable solutions:

• All lessons should involve cold call to find out what students know.

• All lessons should involve turn and talk (15-30 seconds per question), followed by cold call to find out what students think.

• All lessons with writing should include cold call of students work to show call under the visualiser. Ask question from the successes and the failures, and this will reveal misconceptions.

• When you visit lessons, count where these happen and where they don’t. A disorganised teacher might only manage each once a lesson, an organised teacher many times more.

• (You might not like ‘all’, but choose 50% of lessons, or 70% - whatever you choose, go out and count it. Choose a KPI: counting it will make the team keep to it).

Although many schools profess to operating a policy of cold calling students to check understanding, this often fails to permeate into every classroom. Many teachers ask too few questions to too few students. I've lost count of the times I've seen teachers ask a question, take the first answer given (often from a students with their hand up) as evidence that the class's understanding is secure and move on. I make a habit of asking the student next to me if they a) heard the answer their classmate gave and b) if they understood what had been said. Typically, they have little idea of what has happened.

How to make this scaleable?

• The schools are in the Trust. Either you can make cold call mandatory, or you can urge each English team to do so.

• Never lose count! Count everything you value. Record the numbers. ‘I have been into 7 lessons for 10 minutes each. The teachers asked a total of 50 questions. Two teachers asked 15 each, and the other 5 teachers averaged 4 each. All teachers relied on volunteers. Out of 50 questions, 41 went to volunteers. 28 of the volunteers selected were girls.’

• Ask teachers this question: if this pattern continues every day, what are the likely results in student progress for boys, girls, for students who do not volunteer?

• (Do your assessments show this is what is happening? You know they probably will).

• And after that, get them to set a KPI - ‘we will use random questioning X% of the time.’ Put a number on it. Then the department has to count.

As well as labouring under the pressure of having to cover content at the expense of students' learning what's been taught, I've come to the tentative conclusion that many teachers are nervous of carefully probing their class's understanding in case they find out students don't know or understand what has been taught. Lessons are, after all, a performance and if kids 'don't get' what you've spent the last hour teaching it's easy to feel like you've failed.

Make it scaleable.

• Our insights are personal. They are tentative because we aren’t sure, or because we don’t want to offend, or we want to win hearts and minds.

• Numbers solve this for you. You will never be tentative that 50 questions were asked, and 41 went to volunteers etc.

• You’ll win hearts and minds even more quickly if the head of English comes with you, and they count the numbers.

• When you ask the department (for whatever you are counting) if this pattern continues every day, what are the likely results in student progress? you’ll get some honest answers. Enough teachers won’t try to hide behind nuance, and this will mean that the team can negotiate a KPI.

The difference between my poetry lesson and many of the lessons I observe is that I knew students hadn't understood whereas often teacher either have no idea of their class's understanding or, worse, have convinced themselves through insufficient sampling that students know things they're actually still very confused about. It should be clear that the more accurate our understanding of what students are thinking, the better we can address misconceptions and consolidate tricky points.

It may feel exposing to teach in a way that reveals students' understanding, it may feel like a risk to really probe how many students have actually grasped what you've tried to teach but, whether they do or they don't, it's always better to know.

The scaleable solution is a curriculum solution.

• One solution is to plan your hinge questions and put them in the lesson resources.

• But, as Didau’s post revealed, hinge questions can be misunderstood. They are likely to make sense in the mind of an expert, not a novice. I’ll come clean. I don’t think hinge questions move the dial.

• Here’s a scaleable alternative. Teach 3 skills of a genre of writing. Teach from a ‘good model’, where these skills are done reasonably well. Next, teach from an ‘excellent model’, where these skills are done brilliantly. (Yes, you can reverse the order of the models).

• Then provide a new ‘good model’. All students focus on the 3 skills to turn it into ‘excellent’.

• This will always reveal misconceptions. Pair work or the visualiser will always correct the misconceptions. And so will the ‘excellent model’.

• With David’s example, teach from two paragraphs, a good and an excellent, where both deal with the ‘correct’ comparison of both poems. Then give a good paragraph which relies on an ‘incorrect’ comparison. Ask them not just to correct it, but rewrite it as excellent.

• Your teachers need no new skills and no new habits to do this. The lesson resources and structure make them do this. It is 100% scaleable. The 3 most important skills to teach are already identified. Students learn, even with your least effective teacher.

• That’s what scaleable means. And that means improvement is inevitable.

If you are or a know a teacher who wants to become a school improvement partner, I help run a course to train you.

It is the only such course in the country accredited with a quality mark, and the other leaders are both ex lead inspectors, who have been SIPs to over 100 schools between them.