I have been a teacher for 29 years, a Headteacher for 14 years and, at the age of 53, this much I know about how research evidence stopped me judging individual lessons.

Becoming a school which bases its practice on a combination of common sense wisdom and what the evidence says works has been transformative, especially when it comes to lesson observations.
Looking back over the years, the forces of external accountability drove me, at times, to do some stupid things. One of the dumbest was to observe a perfectly good teacher and then assign an OFSTED grade to the lesson. The teacher put on a show-lesson, I ticked lots of boxes and we had some meaningless data which supposedly judged the quality of teaching.
I then gave governors and inspectors the percentage of Outstanding, Good, Satisfactory and Inadequate lessons judged against the criteria for the four OFSTED judgements, all in the name of rigour. I was judging individual lessons, not the quality of teaching. Everyone was happy but no-one was improving their teaching.
If I had thought harder about what we were doing I would have realised the pointlessness of it all.
The Centre for Evaluation and Management at Durham University defines great teaching as that which leads to improved student progress, a startlingly obvious truth. The thing is, there’s no point in teaching that doesn’t result in learning. My old-school lesson observation regime never determined the impact of the teaching upon the students’ learning.
The next sentence is shocking but true. It took me a long time to realise that a Headteacher’s most important priority is improving students’ progress. Once I did appreciate this nugget of common sense, the common sense spread. It became clear that improving the quality of teaching was the key driver to increasing students’ progress.
It then dawned on me that leading our teachers’ learning was the best thing I could do, day-in-day out.
Writing this down is embarrassing, it’s so obvious.
Then Professor Rob Coe’s paper on how difficult it is to be precise when assigning a grade to a lesson was a game-changer.
Once I did comprehend the truth of the matter, I changed most things. Out went graded observations. Instead I asked colleagues a simple question, How would you like to be observed to help you best develop your teaching? I gave my teaching colleagues control over the observation process and changed its purpose from accountability to development.
The thing is, I help develop directly the performance of thirty teachers. I have a commitment to watch each one of those colleagues teach at least once a year. That requires planning time, observation time and debrief time. I calculate I can spend up to 200 hours a year undertaking observation-related activities.
Those 200 hours are precious and if, as a result of those 200 hours, the quality of teaching improves, I have done my job properly. So now, when it comes to seasoned teachers whose classroom management is a competent given, we co-plan small, but very targeted, evidence-based modifications to their teaching and then focus on the impact of those modifications upon students’ outcomes.
These changes to the lesson observation process have profound consequences. I now spend significant amounts of time with teachers planning lessons. I ask lots of questions about their pedagogy tweaks and the intended impact on their students’ learning.
Counter to popular wisdom, seeing teachers teach is not an essential part of the “observation” process. Often, with the securely good teachers, I don’t need to observe the lesson at all! With the help of the IRIS video system, those colleagues will film their lesson and choose whether we need to watch how they delivered the modification to their teaching. I trust their judgement.
With younger, more inexperienced teachers, I may watch a full lesson on video, or be in the lesson like some doddery old teaching assistant, flitting in and out of camera as I try to be as helpful as I can. But I do not make judgements, just help the teachers improve their classroom practice.
For all teachers, the most important part of our revised observation process can happen two or three weeks after the lesson when we look at the students’ work, or the outcomes of a test, to see if we can discern whether the pedagogic tweaks we planned together have had an impact on students’ progress.
If the students’ learning suggests a teaching technique has worked, we replicate it; if not, we work out why it hasn’t worked and either modify it or stop doing it. Whenever possible we try to trace the golden thread from teaching through to students’ outcomes.
Our Assistant Director of Research School, Jane Elsworth, was faced with yet another change to the GCSE specification. She talked me through the tweaks to how she taught the extreme weather unit and taped the lesson for herself. Two weeks later we looked at her students’ responses to an eight mark exam question on the topic. There was overwhelming evidence of an improvement in literacy levels, with one student with a C grade target gaining the full 8 marks.
We are creating a culture where colleagues accept the professional obligation to improve their teaching. When I met with several teachers this autumn to finish their 2016-17 Professional Development cycle, they had all analysed forensically their students’ examination results. The question level analysis allowed them to trace their deliberate practice through to the students’ ultimate outcomes.
One of our mathematicians was able to point out that the metacognition lesson we planned for her Year 11s on histograms seemed to have done the trick. She taught the lesson in the week between the final practice paper and the real exam. The class had gained 7% of the possible marks on the histograms question on the practice paper; in the real exam they gained 70% of the possible marks. Go figure.
With both teachers there was no need to observe them teach formally – I know they can teach. Instead we spent time planning the lessons together; they were then motivated to trace the impact of the tweaks to their teaching upon their students’ learning.
If you remove the fear from observations you can make them developmental. Judging the quality of teaching in your school does not depend upon lesson observations; rather, the quality of teaching can only be assessed over time, with wisdom, tracing the golden thread through to your students’ outcomes.
We have to stop doing the stupid things.

If you would like a place on our ‘Building Confident Research-leads’ programme you can sign up by clicking on the image below:


Previous ArticleNext Article

This post has 5 Comments

  1. Hi John thanks for your great insight as usual. I was just wondering what you would say to school leaders in a secondary school judge to be Special Measures where learning outcomes need drastically improving. They have moved away from grading lessons but are quite often asked about the percentage of good or better lessons. They have moved to colour coding but I think that this is just grading by another means.

    1. I would just keep focusing upon improving teaching. As I point out, lesson by lesson it can all be meaningless, but you can tell from looking in students’ books whether progress and standards are good or not. I think the improving quality of students’ work relative to their age is the best proxy for knowing whether the quality of teaching is getting better. The best test of a school is what is going on when no one is looking. Day-in, day-out good solid teaching is what we are after. No tricks, no show lessons, just solidly good teaching which is reflected in the work students produce.

  2. Hi John,
    I agree wholeheartedly with everything you say. I’m not a Head but I am the PD Lead and as much as I might cite evidence such as Coe et al. the judging process continues. I can see why some might value grading for accountability but I would be interested to know how you managed to change the attitudes of your team.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.