As part of a round of high stakes observations I had two separate members of SLT come to observe my lessons*. In feedback, the first called me out on something which I strongly disagreed with. Right then and there however, I couldn’t quite explain to him why it was that I disagreed; I just knew that I did. So I went away and had a think, and after my second observation I was ready:

SL: I really enjoyed the lesson, the students have clearly been working hard over time and are committed to doing well in science.

Me: That’s good to hear. (context: very difficult year 11 set in the run up to GCSE; I had taught them for 18 months)

SL: However, they didn’t really have a good answer to the question “what do you need to do to improve?” (this is the same thing the first SL had called me out on)

Me: Ok. What did they say?

SL: Well, they all just said that they needed to work hard outside of lessons, focus more in lessons and get all the work done, and make sure they are listening to your advice.

Me: what’s wrong with that?

SL: It isn’t specific enough. It just wasn’t a good enough answer to the question.

This was what I was ready for.

Me: What answer to the question would be good enough?

SL: Um I’m not a subject specialist so I wouldn’t like to say

Me: Fine, but I am a subject specialist, and I don’t think it’s a very good question.


At this point, we had a very polite and cordial discussion about exactly why I didn’t think this was a good question. After the conversation, we went our separate ways. We were both leaving at the end of the year so neither of us were in the mood to push it further; least of all for some petty accountability measure.


When Dylan William and Paul Black published Inside the Black Box, an AfL “revolution” began in schools across the country. This revolution promised higher quality pedagogy and raised standards across the board through the implementation of formative feedback (**). Schools and their masters lapped it up and formative feedback takes pride of place as part of the standards that a teacher must meet before they can be given qualified teacher status.

Unfortunately, this revolution seems to have had limited impact. Robert Coe writes that:

It is now a rare thing, in my experience, to meet any teacher in any school in England who would not claim to be doing Assessment for Learning. And yet, the evidence presented above suggests that during the fifteen years of this intensive intervention to promote AfL, despite its near universal adoption and strong research evidence of substantial impact on attainment, there has been no (or at best limited) effect on learning outcomes nationally.

Many different reasons have been proposed for this lack of improvement. Perhaps, as Coe suggests, the evidence in favour of AfL is not generalisable and too context-dependent to be of wide use.

It is also possible that, once we have made the assessment, the feedback we give is of a poor quality and more likely to retard than to promote progress. Kluger and Denisi famously argued that over one third of feedback interventions actually lead to negative progress. Are all teachers aware of the difference between productive and counter-productive feedback? Are all senior leaders aware? Ofsted inspectors? Educational consultants?

Daisy Christodoulou has argued that in our formative feedback we have been focussing too much on the generic skills that student performance represents. We have tried to establish which generic skills students are progressing through rather than focussing on specific knowledge; an approach which is unlikely to bear fruit.

Dawn Cox and David Didau have pointed out the difference between Learning and Performance. AfL has become bastardised in our classrooms into a process of plenaries to assess student “learning”. This process has been driven by accountability as much as anything else and has seduced teachers into thinking that because their students can adequately address the mini-plenary now, learning has occurred. In reality, students are merely performing in the moment, not demonstrating long term learning. We are fooled into thinking that they have therefore learned the content.


My story above is just that; a story. But it illustrates a further problem with the AfL revolution in that it has, despite researchers best attempts, become domain-general; part of a generic set of skills that every teacher must have. Science teachers have been sent to art lessons and English teachers to maths lessons to observe “best practice” in the realm of AfL. Underlying this is the flawed assumption that such practice would then be readily transferred to their own classrooms.

Christine Counsell’s powerful piece on Bloom’s taxonomy also reflects this problem. In education we have failed to realise that teaching science is not like teaching art, English or geography. We must build a set of tools and techniques specific to our subject.

Coffey et al also make this point in their review of what has been considered to be “best practice” in science AfL. By analysis of teacher transcripts they convincingly demonstrate that in many discussions regarding AfL the “disciplinary substance” of the material under discussion is missing. Advice and “best practice” – even when research based – must be specific to the disciplinary substance of the material to be taught.


Over the next few weeks, myself and a number of colleagues will be hosting an online symposium on the topic of AfL in Science. We will be attempting to discuss and establish what good AfL in science looks like. We are teachers, not professional researchers, but we hope to bring an evidence-based perspective on AfL in Science and that the ensuing discussions can improve all our practice.

Each week, a new blog will be published on a different area. The blogs will aim to be coherent and to appear to flow naturally one to the next. The first piece is written by Rosalind Walker and attempts to understand and define a framework for what scientific learning really is. We cannot implement AfL if we cannot define the “L.” Niki Kaiser will then discuss the importance of “threshold concepts” and successful diagnosis of student understanding of these concepts. Deep Ghataura will then  post on the topic of what we do with assessments once we have made them. Which legitimate inferences can be made and how can we act on those inferences? Ben Rogers will then address the increasingly popular technique of Comparative Judgement and ask what place it might have in a science classroom. Matt Perks continues the symposium by looking back at the previous pieces and suggesting some firm practices which teachers can employ to effectively implement AfL for Science. Dylan Wiliam himself rounds off the symposium with a look back at the history of AfL and the symposium as a whole.

We hope that you find the reading stimulating and we look forward to hearing the voices, opinions and experiences of many different educators.


(*)    This occurred at a previous school

(**)    I use AfL and Formative Assessment interchangeably as I believe teachers generally use them interchangeably. Assessment for learning is a broad process aimed at collecting data on student performance and providing “formative” feedback which enables them to improve. It could be that some see a vital difference but I (and Professor Coe) do not.